improv: colorblind colors

v1.16

.github/workflows/docker.yml

@@ -0,0 +1,34 @@
+name: Docker
+
+on: ["push", "pull_request"]
+
+jobs:
+  build:
+    runs-on: ubuntu-latest
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v2
+      - name: Set up Docker Buildx
+        uses: docker/setup-buildx-action@v1
+      - name: Cache Docker layers
+        uses: actions/cache@v2
+        with:
+          path: /tmp/.buildx-cache
+          key: ${{ runner.os }}-buildx-${{ github.sha }}
+          restore-keys: |
+            ${{ runner.os }}-buildx-
+      - name: Build
+        uses: docker/build-push-action@v2
+        with:
+          context: ./
+          file: ./Dockerfile
+          builder: ${{ steps.buildx.outputs.name }}
+          push: false
+          cache-from: type=local,src=/tmp/.buildx-cache
+          cache-to: type=local,dest=/tmp/.buildx-cache-new
+      - name: Move cache
+        run: |
+          rm -rf /tmp/.buildx-cache
+          mv /tmp/.buildx-cache-new /tmp/.buildx-cache
+      - name: Image digest
+        run: echo ${{ steps.docker_build.outputs.digest }}

.github/workflows/python.yml

@@ -0,0 +1,26 @@
+name: Python
+
+on: ["push", "pull_request"]
+
+jobs:
+  syntax:
+    runs-on: ubuntu-latest
+    strategy:
+      matrix:
+        python-version: [3.7, 3.8, 3.9]
+    steps:
+    - uses: actions/checkout@v2
+    - name: Set up Python ${{ matrix.python-version }}
+      uses: actions/setup-python@v2
+      with:
+        python-version: ${{ matrix.python-version }}
+    - name: Install dependencies
+      run: |
+        python -m pip install --upgrade pip
+        python -m pip install flake8
+    - name: Lint with flake8
+      run: |
+        # stop the build if there are Python syntax errors or undefined names
+        flake8 . --count --select=E9,F63,F7,F82 --show-source --statistics
+        # exit-zero treats all errors as warnings. The GitHub editor is 127 chars wide
+        flake8 . --count --exit-zero --max-complexity=10 --max-line-length=127 --statistics

.gitignore

@@ -5,3 +5,4 @@ __pycache__
 error_*
 *.log
 /logs/
+.vscode

Dockerfile

@@ -1,4 +1,4 @@
-FROM python
+FROM python:3.8.10
 
 # Create app directory
 WORKDIR /usr/src/app
@@ -9,7 +9,7 @@ COPY requirements.txt ./
 
 RUN pip install -r requirements.txt
 
-RUN touch logs/guilds.log && ln -s logs/guilds.log guilds.log
+RUN mkdir -p logs && touch logs/guilds.log && ln -s logs/guilds.log guilds.log
 
 # Bundle app source
 COPY . .

README.md

@@ -18,14 +18,29 @@
 * %freq - frequency analysis
 * %compo - composition analysis
 * %pres - presence analysis
-* %first - read first message
-* %rand - read a random message
-* %last - read last message
-* %emojis - rank emotes by their usage
+* %repeat - repeat last analysis (adding supplied arguments)
+* %mobile - fix @invalid-user for last command but mentions users
+* %gdpr - displays GDPR information
+* %find - find specific words or phrases (you can use quotes to add spaces in queries, backticks define regexes)
+  * arguments:
+    * top - rank users for these queries
+* %first - read first message (add text to filter like %find)
+  * arguments:
+    * image - pull an image instead of a message
+    * spoiler:allow/only - allow spoiler images
+* %rand - read a random message (add text to filter like %find)
+  * arguments:
+    * image - pull an image instead of a message
+    * spoiler:allow/only - allow spoiler images
+* %last - read last message (add text to filter like %find)
+  * arguments:
+    * image - pull an image instead of a message
+    * spoiler:allow/only - allow spoiler images
+* %emojis - rank emojis by their usage
   * arguments:
     * <n> - top <n> emojis, default is 20
     * all - list all common emojis in addition to this guild's
-    * members - show top member for each emote
+    * members - show top member for each emoji
     * sort:usage/reaction - other sorting methods
 * %mentions - rank mentions by their usage
   * arguments:
@@ -43,14 +58,24 @@
 * %react - rank users by their reactions
   * arguments:
     * <n> - top <n> messages, default is 10
-* %cancel - cancel current analysis
+* %words - (BETA) rank words by their usage
+  * arguments:
+    * <n> - words containings <n> or more letters, default is 3
+    * <n2> - top <n2> words, default is 10
+* %cancel - cancel current analysis (not launched with fast)
 
 * Common arguments:
     * @member/me: filter for one or more member
     * #channel/here: filter for one or more channel
+    * <date1> - filter after <date1>
+    * <date2> - filter before <date2>
     * all/everyone - include bots messages
     * fast: only read cache
     * fresh: does not read cache
+    * nsfw:allow/only - allow messages from nsfw channels
+    * mobile/mention: mentions users (fix @invalid-user bug)
+
+(Sample dates: 2020 / 2021-11 / 2021-06-28 / 2020-06-28T23:00 / today / week / 8days / 1y)
 ```
 
 ## Running this bot
@@ -100,8 +125,36 @@ python3 src/main.py
 
 ## Changelog
 
+* **v1.16**
+  * `%freq graph` graph hours frequency along the week
+  * uses discord new time format
+  * `%freq` now shows quietest day of week and hour of day
+  * improvments and bug fix
+* **v1.15**
+  * `nsfw:allow/only` filter nsfw channels
+  * `%find` can use regexes
+  * `%first`, `%rand` and `%last` can be filter with specific keywords
+  * `%first`, `%rand` and `%last` can pull images
+  * bug fix
+* **v1.14**
+  * `mobile/mention` arg to fix mobile bug
+  * `%repeat`, `%mobile` to repeat commands
+  * more scan: `%find`
+  * bug fix
+* **v1.13**
+  * improved scan `%words`
+  * remove old and unused logs at start and guild leaving
+  * GDPR disclaimer before scanning
+  * start and stop dates
+  * bug fix and improvements
+* **v1.12**
+  * more scans: `%words`
+  * concurrent `fast` analysis
+  * assume `fast` if last analysis is fresh
+  * better memory handling
+  * bug fix
 * **v1.11**
-  * more scans `%first`, `%rand`, `%last`
+  * more scans: `%first`, `%rand`, `%last`
   * streak computing in `%pres`
 * **v1.10**
   * multithreading for queries
@@ -122,7 +175,7 @@ python3 src/main.py
   * more scans: `%scan`, `%freq`, `%compo`, `%pres`
   * huge bug fix
 * **v1.5**:
-  * top <n> emotes
+  * top <n> emojis
   * bug fix
 * **v1.4**:
   * integrate miniscord

requirements.txt

@@ -1,3 +1,6 @@
-discord.py
-python-dotenv
+discord.py==1.7.0
+python-dotenv==0.15.0
+python-dateutil==2.8.1
 git+git://github.com/Klemek/miniscord.git
+numpy
+matplotlib

src/data_types/__init__.py

@@ -1,6 +1,6 @@
-from .emote import Emote, get_emote_dict
-from .frequency import Frequency
+from .emoji import Emoji, get_emoji_dict
 from .composition import Composition
-from .presence import Presence
 from .counter import Counter
+from .frequency import Frequency
 from .history import History
+from .presence import Presence

src/data_types/composition.py

@@ -8,9 +8,9 @@ class Composition:
     def __init__(self):
         self.total_characters = 0
         self.plain_text = 0
-        self.emote_msg = 0
-        self.emote_only = 0
-        self.emotes = defaultdict(int)
+        self.emoji_msg = 0
+        self.emoji_only = 0
+        self.emojis = defaultdict(int)
         self.edited = 0
         self.everyone = 0
         self.answers = 0
@@ -23,49 +23,45 @@ class Composition:
         self.spoilers = 0
 
     def to_string(self, msg_count: int) -> List[str]:
-        ret = []
-        ret += [
-            f"- **avg. characters / message**: {self.total_characters/msg_count:.2f}"
+        total_emojis = val_sum(self.emojis)
+        top_emoji = top_key(self.emojis)
+        ret = [
+            f"- **avg. characters / message**: {self.total_characters/msg_count:.2f}",
+            f"- **plain text messages**: {self.plain_text:,} ({percent(self.plain_text/msg_count)})"
+            if self.plain_text > 0
+            else "",
+            f"- **edited messages**: {self.edited:,} ({percent(self.edited/msg_count)})"
+            if self.edited > 0
+            else "",
+            f"- **@\u200beveryone**: {self.everyone:,} ({percent(self.everyone/msg_count)})"
+            if self.everyone > 0
+            else "",
+            f"- **mentions**: {self.mentions:,} (in {percent(self.mention_msg/msg_count)} of msg, avg. {precise(self.mentions/msg_count)}/msg)"
+            if self.mentions > 0
+            else "",
+            f"- **answers**: {self.answers:,} ({percent(self.answers/msg_count)})"
+            if self.answers > 0
+            else "",
+            f"- **emojis**: {total_emojis:,} (in {percent(self.emoji_msg/msg_count)} of msg, avg. {precise(total_emojis/msg_count)}/msg)"
+            if total_emojis > 0
+            else "",
+            f"- **most used emoji**: {top_emoji} ({plural(self.emojis[top_emoji], 'time')}, {percent(self.emojis[top_emoji]/total_emojis)})"
+            if total_emojis > 0
+            else "",
+            f"- **emoji-only messages**: {self.emoji_only:,} ({percent(self.emoji_only/msg_count)})"
+            if self.emoji_only > 0
+            else "",
+            f"- **images**: {self.images:,} ({percent(self.images/msg_count)})"
+            if self.images > 0
+            else "",
+            f"- **links**: {self.links:,} ({percent(self.link_msg/msg_count)})"
+            if self.links > 0
+            else "",
+            f"- **spoilers**: {self.spoilers:,} ({percent(self.spoilers/msg_count)})"
+            if self.spoilers > 0
+            else "",
+            f"- **tts messages**: {self.tts:,} ({percent(self.tts/msg_count)})"
+            if self.tts > 0
+            else "",
         ]
-        if self.plain_text > 0:
-            ret += [
-                f"- **plain text messages**: {self.plain_text:,} ({percent(self.plain_text/msg_count)})"
-            ]
-        if self.edited > 0:
-            ret += [
-                f"- **edited messages**: {self.edited:,} ({percent(self.edited/msg_count)})"
-            ]
-        if self.everyone > 0:
-            ret += [
-                f"- **@\u200beveryone**: {self.everyone:,} ({percent(self.everyone/msg_count)})"
-            ]
-        if self.mentions > 0:
-            ret += [
-                f"- **mentions**: {self.mentions:,} (in {percent(self.mention_msg/msg_count)} of msg, avg. {precise(self.mentions/msg_count)}/msg)",
-            ]
-        if self.answers > 0:
-            ret += [
-                f"- **answers**: {self.answers:,} ({percent(self.answers/msg_count)})"
-            ]
-        total_emotes = val_sum(self.emotes)
-        if total_emotes > 0:
-            top_emote = top_key(self.emotes)
-            ret += [
-                f"- **emojis**: {total_emotes:,} (in {percent(self.emote_msg/msg_count)} of msg, avg. {precise(total_emotes/msg_count)}/msg)",
-                f"- **most used emoji**: {top_emote} ({plural(self.emotes[top_emote], 'time')}, {percent(self.emotes[top_emote]/total_emotes)})",
-            ]
-            if self.emote_only > 0:
-                ret += [
-                    f"- **emoji-only messages**: {self.emote_only:,} ({percent(self.emote_only/msg_count)})"
-                ]
-        if self.images > 0:
-            ret += [f"- **images**: {self.images:,} ({percent(self.images/msg_count)})"]
-        if self.links > 0:
-            ret += [f"- **links**: {self.links:,} ({percent(self.link_msg/msg_count)})"]
-        if self.spoilers > 0:
-            ret += [
-                f"- **spoilers**: {self.spoilers:,} ({percent(self.spoilers/msg_count)})"
-            ]
-        if self.tts > 0:
-            ret += [f"- **tts messages**: {self.tts:,} ({percent(self.tts/msg_count)})"]
         return ret

src/data_types/counter.py

@@ -14,14 +14,16 @@ class Counter:
 
     def update_use(self, count: int, date: datetime, item: int = 0):
         self.usages[item] += count
-        if self.last_used is None or date > self.last_used:
+        if count > 0 and (self.last_used is None or date > self.last_used):
             self.last_used = date
 
     def score(self) -> float:
         # Score is compose of usages + reactions
-        # When 2 emotes have the same score,
+        # When 2 emojis have the same score,
         # the days since last use is stored in the digits
         # (more recent first)
+        if self.last_used is None:
+            return 0
         return self.all_usages() + 1 / (
             100000 * ((datetime.today() - self.last_used).days + 1)
         )
@@ -37,21 +39,29 @@ class Counter:
         total_usage: int,
         counted: str = "time",
         transform: Optional[Callable[[int], str]] = None,
+        ranking: bool = True,
+        top: bool = True,
     ) -> str:
         # place
         output = ""
-        if i == 0:
-            output += ":first_place:"
-        elif i == 1:
-            output += ":second_place:"
-        elif i == 2:
-            output += ":third_place:"
+        if ranking:
+            if i == 0:
+                output += ":first_place: "
+            elif i == 1:
+                output += ":second_place: "
+            elif i == 2:
+                output += ":third_place: "
+            else:
+                output += f"**#{i + 1}** "
         else:
-            output += f"**#{i + 1}**"
+            output += f"- "
         sum = val_sum(self.usages)
-        output += f" {name} - {plural(sum, counted)} ({percent(sum/total_usage)}, last {from_now(self.last_used)})"
+        if sum > 0:
+            output += f"{name} - {plural(sum, counted)} ({percent(sum/total_usage)}, last {from_now(self.last_used)})"
+        else:
+            output += f"{name} - unused"
         top_item = top_key(self.usages)
-        if top_item != 0 and transform is not None:
+        if sum > 0 and top and top_item != 0 and transform is not None:
             if self.usages[top_item] == sum:
                 output += f" (all{transform(top_item)})"
             else:

src/data_types/emoji.py

@@ -8,9 +8,9 @@ import discord
 from utils import mention, plural, from_now, top_key, percent
 
 
-class Emote:
+class Emoji:
     """
-    Custom class to store emotes data
+    Custom class to store emojis data
     """
 
     def __init__(self, emoji: Optional[discord.Emoji] = None):
@@ -34,7 +34,7 @@ class Emote:
 
     def score(self, *, usage_weight: int = 1, react_weight: int = 1) -> float:
         # Score is compose of usages + reactions
-        # When 2 emotes have the same score,
+        # When 2 emojis have the same score,
         # the days since last use is stored in the digits
         # (more recent first)
         return (
@@ -99,8 +99,8 @@ class Emote:
         return output
 
 
-def get_emote_dict(guild: discord.Guild) -> Dict[str, Emote]:
-    emotes = defaultdict(Emote)
+def get_emoji_dict(guild: discord.Guild) -> Dict[str, Emoji]:
+    emojis = defaultdict(Emoji)
     for emoji in guild.emojis:
-        emotes[str(emoji)] = Emote(emoji)
-    return emotes
+        emojis[str(emoji)] = Emoji(emoji)
+    return emojis

src/data_types/frequency.py

@@ -1,10 +1,13 @@
 from typing import List
 from datetime import timedelta
 import calendar
+import matplotlib.pyplot as plt
+import numpy as np
+from io import BytesIO
+import discord
+import time
 
 from utils import (
-    str_date,
-    str_datetime,
     from_now,
     plural,
     percent,
@@ -13,14 +16,25 @@ from utils import (
     mention,
 )
 
+CB_color_cycle = [
+    "#e41a1c",
+    "#984ea3",
+    "#377eb8",
+    "#4daf4a",
+    "#dede00",
+    "#ff7f00",
+    "#a65628",
+    "#f781bf",
+    "#999999",
+]
+
 
 class Frequency:
     def __init__(self):
         self.dates = []
         self.longest_break = timedelta(seconds=0)
         self.longest_break_start = None
-        self.week = {i: 0 for i in range(7)}
-        self.day = {i: 0 for i in range(24)}
+        self.hours = {i: {j: 0 for j in range(24)} for i in range(7)}
         self.busiest_day = None
         self.busiest_day_count = 0
         self.busiest_hour = None
@@ -33,42 +47,109 @@ class Frequency:
         self.longest_streak_start = None
         self.longest_streak_author = None
 
+    def to_graph(self) -> List[str]:
+        self.dates.sort()
+        delta = self.dates[-1] - self.dates[0]
+        if delta.days == 0:
+            delta = timedelta(days=1)
+        day = {j: sum(self.hours[i][j] for i in range(7)) for j in range(24)}
+        busiest_hour = top_key(day)
+        n_hours = delta.days
+        if self.dates[0].hour <= busiest_hour and self.dates[-1].hour >= busiest_hour:
+            n_hours += 1
+
+        plt.style.use("dark_background")
+
+        fig, ax = plt.subplots()
+
+        times = range(25)
+        ax.set_xticks(times)
+        ax.set_xticklabels([f"{t:0>2}h" if t % 2 == 0 else "" for t in times])
+
+        for i in range(7):
+            hours = [self.hours[i][hour] * 7 / n_hours for hour in range(24)] + [
+                self.hours[i][0] * 7 / n_hours
+            ]
+            ax.plot(
+                times,
+                hours,
+                label=calendar.day_name[i],
+                linestyle="--",
+                linewidth=0.8,
+                c=CB_color_cycle[i],
+            )
+
+        hours = [day[hour] / n_hours for hour in range(24)] + [day[0] / n_hours]
+        ax.plot(times, hours, c="r", label="average", linewidth=1.5)
+
+        fig.patch.set_facecolor("#36393F")
+        ax.patch.set_alpha(0)
+        ax.set_xlim([0, 24])
+        ax.set_ylim([0, None])
+        ax.set_ylabel("average messages")
+        ax.legend(framealpha=0)
+        ax.grid(True, alpha=0.1)
+
+        with BytesIO() as f:
+            plt.savefig(
+                f,
+                format="png",
+                facecolor=fig.get_facecolor(),
+                edgecolor="none",
+                bbox_inches="tight",
+                dpi=300,
+            )
+            f.seek(0)
+            return [discord.File(f, f"{time.time()}-plot.png")]
+
     def to_string(
         self,
         *,
         member_specific: bool,
     ) -> List[str]:
+        self.dates.sort()
         delta = self.dates[-1] - self.dates[0]
+        if delta.days == 0:
+            delta = timedelta(days=1)
         total_msg = len(self.dates)
-        busiest_weekday = top_key(self.week)
-        busiest_hour = top_key(self.day)
+
+        week = {i: sum(self.hours[i].values()) for i in range(7)}
+        day = {j: sum(self.hours[i][j] for i in range(7)) for j in range(24)}
+
+        busiest_weekday = top_key(week)
+        busiest_hour = top_key(day)
+        quietest_weekday = top_key(week, reverse=True)
+        quietest_hour = top_key(day, reverse=True)
         n_weekdays = delta.days // 7
         if (
             self.dates[0].weekday() <= busiest_weekday
             and self.dates[-1].weekday() >= busiest_weekday
-        ):
+        ) or n_weekdays == 0:
             n_weekdays += 1
         n_hours = delta.days
         if self.dates[0].hour <= busiest_hour and self.dates[-1].hour >= busiest_hour:
             n_hours += 1
         ret = [
-            f"- **earliest message**: {str_datetime(self.dates[0])} ({from_now(self.dates[0])})",
-            f"- **latest message**: {str_datetime(self.dates[-1])} ({from_now(self.dates[-1])})",
+            f"- **earliest message**: {from_now(self.dates[0])}",
+            f"- **latest message**: {from_now(self.dates[-1])}",
             f"- **messages/day**: {precise(total_msg/delta.days, precision=3)}",
-            f"- **busiest day of week**: {calendar.day_name[busiest_weekday]} (~{precise(self.week[busiest_weekday]/n_weekdays, precision=3)} msg, {percent(self.week[busiest_weekday]/total_msg)})",
-            f"- **busiest day ever**: {str_date(self.busiest_day)} ({from_now(self.busiest_day)}, {self.busiest_day_count} msg)",
+            f"- **busiest day of week**: {calendar.day_name[busiest_weekday]} (~{precise(week[busiest_weekday]/n_weekdays, precision=3)} msg, {percent(week[busiest_weekday]/total_msg)})",
+            f"- **quietest day of week**: {calendar.day_name[quietest_weekday]} (~{precise(week[quietest_weekday]/n_weekdays, precision=3)} msg, {percent(week[quietest_weekday]/total_msg)})"
+            if week[quietest_weekday] > 0
+            else "",
+            f"- **busiest day ever**: {from_now(self.busiest_day)} ({self.busiest_day_count} msg)"
+            if self.busiest_day is not None
+            else "",
             f"- **messages/hour**: {precise(total_msg*3600/delta.total_seconds(), precision=3)}",
-            f"- **busiest hour of day**: {busiest_hour:0>2}:00 (~{precise(self.day[busiest_hour]/n_hours, precision=3)} msg, {percent(self.day[busiest_hour]/total_msg)})",
-            f"- **busiest hour ever**: {str_datetime(self.busiest_hour)} ({from_now(self.busiest_hour)}, {self.busiest_hour_count} msg)",
-            f"- **longest break**: {plural(round(self.longest_break.total_seconds()/3600), 'hour')} ({plural(self.longest_break.days,'day')}) from {str_datetime(self.longest_break_start)} ({from_now(self.longest_break_start)})",
+            f"- **busiest hour of day**: {busiest_hour:0>2}:00 (~{precise(day[busiest_hour]/n_hours, precision=3)} msg, {percent(day[busiest_hour]/total_msg)})",
+            f"- **quietest hour of day**: {quietest_hour:0>2}:00 (~{precise(day[quietest_hour]/n_hours, precision=3)} msg, {percent(day[quietest_hour]/total_msg)})"
+            if day[quietest_hour] > 0
+            else "",
+            f"- **busiest hour ever**: {from_now(self.busiest_hour)} ({self.busiest_hour_count} msg)",
+            f"- **longest break**: {plural(round(self.longest_break.total_seconds()/3600), 'hour')} ({plural(self.longest_break.days,'day')}), started {from_now(self.longest_break_start)}",
             f"- **avg. streak**: {precise(sum(self.streaks)/len(self.streaks), precision=3)} msg",
+            f"- **longest streak**: {self.longest_streak:,} msg, started {from_now(self.longest_streak_start)}"
+            if member_specific
+            else f"- **longest streak**: {mention(self.longest_streak_author)} ({self.longest_streak:,} msg, started {from_now(self.longest_streak_start)})",
         ]
-        if member_specific:
-            ret += [
-                f"- **longest streak**: {self.longest_streak:,} msg from {str_datetime(self.longest_streak_start)} ({from_now(self.longest_streak_start)})"
-            ]
-        else:
-            ret += [
-                f"- **longest streak**: {mention(self.longest_streak_author)} ({self.longest_streak:,} msg from {str_datetime(self.longest_streak_start)}, {from_now(self.longest_streak_start)})"
-            ]
         return ret

src/data_types/history.py

@@ -3,13 +3,86 @@ import random
 
 # Custom libs
 
-from utils import mention, from_now, str_datetime, message_link
+from utils import (
+    mention,
+    from_now,
+    message_link,
+    SPLIT_TOKEN,
+    FilterLevel,
+    should_allow_spoiler,
+    is_image_gif,
+)
+
+MAX_RANDOM_TRIES = 100
 
 
 class History:
     def __init__(self):
         self.messages = []
 
+    async def to_string_image(
+        self, *, type: str, spoiler: FilterLevel, gif_only: bool
+    ) -> List[str]:
+        if len(self.messages) == 0:
+            return ["There was no messages matching your filters"]
+        message = None
+        intro = None
+        real_message = None
+        if type == "first":
+            self.messages.sort(key=lambda m: m.created_at)
+            index = 0
+            while real_message is None and index < len(self.messages):
+                message = self.messages[index]
+                real_message = await message.fetch()
+                if real_message is not None and (
+                    not should_allow_spoiler(real_message, spoiler)
+                    or (gif_only and not is_image_gif(real_message))
+                ):
+                    real_message = None
+                index += 1
+            intro = f"First image out of {len(self.messages):,}"
+        elif type == "last":
+            self.messages.sort(key=lambda m: m.created_at, reverse=True)
+            index = 0
+            while real_message is None and index < len(self.messages):
+                message = self.messages[index]
+                real_message = await message.fetch()
+                if real_message is not None and (
+                    not should_allow_spoiler(real_message, spoiler)
+                    or (gif_only and not is_image_gif(real_message))
+                ):
+                    real_message = None
+                index += 1
+            intro = f"Last image out of {len(self.messages):,}"
+        elif type == "random":
+            intro = f"Random image out of {len(self.messages):,}"
+            tries = 0
+            while real_message is None and tries < MAX_RANDOM_TRIES:
+                message = random.choice(self.messages)
+                real_message = await message.fetch()
+                if real_message is not None and (
+                    not should_allow_spoiler(real_message, spoiler)
+                    or (gif_only and not is_image_gif(real_message))
+                ):
+                    real_message = None
+                tries += 1
+
+        if real_message is None:
+            return ["There was no messages matching your filters"]
+        image = "<Error>"
+        if len(real_message.attachments) > 0:
+            image = real_message.attachments[0].url
+        elif len(real_message.embeds) > 0:
+            image = real_message.embeds[0].url
+
+        return [
+            intro,
+            f"{from_now(message.created_at)}, {mention(message.author)} sent:",
+            f"<{message_link(message)}>",
+            SPLIT_TOKEN,
+            image,
+        ]
+
     def to_string(self, *, type: str) -> List[str]:
         if len(self.messages) == 0:
             return ["There was no messages matching your filters"]
@@ -33,7 +106,7 @@ class History:
 
         return [
             intro,
-            f"{str_datetime(message.created_at)} ({from_now(message.created_at)}) {mention(message.author)} said:",
+            f"{from_now(message.created_at)}, {mention(message.author)} said:",
             *text,
             f"<{message_link(message)}>",
         ]

src/data_types/presence.py

@@ -25,74 +25,70 @@ class Presence:
         show_top_channel: bool,
         member_specific: bool,
     ) -> List[str]:
-        ret = []
         if chan_count is None:
             type = "server's"
         elif chan_count == 1:
             type = "channel's"
         else:
             type = "channels'"
-        if member_specific:
-            ret += [
-                f"- **messages**: {msg_count:,} ({percent(msg_count/total_msg)} of {type})"
-            ]
-        else:
-            top_member = top_key(self.messages)
-            ret += [
-                f"- **top messages**:  {mention(top_member)} ({self.messages[top_member]:,} msg, {percent(self.messages[top_member]/val_sum(self.messages))})"
-            ]
-        if show_top_channel:
-            top_channel = top_key(self.channel_usage)
-            channel_sum = val_sum(self.channel_usage)
-            found_in = sorted(
-                self.channel_usage,
-                key=lambda k: self.channel_usage[k] / self.channel_total[k],
-            )[-1]
-            ret += [
-                f"- **most visited channel**: {channel_mention(top_channel)} ({self.channel_usage[top_channel]:,} msg, {percent(self.channel_usage[top_channel]/channel_sum)})",
-            ]
-            if member_specific:
-                ret += [
-                    f"- **most contributed channel**: {channel_mention(found_in)} ({self.channel_usage[found_in]:,} msg, {percent(self.channel_usage[found_in]/self.channel_total[found_in])} of {type})"
-                ]
-        if member_specific:
-            if len(self.mentions) > 0:
-                top_mention = top_key(self.mentions)
-                mention_sum = val_sum(self.mentions)
-                ret += [
-                    f"- **was mentioned**: {plural(mention_sum, 'time')} ({percent(mention_sum/val_sum(self.mention_count))} of {type})",
-                    f"- **mostly mentioned by**: {mention(top_mention)} ({plural(self.mentions[top_mention], 'time')}, {percent(self.mentions[top_mention]/mention_sum)})",
-                ]
-        if len(self.mention_others) > 0:
-            top_mention = top_key(self.mention_others)
-            mention_sum = val_sum(self.mention_others)
-            if member_specific:
-                ret += [
-                    f"- **mentioned others**: {plural(mention_sum, 'time')} ({percent(mention_sum/val_sum(self.mention_count))} of {type})",
-                    f"- **mostly mentioned**: {mention(top_mention)} ({plural(self.mention_others[top_mention], 'time')}, {percent(self.mention_others[top_mention]/mention_sum)})",
-                ]
-            else:
-                top_member = top_key(self.mention_count)
-                ret += [
-                    f"- **mentioned**: {plural(mention_sum, 'time')} ({mention(top_member)}, {percent(self.mention_count[top_member]/val_sum(self.mention_count))})",
-                    f"- **top mentions**: {mention(top_member)} ({plural(self.mention_count[top_member], 'time')}, {percent(self.mention_count[top_member]/val_sum(self.mention_count))})",
-                    f"- **most mentioned**: {mention(top_mention)} ({plural(self.mention_others[top_mention], 'time')}, {percent(self.mention_others[top_mention]/mention_sum)})",
-                ]
-        if len(self.reactions) > 0:
-            total_used = val_sum(self.reactions)
-            top_reaction = top_key(self.reactions)
-            ret += [
-                f"- **reactions**: {plural(total_used, 'time')}",
-                f"- **most used reaction**: {top_reaction} ({plural(self.reactions[top_reaction], 'time')}, {percent(self.reactions[top_reaction]/total_used)})",
-            ]
-            if member_specific:
-                ret[
-                    -2
-                ] += f" ({percent(total_used/val_sum(self.used_reaction))} of {type})"
-            else:
-                top_member = top_key(self.used_reaction)
-                ret.insert(
-                    -1,
-                    f"- **top reactions**: {mention(top_member)} ({plural(self.used_reaction[top_member], 'time')}, {percent(self.used_reaction[top_member]/val_sum(self.used_reaction))})",
-                )
+        top_member = top_key(self.messages)
+        top_channel = top_key(self.channel_usage)
+        channel_sum = val_sum(self.channel_usage)
+        found_in = top_key(
+            self.channel_usage,
+            key=lambda k: self.channel_usage[k] / self.channel_total[k],
+        )
+        top_mention = top_key(self.mentions)
+        mention_sum = val_sum(self.mentions)
+        top_mention_others = top_key(self.mention_others)
+        mention_others_sum = val_sum(self.mention_others)
+        top_member_mentioned = top_key(self.mention_count)
+        total_reaction_used = val_sum(self.reactions)
+        top_reaction = top_key(self.reactions)
+        top_reaction_member = top_key(self.used_reaction)
+
+        ret = [
+            f"- **messages**: {msg_count:,} ({percent(msg_count/total_msg)} of {type})"
+            if member_specific
+            else f"- **top messages**:  {mention(top_member)} ({self.messages[top_member]:,} msg, {percent(self.messages[top_member]/val_sum(self.messages))})",
+            f"- **most visited channel**: {channel_mention(top_channel)} ({self.channel_usage[top_channel]:,} msg, {percent(self.channel_usage[top_channel]/channel_sum)})"
+            if show_top_channel
+            else "",
+            f"- **most contributed channel**: {channel_mention(found_in)} ({self.channel_usage[found_in]:,} msg, {percent(self.channel_usage[found_in]/self.channel_total[found_in])} of {type})"
+            if show_top_channel and member_specific
+            else "",
+            f"- **was mentioned**: {plural(mention_sum, 'time')} ({percent(mention_sum/val_sum(self.mention_count))} of {type})"
+            if member_specific and len(self.mentions) > 0
+            else "",
+            f"- **mostly mentioned by**: {mention(top_mention)} ({plural(self.mentions[top_mention], 'time')}, {percent(self.mentions[top_mention]/mention_sum)})"
+            if member_specific and len(self.mentions) > 0
+            else "",
+            f"- **mentioned others**: {plural(mention_others_sum, 'time')} ({percent(mention_others_sum/val_sum(self.mention_count))} of {type})"
+            if len(self.mention_others) > 0 and member_specific
+            else "",
+            f"- **mostly mentioned**: {mention(top_mention_others)} ({plural(self.mention_others[top_mention_others], 'time')}, {percent(self.mention_others[top_mention_others]/mention_others_sum)})"
+            if len(self.mention_others) > 0 and member_specific
+            else "",
+            f"- **mentioned**: {plural(mention_others_sum, 'time')} ({mention(top_member_mentioned)}, {percent(self.mention_count[top_member_mentioned]/val_sum(self.mention_count))})"
+            if len(self.mention_others) > 0 and not member_specific
+            else "",
+            f"- **top mentions**: {mention(top_member_mentioned)} ({plural(self.mention_count[top_member_mentioned], 'time')}, {percent(self.mention_count[top_member_mentioned]/val_sum(self.mention_count))})"
+            if len(self.mention_others) > 0 and not member_specific
+            else "",
+            f"- **most mentioned**: {mention(top_mention_others)} ({plural(self.mention_others[top_mention_others], 'time')}, {percent(self.mention_others[top_mention_others]/mention_others_sum)})"
+            if len(self.mention_others) > 0 and not member_specific
+            else "",
+            f"- **reactions**: {plural(total_reaction_used, 'time')}"
+            if len(self.reactions) > 0 and not member_specific
+            else "",
+            f"- **reactions**: {plural(total_reaction_used, 'time')} ({percent(total_reaction_used/val_sum(self.used_reaction))} of {type})"
+            if len(self.reactions) > 0 and member_specific
+            else "",
+            f"- **top reactions**: {mention(top_reaction_member)} ({plural(self.used_reaction[top_reaction_member], 'time')}, {percent(self.used_reaction[top_reaction_member]/val_sum(self.used_reaction))})"
+            if len(self.reactions) > 0 and not member_specific
+            else "",
+            f"- **most used reaction**: {top_reaction} ({plural(self.reactions[top_reaction], 'time')}, {percent(self.reactions[top_reaction]/total_reaction_used)})"
+            if len(self.reactions) > 0
+            else "",
+        ]
         return ret

src/logs/__init__.py

@@ -1,3 +1,3 @@
 from .message_log import MessageLog
 from .channel_logs import ChannelLogs
-from .guild_logs import GuildLogs, ALREADY_RUNNING, CANCELLED
+from .guild_logs import GuildLogs, ALREADY_RUNNING, CANCELLED, NO_FILE

src/logs/channel_logs.py

@@ -1,8 +1,10 @@
+import logging
 from typing import Union, Tuple, Any
 import discord
+from datetime import datetime
 
 from . import MessageLog
-from utils import FakeMessage
+from utils import serialize, FakeMessage
 
 CHUNK_SIZE = 2000
 FORMAT = 3
@@ -15,8 +17,10 @@ class ChannelLogs:
             self.id = channel.id
             self.name = channel.name
             self.last_message_id = None
+            self.first_message_id = None
             self.format = FORMAT
-            self.messages = []
+            self.messages = set()
+            self.start_date = None
         elif isinstance(channel, dict):
             self.format = channel["format"] if "format" in channel else None
             if not self.is_format():
@@ -28,55 +32,110 @@ class ChannelLogs:
                 if channel["last_message_id"] is not None
                 else None
             )
-            self.messages = [MessageLog(message, self) for message in channel["messages"]]
+            self.first_message_id = (
+                int(channel["first_message_id"])
+                if "first_message_id" in channel
+                and channel["first_message_id"] is not None
+                else None
+            )
+            self.messages = {
+                MessageLog(message, self) for message in channel["messages"]
+            }
+            self.start_date = (
+                self.sorted_messages[0].created_at if len(self.messages) > 0 else None
+            )
 
     def is_format(self):
         return self.format == FORMAT
 
-    async def load(self, channel: discord.TextChannel) -> Tuple[int, int]:
+    def preload(self, channel: discord.TextChannel):
         self.name = channel.name
         self.channel = channel
+
+    @property
+    def sorted_messages(self):
+        return sorted(self.messages)
+
+    @property
+    def nsfw(self):
+        self.channel.nsfw
+
+    async def load(
+        self, channel: discord.TextChannel, start_date: datetime, stop_date: datetime
+    ) -> Tuple[int, int]:
+        is_empty = self.last_message_id is None
         try:
-            if self.last_message_id is not None:  # append
-                while self.last_message_id != channel.last_message_id:
+            if is_empty:
+                sanity_check = len(await channel.history(limit=1).flatten())
+                if sanity_check != 1:
+                    yield len(self.messages), True
+                    return
+            # load backward
+            if is_empty or (
+                self.first_message_id is not None
+                and (
+                    start_date is None
+                    or (self.start_date is not None and self.start_date > start_date)
+                )
+            ):
+                first_message_date = None
+                tmp_message_id = 0
+                done = 0
+                while (
+                    first_message_date is None
+                    or (
+                        done >= CHUNK_SIZE
+                        and (start_date is None or first_message_date > start_date)
+                    )
+                ) and tmp_message_id != self.first_message_id:
+                    tmp_message_id = self.first_message_id
+                    done = 0
                     async for message in channel.history(
                         limit=CHUNK_SIZE,
-                        after=FakeMessage(self.last_message_id),
+                        before=FakeMessage(self.first_message_id)
+                        if self.first_message_id is not None
+                        else None,
+                        oldest_first=False,
+                    ):
+                        done += 1
+                        self.first_message_id = message.id
+                        first_message_date = message.created_at
+                        m = MessageLog(message, self)
+                        await m.load(message)
+                        self.messages.add(m)
+                    yield len(self.messages), False
+                if done < CHUNK_SIZE:  # reached bottom
+                    self.first_message_id = None
+                self.last_message_id = channel.last_message_id
+            # load forward
+            last_message_date = self.sorted_messages[-1].created_at
+            if not is_empty and (stop_date is None or last_message_date < stop_date):
+                tmp_message_id = None
+                while (
+                    self.last_message_id != channel.last_message_id
+                    and (stop_date is None or last_message_date < stop_date)
+                ) and self.last_message_id != tmp_message_id:
+                    tmp_message_id = self.last_message_id
+                    async for message in channel.history(
+                        limit=CHUNK_SIZE,
+                        after=FakeMessage(self.first_message_id),
                         oldest_first=True,
                     ):
+                        last_message_date = message.created_at
                         self.last_message_id = message.id
                         m = MessageLog(message, self)
                         await m.load(message)
-                        self.messages.insert(0, m)
+                        self.messages.add(m)
                     yield len(self.messages), False
-            else:  # first load
-                last_message_id = None
-                done = 0
-                sanity_check = len(await channel.history(limit=1).flatten())
-                if sanity_check == 1:
-                    while done >= CHUNK_SIZE or last_message_id is None:
-                        done = 0
-                        async for message in channel.history(
-                            limit=CHUNK_SIZE,
-                            before=FakeMessage(last_message_id)
-                            if last_message_id is not None
-                            else None,
-                            oldest_first=False,
-                        ):
-                            done += 1
-                            last_message_id = message.id
-                            m = MessageLog(message, self)
-                            await m.load(message)
-                            self.messages += [m]
-                        yield len(self.messages), False
-                    self.last_message_id = channel.last_message_id
-        except discord.errors.HTTPException:
+        except discord.errors.HTTPException as e:
             yield -1, True
             return  # When an exception occurs (like Forbidden)
+        self.start_date = (
+            self.sorted_messages[0].created_at if len(self.messages) > 0 else None
+        )
         yield len(self.messages), True
 
     def dict(self) -> dict:
-        channel = dict(self.__dict__)
-        channel.pop("channel", None)
+        channel = serialize(self, not_serialized=["channel", "guild", "start_date"])
         channel["messages"] = [message.dict() for message in self.messages]
         return channel

src/logs/guild_logs.py

@@ -4,6 +4,7 @@ import discord
 import json
 import gzip
 from datetime import datetime
+import time
 import logging
 import asyncio
 import threading
@@ -14,6 +15,7 @@ from utils import code_message, delta, deltas
 
 
 LOG_DIR = "logs"
+LOG_EXT = ".logz"
 
 current_analysis = []
 current_analysis_lock = threading.Lock()
@@ -21,10 +23,22 @@ current_analysis_lock = threading.Lock()
 
 ALREADY_RUNNING = -100
 CANCELLED = -200
+NO_FILE = -300
+
+# 5 minutes, assume 'fast' arg
+MIN_MODIFICATION_TIME = 5 * 60
+# ~1 year, remove log file
+MAX_MODIFICATION_TIME = 365 * 24 * 60 * 60
 
 
 class Worker:
-    def __init__(self, channel_log: ChannelLogs, channel: discord.TextChannel):
+    def __init__(
+        self,
+        channel_log: ChannelLogs,
+        channel: discord.TextChannel,
+        start_date: datetime,
+        stop_date: datetime,
+    ):
         self.channel_log = channel_log
         self.channel = channel
         self.start_msg = len(channel_log.messages)
@@ -33,12 +47,16 @@ class Worker:
         self.done = False
         self.cancelled = False
         self.loop = asyncio.get_event_loop()
+        self.start_date = start_date
+        self.stop_date = stop_date
 
     def start(self):
         asyncio.run_coroutine_threadsafe(self.process(), self.loop)
 
     async def process(self):
-        async for count, done in self.channel_log.load(self.channel):
+        async for count, done in self.channel_log.load(
+            self.channel, self.start_date, self.stop_date
+        ):
             if count > 0:
                 self.queried_msg = count - self.start_msg
                 self.total_msg = count
@@ -51,102 +69,159 @@ class GuildLogs:
     def __init__(self, guild: discord.Guild):
         self.id = guild.id
         self.guild = guild
-        self.log_file = os.path.join(LOG_DIR, f"{guild.id}.logz")
+        self.log_file = os.path.join(LOG_DIR, f"{guild.id}{LOG_EXT}")
         self.channels = {}
+        self.locked = False
+
+    def __enter__(self):
+        return self
+
+    def __exit__(self, type, value, tb):
+        del self.channels
+        del self.guild
+        if self.locked:
+            self.unlock()
 
     def dict(self) -> dict:
         return {id: self.channels[id].dict() for id in self.channels}
 
     def check_cancelled(self) -> bool:
-        return self.log_file not in current_analysis
+        return self.locked and self.log_file not in current_analysis
+
+    def lock(self) -> bool:
+        current_analysis_lock.acquire()
+        if self.log_file in current_analysis:
+            current_analysis_lock.release()
+            return False
+        self.locked = True
+        current_analysis.append(self.log_file)
+        current_analysis_lock.release()
+        return True
+
+    def unlock(self):
+        if self.locked:
+            self.locked = False
+            current_analysis_lock.acquire()
+            if self.log_file in current_analysis:
+                current_analysis.remove(self.log_file)
+            current_analysis_lock.release()
 
     async def load(
         self,
         progress: discord.Message,
-        target_channels: List[discord.TextChannel] = [],
+        target_channels: List[discord.TextChannel],
+        start_date: datetime,
+        stop_date: datetime,
         *,
         fast: bool,
         fresh: bool,
     ) -> Tuple[int, int]:
-        current_analysis_lock.acquire()
-        if self.log_file in current_analysis:
-            current_analysis_lock.release()
+        self.locked = False
+        if not fast and not self.lock():
             return ALREADY_RUNNING, 0
-        current_analysis.append(self.log_file)
-        current_analysis_lock.release()
         t00 = datetime.now()
         # read logs
         if not os.path.exists(LOG_DIR):
             os.mkdir(LOG_DIR)
-        if os.path.exists(self.log_file):
-            channels = {}
-            try:
-                gziped_data = None
-                await code_message(progress, "Reading saved history (1/4)...")
-                t0 = datetime.now()
-                with open(self.log_file, mode="rb") as f:
-                    gziped_data = f.read()
-                logging.info(f"log {self.guild.id} > read in {delta(t0):,}ms")
-                if self.check_cancelled():
-                    return CANCELLED, 0
-                await code_message(progress, "Reading saved history (2/4)...")
-                t0 = datetime.now()
-                json_data = gzip.decompress(gziped_data)
-                logging.info(
-                    f"log {self.guild.id} > gzip decompress in {delta(t0):,}ms"
-                )
-                if self.check_cancelled():
-                    return CANCELLED, 0
-                await code_message(progress, "Reading saved history (3/4)...")
-                t0 = datetime.now()
-                channels = json.loads(json_data)
-                logging.info(f"log {self.guild.id} > json parse in {delta(t0):,}ms")
-                if self.check_cancelled():
-                    return CANCELLED, 0
-                await code_message(progress, "Reading saved history (4/4)...")
-                t0 = datetime.now()
-                self.channels = {
-                    int(id): ChannelLogs(channels[id], self) for id in channels
-                }
-                # remove invalid format
-                self.channels = {
-                    id: self.channels[id]
-                    for id in self.channels
-                    if self.channels[id].is_format()
-                }
-                logging.info(f"log {self.guild.id} > loaded in {delta(t0):,}ms")
-            except json.decoder.JSONDecodeError:
-                logging.error(f"log {self.guild.id} > invalid JSON")
-            except IOError:
-                logging.error(f"log {self.guild.id} > cannot read")
-        else:
-            fast = False
+        last_time = None
+        if not os.path.exists(self.log_file):
+            return NO_FILE, 0
+        channels = {}
+        try:
+            last_time = os.path.getmtime(self.log_file)
+            gziped_data = None
+            await code_message(progress, "Reading saved history (1/4)...")
+            t0 = datetime.now()
+            with open(self.log_file, mode="rb") as f:
+                gziped_data = f.read()
+            logging.info(f"log {self.guild.id} > read in {delta(t0):,}ms")
+            if self.check_cancelled():
+                return CANCELLED, 0
+            await code_message(progress, "Reading saved history (2/4)...")
+            t0 = datetime.now()
+            json_data = gzip.decompress(gziped_data)
+            del gziped_data
+            logging.info(f"log {self.guild.id} > gzip decompress in {delta(t0):,}ms")
+            if self.check_cancelled():
+                return CANCELLED, 0
+            await code_message(progress, "Reading saved history (3/4)...")
+            t0 = datetime.now()
+            channels = json.loads(json_data)
+            del json_data
+            logging.info(f"log {self.guild.id} > json parse in {delta(t0):,}ms")
+            if self.check_cancelled():
+                return CANCELLED, 0
+            await code_message(progress, "Reading saved history (4/4)...")
+            t0 = datetime.now()
+            self.channels = {
+                int(id): ChannelLogs(channels[id], self) for id in channels
+            }
+            # remove invalid format
+            self.channels = {
+                id: self.channels[id]
+                for id in self.channels
+                if self.channels[id].is_format()
+            }
+            logging.info(f"log {self.guild.id} > loaded in {delta(t0):,}ms")
+        except json.decoder.JSONDecodeError:
+            logging.error(f"log {self.guild.id} > invalid JSON")
+        except IOError:
+            logging.error(f"log {self.guild.id} > cannot read")
+
+        if len(target_channels) == 0:
+            target_channels = (
+                self.channels.values() if fast else self.guild.text_channels
+            )
+        elif fast:
+            # select already loaded channels only
+            target_channels_tmp = [
+                channel for channel in target_channels if channel.id in self.channels
+            ]
+            if len(target_channels_tmp) == 0:
+                fast = False
+            else:
+                target_channels = target_channels_tmp
+
+        # assume fast if file is fresh
+        if (
+            not fast
+            and not fresh
+            and start_date is None
+            and stop_date is None
+            and last_time is not None
+            and (time.time() - last_time) < MIN_MODIFICATION_TIME
+        ):
+            invalid_target_channels = [
+                channel
+                for channel in target_channels
+                if channel.id not in self.channels
+                or self.channels[channel.id].first_message_id is not None
+            ]
+            if len(invalid_target_channels) == 0:
+                logging.info(f"log {self.guild.id} > assumed fast")
+                fast = True
+                if self.locked:
+                    self.unlock()
 
         total_msg = 0
         total_chan = 0
         if fast:
-            if len(target_channels) == 0:
-                total_msg = sum(
-                    [len(channel.messages) for channel in self.channels.values()]
-                )
-                total_chan = len(self.channels)
-            else:
-                target_channels_id = [channel.id for channel in target_channels]
-                total_msg = sum(
-                    [
-                        len(channel.messages)
-                        for channel in self.channels.values()
-                        if channel.id in target_channels_id
-                    ]
-                )
-                total_chan = len(target_channels)
+            target_channels_id = [channel.id for channel in target_channels]
+            total_msg = sum(
+                [
+                    len(channel.messages)
+                    for channel in self.channels.values()
+                    if channel.id in target_channels_id
+                ]
+            )
+            total_chan = len(target_channels)
+            for channel in target_channels:
+                self.channels[channel.id].preload(channel)
         else:
+            if not self.locked and not self.lock():
+                return ALREADY_RUNNING, 0
             # load channels
             t0 = datetime.now()
-            if len(target_channels) == 0:
-                target_channels = (
-                    self.guild.text_channels if not fast else self.channels.keys()
-                )
             loading_new = 0
             queried_msg = 0
             total_chan = 0
@@ -158,7 +233,10 @@ class GuildLogs:
                 if channel.id not in self.channels or fresh:
                     loading_new += 1
                     self.channels[channel.id] = ChannelLogs(channel, self)
-                workers += [Worker(self.channels[channel.id], channel)]
+                self.channels[channel.id].preload(channel)
+                workers += [
+                    Worker(self.channels[channel.id], channel, start_date, stop_date)
+                ]
             warning_msg = "(this might take a while)"
             if len(target_channels) > 5 and loading_new > 5:
                 warning_msg = "(most channels are new, this will take a long while)"
@@ -199,7 +277,7 @@ class GuildLogs:
                     f"Reading new history...\n{total_msg:,} messages in {total_chan:,}/{max_chan:,} channels ({round(queried_msg/deltas(t0)):,}m/s)\n{warning_msg}{remaining_msg}",
                 )
             logging.info(
-                f"log {self.guild.id} > queried in {delta(t0):,}ms -> {queried_msg / deltas(t0):,.3f} m/s"
+                f"log {self.guild.id} > queried {queried_msg} in {delta(t0):,}ms -> {queried_msg / deltas(t0):,.3f} m/s"
             )
             # write logs
             real_total_msg = sum(
@@ -225,6 +303,7 @@ class GuildLogs:
             )
             t0 = datetime.now()
             gziped_data = gzip.compress(json_data)
+            del json_data
             logging.info(
                 f"log {self.guild.id} > gzip in {delta(t0):,}ms -> {real_total_msg / deltas(t0):,.3f} m/s"
             )
@@ -237,6 +316,7 @@ class GuildLogs:
             t0 = datetime.now()
             with open(self.log_file, mode="wb") as f:
                 f.write(gziped_data)
+            del gziped_data
             logging.info(
                 f"log {self.guild.id} > saved in {delta(t0):,}ms -> {real_total_msg / deltas(t0):,.3f} m/s"
             )
@@ -247,9 +327,10 @@ class GuildLogs:
             f"Analysing...\n{total_msg:,} messages in {total_chan:,} channels",
         )
         logging.info(f"log {self.guild.id} > TOTAL TIME: {delta(t00):,}ms")
-        current_analysis_lock.acquire()
-        current_analysis.remove(self.log_file)
-        current_analysis_lock.release()
+        if self.locked:
+            current_analysis_lock.acquire()
+            current_analysis.remove(self.log_file)
+            current_analysis_lock.release()
         return total_msg, total_chan
 
     @staticmethod
@@ -262,5 +343,49 @@ class GuildLogs:
         else:
             current_analysis_lock.release()
             await message.channel.send(
-                f"No analysis are currently running on this server", reference=message
+                f"No cancellable analysis are currently running on this server",
+                reference=message,
             )
+
+    @staticmethod
+    def init_log(guild: List[discord.Guild]):
+        if not os.path.exists(LOG_DIR):
+            os.mkdir(LOG_DIR)
+        filename = os.path.join(LOG_DIR, f"{guild.id}{LOG_EXT}")
+        if not os.path.exists(filename):
+            with open(filename, mode="wb") as f:
+                f.write(gzip.compress(bytes("{}", "utf-8")))
+            logging.info(f"log {guild.id} > created")
+        else:
+            logging.info(f"log {guild.id} > already exists")
+
+    @staticmethod
+    def remove_log(guild: List[discord.Guild]):
+        if not os.path.exists(LOG_DIR):
+            os.mkdir(LOG_DIR)
+        filename = os.path.join(LOG_DIR, f"{guild.id}{LOG_EXT}")
+        if os.path.exists(filename):
+            os.unlink(filename)
+            logging.info(f"log {guild.id} > removed")
+        else:
+            logging.info(f"log {guild.id} > does not exists")
+
+    @staticmethod
+    def check_logs(guilds: List[discord.Guild]):
+        logging.info(f"checking logs...")
+        if not os.path.exists(LOG_DIR):
+            os.mkdir(LOG_DIR)
+        guild_ids = [str(guild.id) for guild in guilds]
+        for item in os.listdir(LOG_DIR):
+            path = os.path.join(LOG_DIR, item)
+            name, ext = os.path.splitext(item)
+            if os.path.isfile(path) and ext == LOG_EXT:
+                if (
+                    name in guild_ids
+                    and (time.time() - os.path.getmtime(path)) > MAX_MODIFICATION_TIME
+                ):
+                    logging.info(f"> removing old log '{path}'")
+                    os.unlink(path)
+                elif name not in guild_ids:
+                    logging.info(f"> removing unused log '{path}'")
+                    os.unlink(path)

src/logs/message_log.py

@@ -1,11 +1,11 @@
-from typing import Union, Any
+from typing import Optional, Union, Any
 import discord
 from datetime import datetime
 
-from utils import is_extension
-
-IMAGE_FORMAT = [".gif", ".gifv", ".png", ".jpg", ".jpeg", ".bmp"]
-EMBED_IMAGES = ["image", "gifv"]
+from utils import (
+    serialize,
+    has_image,
+)
 
 
 class MessageLog:
@@ -36,15 +36,7 @@ class MessageLog:
             self.image = False
             self.attachment = len(message.attachments) > 0
             self.embed = len(message.embeds) > 0
-            for attachment in message.attachments:
-                if is_extension(attachment.filename, IMAGE_FORMAT):
-                    self.image = True
-                    break
-            else:
-                for embed in message.embeds:
-                    if embed.type in EMBED_IMAGES:
-                        self.image = True
-                        break
+            self.image = has_image(message)
             self.reactions = {}
         elif isinstance(message, dict):
             self.id = int(message["id"])
@@ -71,16 +63,28 @@ class MessageLog:
             self.attachment = message["attachment"]
             self.reactions = message["reactions"]
 
+    def __eq__(self, other: object) -> bool:
+        return isinstance(other, self.__class__) and other.id == self.id
+
+    def __gt__(self, other: "MessageLog") -> bool:
+        return self.created_at > other.created_at
+
+    def __hash__(self) -> int:
+        return self.id
+
     async def load(self, message: discord.Message):
         for reaction in message.reactions:
             self.reactions[str(reaction.emoji)] = []
             async for user in reaction.users():
                 self.reactions[str(reaction.emoji)] += [user.id]
 
+    async def fetch(self) -> Optional[discord.Message]:
+        try:
+            return await self.channel.channel.fetch_message(self.id)
+        except (discord.NotFound, discord.Forbidden, discord.HTTPException):
+            return None
+
     def dict(self) -> dict:
-        message = dict(self.__dict__)
-        message["created_at"] = self.created_at.isoformat()
-        message["edited_at"] = (
-            self.edited_at.isoformat() if self.edited_at is not None else None
+        return serialize(
+            self, not_serialized=["channel"], dates=["created_at", "edited_at"]
         )
-        return message

src/main.py

@@ -6,22 +6,8 @@ if sys.version_info < (3, 7):
     print("Please upgrade your Python version to 3.7.0 or higher")
     sys.exit(1)
 
-from utils import emojis
-from scanners import (
-    EmotesScanner,
-    FullScanner,
-    FrequencyScanner,
-    CompositionScanner,
-    PresenceScanner,
-    MentionsScanner,
-    MentionedScanner,
-    MessagesScanner,
-    ChannelsScanner,
-    ReactionsScanner,
-    FirstScanner,
-    RandomScanner,
-    LastScanner,
-)
+from utils import emojis, gdpr, command_cache
+import scanners
 from logs import GuildLogs
 
 logging.basicConfig(
@@ -32,95 +18,139 @@ emojis.load_emojis()
 
 bot = Bot(
     "Discord Analyst",
-    "1.11",
+    "1.16.1",
     alias="%",
 )
 
 bot.log_calls = True
 
+
+async def on_ready():
+    GuildLogs.check_logs(bot.client.guilds)
+    return True
+
+
+async def on_guild_remove():
+    GuildLogs.check_logs(bot.client.guilds)
+    return True
+
+
+bot.register_event(on_ready)
+bot.register_event(on_guild_remove)
+
 bot.register_command(
     "(cancel|stop)",
     GuildLogs.cancel,
-    "cancel: stop current analysis",
-    "```\n" + "%cancel: Stop current analysis\n" + "```",
+    "cancel: stop current analysis (not launched with fast)",
+    "```\n%cancel: Stop current analysis (not launched with fast)\n```",
+)
+bot.register_command(
+    "gdpr",
+    gdpr.process,
+    "gdpr: displays GDPR information",
+    gdpr.HELP,
+)
+bot.register_command(
+    "words",
+    lambda *args: scanners.WordsScanner().compute(*args),
+    "words: (BETA) rank words by their usage",
+    scanners.WordsScanner.help(),
+)
+bot.register_command(
+    "repeat",
+    command_cache.repeat,
+    "repeat: repeat last analysis (adding supplied arguments)",
+    "```\n%repeat: repeat last analysis (adding supplied arguments)\n```",
+)
+bot.register_command(
+    "mobile",
+    lambda *args: command_cache.repeat(*args, add_args=["mobile"]),
+    "mobile: fix @invalid-user for last command but mentions users",
+    "```\n%mobile: fix @invalid-user for last command but mentions users\n```",
+)
+bot.register_command(
+    "find",
+    lambda *args: scanners.FindScanner().compute(*args),
+    "find: find specific words or phrases",
+    scanners.FindScanner.help(),
 )
 bot.register_command(
     "last",
-    lambda *args: LastScanner().compute(*args),
+    lambda *args: scanners.LastScanner().compute(*args),
     "last: read last message",
-    LastScanner.help(),
+    scanners.LastScanner.help(),
 )
 bot.register_command(
-    "rand(om)?",
-    lambda *args: RandomScanner().compute(*args),
+    "(rand(om)?|mood)",
+    lambda *args: scanners.RandomScanner().compute(*args),
     "rand: read a random message",
-    RandomScanner.help(),
+    scanners.RandomScanner.help(),
 )
 bot.register_command(
     "first",
-    lambda *args: FirstScanner().compute(*args),
+    lambda *args: scanners.FirstScanner().compute(*args),
     "first: read first message",
-    FirstScanner.help(),
+    scanners.FirstScanner.help(),
 )
 bot.register_command(
     "mentioned",
-    lambda *args: MentionedScanner().compute(*args),
+    lambda *args: scanners.MentionedScanner().compute(*args),
     "mentioned: rank specific user mentions by their usage",
-    MentionedScanner.help(),
+    scanners.MentionedScanner.help(),
 )
 bot.register_command(
     "(mentions?)",
-    lambda *args: MentionsScanner().compute(*args),
+    lambda *args: scanners.MentionsScanner().compute(*args),
     "mentions: rank mentions by their usage",
-    MentionsScanner.help(),
+    scanners.MentionsScanner.help(),
 )
 bot.register_command(
     "(emojis?|emotes?)",
-    lambda *args: EmotesScanner().compute(*args),
+    lambda *args: scanners.EmojisScanner().compute(*args),
     "emojis: rank emojis by their usage",
-    EmotesScanner.help(),
+    scanners.EmojisScanner.help(),
 )
 bot.register_command(
     "(react(ions?)?)",
-    lambda *args: ReactionsScanner().compute(*args),
+    lambda *args: scanners.ReactionsScanner().compute(*args),
     "react: rank users by their reactions",
-    ReactionsScanner.help(),
+    scanners.ReactionsScanner.help(),
 )
 bot.register_command(
     "(channels?|chan)",
-    lambda *args: ChannelsScanner().compute(*args),
+    lambda *args: scanners.ChannelsScanner().compute(*args),
     "chan: rank channels by their messages",
-    ChannelsScanner.help(),
+    scanners.ChannelsScanner.help(),
 )
 bot.register_command(
     "(messages?|msg)",
-    lambda *args: MessagesScanner().compute(*args),
+    lambda *args: scanners.MessagesScanner().compute(*args),
     "msg: rank users by their messages",
-    MessagesScanner.help(),
+    scanners.MessagesScanner.help(),
 )
 bot.register_command(
     "pres(ence)?",
-    lambda *args: PresenceScanner().compute(*args),
+    lambda *args: scanners.PresenceScanner().compute(*args),
     "pres: presence analysis",
-    PresenceScanner.help(),
+    scanners.PresenceScanner.help(),
 )
 bot.register_command(
     "compo(sition)?",
-    lambda *args: CompositionScanner().compute(*args),
+    lambda *args: scanners.CompositionScanner().compute(*args),
     "compo: composition analysis",
-    CompositionScanner.help(),
+    scanners.CompositionScanner.help(),
 )
 bot.register_command(
     "freq(ency)?",
-    lambda *args: FrequencyScanner().compute(*args),
+    lambda *args: scanners.FrequencyScanner().compute(*args),
     "freq: frequency analysis",
-    FrequencyScanner.help(),
+    scanners.FrequencyScanner.help(),
 )
 bot.register_command(
     "(full|scan)",
-    lambda *args: FullScanner().compute(*args),
+    lambda *args: scanners.FullScanner().compute(*args),
     "scan: full analysis",
-    FullScanner.help(),
+    scanners.FullScanner.help(),
 )
 
 bot.start()

src/scanners/__init__.py

@@ -1,13 +1,17 @@
-from .emotes_scanner import EmotesScanner
-from .frequency_scanner import FrequencyScanner
-from .composition_scanner import CompositionScanner
-from .presence_scanner import PresenceScanner
-from .full_scanner import FullScanner
-from .mentions_scanner import MentionsScanner
-from .mentioned_scanner import MentionedScanner
-from .messages_scanner import MessagesScanner
+from .scanner import Scanner
+
 from .channels_scanner import ChannelsScanner
-from .reactions_scanner import ReactionsScanner
+from .composition_scanner import CompositionScanner
+from .emojis_scanner import EmojisScanner
+from .find_scanner import FindScanner
 from .first_scanner import FirstScanner
+from .frequency_scanner import FrequencyScanner
+from .full_scanner import FullScanner
 from .last_scanner import LastScanner
-from .random_scanner import RandomScanner
+from .mentioned_scanner import MentionedScanner
+from .mentions_scanner import MentionsScanner
+from .messages_scanner import MessagesScanner
+from .presence_scanner import PresenceScanner
+from .random_scanner import RandomScanner
+from .reactions_scanner import ReactionsScanner
+from .words_scanner import WordsScanner

src/scanners/channels_scanner.py

@@ -8,21 +8,17 @@ import discord
 from logs import ChannelLogs, MessageLog
 from .scanner import Scanner
 from data_types import Counter
-from utils import COMMON_HELP_ARGS, mention, channel_mention
+from utils import generate_help, mention, channel_mention
 
 
 class ChannelsScanner(Scanner):
     @staticmethod
     def help() -> str:
-        return (
-            "```\n"
-            + "%chan: Rank channels by their messages\n"
-            + "arguments:\n"
-            + COMMON_HELP_ARGS
-            + "* <n> - top <n>, default is 10\n"
-            + "* all/everyone - include bots\n"
-            + "Example: %chan 10 @user\n"
-            + "```"
+        return generate_help(
+            "chan",
+            "Rank channels by their messages",
+            args=["<n> - top <n>, default is 10", "all/everyone - include bots"],
+            example="5 @user",
         )
 
     def __init__(self):
@@ -34,7 +30,6 @@ class ChannelsScanner(Scanner):
         )
 
     async def init(self, message: discord.Message, *args: str) -> bool:
-        # get max emotes to view
         self.top = 10
         for arg in args:
             if arg.isdigit():
@@ -66,6 +61,7 @@ class ChannelsScanner(Scanner):
                 total_usage=usage_count,
                 counted="message",
                 transform=lambda id: f" by {mention(id)}",
+                top=len(self.members) != 1,
             )
             for name in names
         ]

src/scanners/composition_scanner.py

@@ -8,21 +8,13 @@ import discord
 from .scanner import Scanner
 from data_types import Composition
 from logs import ChannelLogs, MessageLog
-from utils import emojis, COMMON_HELP_ARGS
+from utils import emojis, generate_help
 
 
 class CompositionScanner(Scanner):
     @staticmethod
     def help() -> str:
-        return (
-            "```\n"
-            + "%compo: Show composition statistics\n"
-            + "arguments:\n"
-            + COMMON_HELP_ARGS
-            + "* all/everyone - include bots\n"
-            + "Example: %compo #mychannel1 @user\n"
-            + "```"
-        )
+        return generate_help("compo", "Show composition statistics")
 
     def __init__(self):
         super().__init__(
@@ -65,19 +57,19 @@ class CompositionScanner(Scanner):
             impacted = True
             compo.total_characters += len(message.content)
 
-            emotes_found = emojis.regex.findall(message.content)
-            without_emote = message.content
-            for name in emotes_found:
+            emojis_found = emojis.regex.findall(message.content)
+            without_emoji = message.content
+            for name in emojis_found:
                 if name in emojis.unicode_list or re.match(
                     r"(<a?:[\w\-\~]+:\d+>|:[\w\\-\~]+:)", name
                 ):
-                    compo.emotes[name] += 1
-                    i = without_emote.index(name)
-                    without_emote = without_emote[:i] + without_emote[i + len(name) :]
-            if len(message.content.strip()) > 0 and len(without_emote.strip()) == 0:
-                compo.emote_only += 1
-            if len(emotes_found) > 0:
-                compo.emote_msg += 1
+                    compo.emojis[name] += 1
+                    i = without_emoji.index(name)
+                    without_emoji = without_emoji[:i] + without_emoji[i + len(name) :]
+            if len(message.content.strip()) > 0 and len(without_emoji.strip()) == 0:
+                compo.emoji_only += 1
+            if len(emojis_found) > 0:
+                compo.emoji_msg += 1
 
             links_found = re.findall(r"https?:\/\/", message.content)
             compo.links += len(links_found)
@@ -110,7 +102,7 @@ class CompositionScanner(Scanner):
                 compo.tts += 1
 
             if (
-                len(emotes_found) == 0
+                len(emojis_found) == 0
                 and message.reference is None
                 and not message.image
                 and len(message.mentions) == 0

src/scanners/emojis_scanner.py

@@ -1,44 +1,42 @@
 from typing import Dict, List
-from collections import defaultdict
 import discord
 
 
 # Custom libs
 
 from logs import ChannelLogs, MessageLog
-from data_types import Emote, get_emote_dict
+from data_types import Emoji, get_emoji_dict
 from .scanner import Scanner
-from utils import emojis, COMMON_HELP_ARGS, plural, precise
+from utils import emojis, generate_help, plural, precise
 
 
-class EmotesScanner(Scanner):
+class EmojisScanner(Scanner):
     @staticmethod
     def help() -> str:
-        return (
-            "```\n"
-            + "%emojis: Rank emojis by their usage\n"
-            + "arguments:\n"
-            + COMMON_HELP_ARGS
-            + "* <n> - top <n> emojis, default is 20\n"
-            + "* all - list all common emojis in addition to this guild's\n"
-            + "* members - show top member for each emojis\n"
-            + "* sort:usage/reaction - other sorting methods\n"
-            + "* everyone - include bots\n"
-            + "Example: %emojis 10 all #mychannel1 #mychannel2 @user\n"
-            + "```"
+        return generate_help(
+            "emojis",
+            "Rank emojis by their usage",
+            args=[
+                "<n> - top <n> emojis, default is 20",
+                "all - list all common emojis in addition to this guild's",
+                "members - show top member for each emojis",
+                "sort:usage/reaction - other sorting methods",
+                "everyone - include bots",
+            ],
+            example="10 all #mychannel1 #mychannel2 @user",
         )
 
     def __init__(self):
         super().__init__(
             has_digit_args=True,
             valid_args=["all", "members", "sort:usage", "sort:reaction", "everyone"],
-            help=EmotesScanner.help(),
+            help=EmojisScanner.help(),
             intro_context="Emoji usage",
         )
 
     async def init(self, message: discord.Message, *args: str) -> bool:
         guild = message.channel.guild
-        # get max emotes to view
+        # get max emojis to view
         self.top = 20
         for arg in args:
             if arg.isdigit():
@@ -48,8 +46,8 @@ class EmotesScanner(Scanner):
         self.show_members = "members" in args and (
             len(self.members) == 0 or len(self.members) > 1
         )
-        # Create emotes dict from custom emojis of the guild
-        self.emotes = get_emote_dict(guild)
+        # Create emojis dict from custom emojis of the guild
+        self.emojis = get_emoji_dict(guild)
         self.sort = None
         if "sort:usage" in args:
             self.sort = "usage"
@@ -59,36 +57,36 @@ class EmotesScanner(Scanner):
         return True
 
     def compute_message(self, channel: ChannelLogs, message: MessageLog):
-        return EmotesScanner.analyse_message(
+        return EmojisScanner.analyse_message(
             message,
-            self.emotes,
+            self.emojis,
             self.raw_members,
             all_emojis=self.all_emojis,
             all_messages=self.all_messages,
         )
 
     def get_results(self, intro: str) -> List[str]:
-        names = [name for name in self.emotes]
+        names = [name for name in self.emojis]
         names.sort(
-            key=lambda name: self.emotes[name].score(
+            key=lambda name: self.emojis[name].score(
                 usage_weight=(0 if self.sort == "reaction" else 1),
                 react_weight=(0 if self.sort == "usage" else 1),
             ),
             reverse=True,
         )
         names = names[: self.top]
-        # Get the total of all emotes used
+        # Get the total of all emojis used
         usage_count = 0
         reaction_count = 0
-        for name in self.emotes:
-            usage_count += self.emotes[name].usages
-            reaction_count += self.emotes[name].reactions
+        for name in self.emojis:
+            usage_count += self.emojis[name].usages
+            reaction_count += self.emojis[name].reactions
         res = [intro]
         allow_unused = self.full and len(self.members) == 0
         if self.sort is not None:
             res += [f"(Sorted by {self.sort})"]
         res += [
-            self.emotes[name].to_string(
+            self.emojis[name].to_string(
                 names.index(name),
                 name,
                 total_usage=usage_count,
@@ -97,7 +95,7 @@ class EmotesScanner(Scanner):
                 show_members=self.show_members or len(self.raw_members) == 0,
             )
             for name in names
-            if allow_unused or self.emotes[name].used()
+            if allow_unused or self.emojis[name].used()
         ]
         res += [
             f"Total: {plural(usage_count,'time')} ({precise(usage_count/self.msg_count)}/msg)"
@@ -109,7 +107,7 @@ class EmotesScanner(Scanner):
     @staticmethod
     def analyse_message(
         message: MessageLog,
-        emotes: Dict[str, Emote],
+        emojis_dict: Dict[str, Emoji],
         raw_members: List[int],
         *,
         all_emojis: bool,
@@ -123,27 +121,29 @@ class EmotesScanner(Scanner):
             or message.author in raw_members
         ):
             impacted = True
-            # Find all emotes un the current message in the form "<:emoji:123456789>"
-            # Filter for known emotes
+            # Find all emojis un the current message in the form "<:emoji:123456789>"
+            # Filter for known emojis
             found = emojis.regex.findall(message.content)
-            # For each emote, update its usage
+            # For each emoji, update its usage
             for name in found:
-                if name not in emotes:
+                if name not in emojis_dict:
                     if not all_emojis or name not in emojis.unicode_list:
                         continue
-                emotes[name].usages += 1
-                emotes[name].update_use(message.created_at, [message.author])
-        # For each reaction of this message, test if known emote and update when it's the case
+                emojis_dict[name].usages += 1
+                emojis_dict[name].update_use(message.created_at, [message.author])
+        # For each reaction of this message, test if known emoji and update when it's the case
         for name in message.reactions:
-            if name not in emotes:
+            if name not in emojis_dict:
                 if not all_emojis or name not in emojis.unicode_list:
                     continue
             if len(raw_members) == 0:
-                emotes[name].reactions += len(message.reactions[name])
-                emotes[name].update_use(message.created_at, message.reactions[name])
+                emojis_dict[name].reactions += len(message.reactions[name])
+                emojis_dict[name].update_use(
+                    message.created_at, message.reactions[name]
+                )
             else:
                 for member in raw_members:
                     if member in message.reactions[name]:
-                        emotes[name].reactions += 1
-                        emotes[name].update_use(message.created_at, [member])
+                        emojis_dict[name].reactions += 1
+                        emojis_dict[name].update_use(message.created_at, [member])
         return impacted

src/scanners/find_scanner.py

@@ -0,0 +1,134 @@
+from typing import Dict, List, Optional, Tuple
+from collections import defaultdict
+import discord
+import re
+
+# Custom libs
+
+from logs import ChannelLogs, MessageLog
+from .scanner import Scanner
+from data_types import Counter
+from utils import (
+    generate_help,
+    plural,
+    precise,
+    mention,
+    escape_text,
+)
+
+
+class FindScanner(Scanner):
+    @staticmethod
+    def help() -> str:
+        return generate_help(
+            "find",
+            "Find specific words or phrases (you can use quotes to add spaces in queries, backticks define regexes)",
+            args=[
+                "top - rank users for these queries",
+                "all/everyone - include bots",
+            ],
+            example='#mychannel1 #mychannel2 @user "I love you" "you too"',
+        )
+
+    def __init__(self):
+        super().__init__(
+            all_args=True,
+            valid_args=["all", "everyone", "top"],
+            help=FindScanner.help(),
+            intro_context="Matches",
+        )
+
+    async def init(self, message: discord.Message, *args: str) -> bool:
+        self.matches = defaultdict(Counter)
+        self.all_messages = "all" in args or "everyone" in args
+        self.top = "top" in args or len(self.other_args) == 1
+        if len(self.other_args) == 0:
+            await message.channel.send(
+                "You need to add a query to find (you can use quotes to add spaces in queries, backticks define regexes)",
+                reference=message,
+            )
+            return False
+        self.queries = [
+            (query, query.strip("`") if re.match(r"^`.*`$", query) else None)
+            for query in self.other_args
+        ]
+        return True
+
+    def compute_message(self, channel: ChannelLogs, message: MessageLog):
+        return FindScanner.analyse_message(
+            message,
+            self.matches,
+            self.queries,
+            self.raw_members,
+            all_messages=self.all_messages,
+            top=self.top,
+        )
+
+    def get_results(self, intro: str) -> List[str]:
+        res = [intro]
+        matches = [match for match in self.matches]
+        matches.sort(key=lambda match: self.matches[match].score(), reverse=True)
+        usage_count = Counter.total(self.matches)
+        if self.top:
+            res += [
+                self.matches[match].to_string(
+                    matches.index(match),
+                    mention(match),
+                    total_usage=usage_count,
+                )
+                for match in matches
+            ]
+        else:
+            res += [
+                self.matches[match].to_string(
+                    matches.index(match),
+                    f'"{escape_text(match)}"'
+                    if len(match.strip("`")) == len(match)
+                    else match,
+                    total_usage=self.msg_count,
+                    ranking=False,
+                    transform=lambda id: f" by {mention(id)}",
+                    top=len(self.members) != 1,
+                )
+                for match in matches
+            ]
+        if self.top or len(matches) > 1:
+            res += [
+                f"Total: {plural(usage_count,'time')} ({precise(usage_count/self.msg_count)}/msg)"
+            ]
+        return res
+
+    special_cases = ["'s", "s"]
+
+    @staticmethod
+    def analyse_message(
+        message: MessageLog,
+        matches: Dict[str, Counter],
+        queries: List[Tuple[str, Optional[str]]],
+        raw_members: List[int],
+        *,
+        all_messages: bool,
+        top: bool,
+    ) -> bool:
+        impacted = False
+        # If author is included in the selection (empty list is all)
+        if (
+            (not message.bot or all_messages)
+            and len(raw_members) == 0
+            or message.author in raw_members
+        ):
+            impacted = True
+            content = message.content.lower()
+            for query in queries:
+                if query[1] is not None:
+                    count = len(re.findall(query[1], message.content))
+                else:
+                    count = content.count(query[0].lower())
+                if top:
+                    if count > 0:
+                        matches[message.author].update_use(count, message.created_at)
+                else:
+                    matches[query[0]].update_use(
+                        count, message.created_at, message.author
+                    )
+        return impacted

src/scanners/first_scanner.py

@@ -3,17 +3,28 @@ from typing import List
 # Custom libs
 
 from .history_scanner import HistoryScanner
+from utils import generate_help
 
 
 class FirstScanner(HistoryScanner):
     @staticmethod
     def help() -> str:
-        return super(FirstScanner, FirstScanner).help(
-            cmd="first", text="Read first message"
+        return generate_help(
+            "first",
+            "Read first message  (add text to filter like %find)",
+            args=[
+                "image/gif - pull an image instead of a message",
+                "spoiler:allow/only - allow spoiler images",
+            ],
         )
 
     def __init__(self):
         super().__init__(help=FirstScanner.help())
 
-    def get_results(self, intro: str) -> List[str]:
-        return self.history.to_string(type="first")
+    async def get_results(self, intro: str) -> List[str]:
+        if self.images_only:
+            return await self.history.to_string_image(
+                type="first", spoiler=self.spoiler, gif_only=self.gif_only
+            )
+        else:
+            return self.history.to_string(type="first")

src/scanners/frequency_scanner.py

@@ -8,25 +8,23 @@ import discord
 from .scanner import Scanner
 from data_types import Frequency
 from logs import ChannelLogs, MessageLog
-from utils import COMMON_HELP_ARGS
+from utils import generate_help
 
 
 class FrequencyScanner(Scanner):
     @staticmethod
     def help() -> str:
-        return (
-            "```\n"
-            + "%freq: Show frequency-related statistics\n"
-            + "arguments:\n"
-            + COMMON_HELP_ARGS
-            + "* all/everyone - include bots\n"
-            + "Example: %freq #mychannel1 @user\n"
-            + "```"
+        return generate_help(
+            "freq",
+            "(BETA) Show frequency-related statistics",
+            args=[
+                "graph - plot hours of week",
+            ],
         )
 
     def __init__(self):
         super().__init__(
-            valid_args=["all", "everyone"],
+            valid_args=["all", "everyone", "graph"],
             help=FrequencyScanner.help(),
             intro_context="Frequency",
         )
@@ -34,6 +32,8 @@ class FrequencyScanner(Scanner):
     async def init(self, message: discord.Message, *args: str) -> bool:
         self.freq = Frequency()
         self.all_messages = "all" in args or "everyone" in args
+        self.member_specific = len(self.members) > 0
+        self.to_graph = "graph" in args
         return True
 
     def compute_message(self, channel: ChannelLogs, message: MessageLog):
@@ -43,10 +43,13 @@ class FrequencyScanner(Scanner):
 
     def get_results(self, intro: str) -> List[str]:
         FrequencyScanner.compute_results(self.freq)
-        res = [intro]
-        res += self.freq.to_string(
-            member_specific=self.member_specific,
-        )
+        if self.to_graph:
+            res = self.freq.to_graph()
+        else:
+            res = [intro]
+            res += self.freq.to_string(
+                member_specific=self.member_specific,
+            )
         return res
 
     @staticmethod
@@ -55,7 +58,7 @@ class FrequencyScanner(Scanner):
         freq: Frequency,
         raw_members: List[int],
         *,
-        all_messages: bool
+        all_messages: bool,
     ) -> bool:
         impacted = False
         # If author is included in the selection (empty list is all)
@@ -98,8 +101,7 @@ class FrequencyScanner(Scanner):
                 freq.longest_break_start = latest
             latest = date
             # calculate busiest weekday / hours
-            freq.week[date.weekday()] += 1
-            freq.day[date.hour] += 1
+            freq.hours[date.weekday()][date.hour] += 1
             # calculate busiest day ever
             start_delta = date - freq.dates[0]
             if start_delta.days > current_day:

src/scanners/full_scanner.py

@@ -5,24 +5,18 @@ import discord
 # Custom libs
 
 from .scanner import Scanner
-from . import FrequencyScanner, CompositionScanner, PresenceScanner
+from .composition_scanner import CompositionScanner
+from .frequency_scanner import FrequencyScanner
+from .presence_scanner import PresenceScanner
 from data_types import Frequency, Composition, Presence
 from logs import ChannelLogs, MessageLog
-from utils import COMMON_HELP_ARGS
+from utils import generate_help
 
 
 class FullScanner(Scanner):
     @staticmethod
     def help() -> str:
-        return (
-            "```\n"
-            + "%scan: Show full statistics\n"
-            + "arguments:\n"
-            + COMMON_HELP_ARGS
-            + "* all/everyone - include bots\n"
-            + "Example: %scan #mychannel1 @user\n"
-            + "```"
-        )
+        return generate_help("scan", "Show full statistics")
 
     def __init__(self):
         super().__init__(

src/scanners/history_scanner.py

@@ -1,39 +1,56 @@
 from abc import ABC, abstractmethod
-from typing import List
+from typing import List, Tuple, Optional
 import discord
+import re
 
 # Custom libs
 
 from .scanner import Scanner
 from data_types import History
 from logs import ChannelLogs, MessageLog
-from utils import COMMON_HELP_ARGS
+from utils import FilterLevel
 
 
 class HistoryScanner(Scanner, ABC):
-    @staticmethod
-    def help(*, cmd: str, text: str) -> str:
-        return (
-            "```\n"
-            + f"%{cmd}: {text}\n"
-            + "arguments:\n"
-            + COMMON_HELP_ARGS
-            + "* all/everyone - include bots\n"
-            + "Example: %{cmd} #mychannel1 @user\n"
-            + "```"
-        )
-
     def __init__(self, *, help: str):
         super().__init__(
             has_digit_args=True,
-            valid_args=["all", "everyone"],
+            valid_args=[
+                "all",
+                "everyone",
+                "spoiler",
+                "spoiler:allow",
+                "spoiler:only",
+                "image",
+                "img",
+                "gif",
+            ],
             help=help,
             intro_context="",
+            all_args=True,
         )
 
     async def init(self, message: discord.Message, *args: str) -> bool:
         self.history = History()
         self.all_messages = "all" in args or "everyone" in args
+        self.images_only = "image" in args or "img" in args or "gif" in args
+        self.gif_only = "gif" in args
+        if "spoiler" in args or "spoiler:allow" in args:
+            self.spoiler = FilterLevel.ALLOW
+        elif "spoiler:only" in args:
+            self.spoiler = FilterLevel.ONLY
+        else:
+            self.spoiler = FilterLevel.NONE
+        if not self.images_only:
+            self.queries = [
+                (
+                    query.lower(),
+                    query.strip("`") if re.match(r"^`.*`$", query) else None,
+                )
+                for query in self.other_args
+            ]
+        else:
+            self.queries = []
         return True
 
     def compute_message(self, channel: ChannelLogs, message: MessageLog):
@@ -43,6 +60,8 @@ class HistoryScanner(Scanner, ABC):
             self.history,
             self.raw_members,
             all_messages=self.all_messages,
+            queries=self.queries,
+            images_only=self.images_only,
         )
 
     @abstractmethod
@@ -57,14 +76,28 @@ class HistoryScanner(Scanner, ABC):
         raw_members: List[int],
         *,
         all_messages: bool,
+        queries: List[Tuple[str, Optional[str]]],
+        images_only: bool,
     ) -> bool:
         impacted = False
         # If author is included in the selection (empty list is all)
         if (
-            (not message.bot or all_messages)
-            and len(raw_members) == 0
-            or message.author in raw_members
-        ) and (message.content or message.attachment):
+            (
+                (not message.bot or all_messages)
+                and len(raw_members) == 0
+                or message.author in raw_members
+            )
+            and (message.content or message.attachment)
+            and (not images_only or message.image)
+        ):
+            if not images_only:
+                content = message.content.lower()
+                for query in queries:
+                    if query[1] is not None:
+                        if not re.match(query[1], message.content):
+                            return False
+                    elif not query[0] in content:
+                        return False
             impacted = True
             history.messages += [message]
         return impacted

src/scanners/last_scanner.py

@@ -3,17 +3,28 @@ from typing import List
 # Custom libs
 
 from .history_scanner import HistoryScanner
+from utils import generate_help
 
 
 class LastScanner(HistoryScanner):
     @staticmethod
     def help() -> str:
-        return super(LastScanner, LastScanner).help(
-            cmd="last", text="Read last message"
+        return generate_help(
+            "last",
+            "Read last message (add text to filter like %find)",
+            args=[
+                "image/gif - pull an image instead of a message",
+                "spoiler:allow/only - allow spoiler images",
+            ],
         )
 
     def __init__(self):
         super().__init__(help=LastScanner.help())
 
-    def get_results(self, intro: str) -> List[str]:
-        return self.history.to_string(type="last")
+    async def get_results(self, intro: str) -> List[str]:
+        if self.images_only:
+            return await self.history.to_string_image(
+                type="last", spoiler=self.spoiler, gif_only=self.gif_only
+            )
+        else:
+            return self.history.to_string(type="last")

src/scanners/mentioned_scanner.py

@@ -8,22 +8,18 @@ import discord
 from logs import ChannelLogs, MessageLog
 from .scanner import Scanner
 from data_types import Counter
-from utils import COMMON_HELP_ARGS, plural, precise, mention, alt_mention
+from utils import generate_help, plural, precise, mention, alt_mention
 
 
 class MentionedScanner(Scanner):
     @staticmethod
     def help() -> str:
-        return (
-            "```\n"
-            + "%mentioned: Rank specific user's mentions by their usage\n"
-            + "arguments:\n"
-            + "* @member/me - (required) one or more member\n"
-            + "\n".join(COMMON_HELP_ARGS.split("\n")[1:])
-            + "* <n> - top <n> mentions, default is 10\n"
-            + "* all - include bots mentions\n"
-            + "Example: %mentioned 10 @user\n"
-            + "```"
+        return generate_help(
+            "mentioned",
+            "Rank specific user's mentions by their usage",
+            args=["<n> - top <n>, default is 10", "all/everyone - include bots"],
+            example="5 @user",
+            replace_args=[" @member/me - (required) one or more member"],
         )
 
     def __init__(self):
@@ -35,7 +31,6 @@ class MentionedScanner(Scanner):
         )
 
     async def init(self, message: discord.Message, *args: str) -> bool:
-        # get max emotes to view
         self.top = 10
         for arg in args:
             if arg.isdigit():
@@ -45,7 +40,7 @@ class MentionedScanner(Scanner):
                 "You need to mention at least one member or use `me`", reference=message
             )
             return False
-        self.all_mentions = "all" in args
+        self.all_mentions = "all" in args or "everyone" in args
         # Create mentions dict
         self.mentions = defaultdict(Counter)
         return True
@@ -59,7 +54,6 @@ class MentionedScanner(Scanner):
         names = [name for name in self.mentions]
         names.sort(key=lambda name: self.mentions[name].score(), reverse=True)
         names = names[: self.top]
-        # Get the total of all emotes used
         usage_count = Counter.total(self.mentions)
         res = [intro]
         res += [
@@ -67,6 +61,8 @@ class MentionedScanner(Scanner):
                 names.index(name),
                 name,
                 total_usage=usage_count,
+                transform=lambda id: f" for {mention(id)}",
+                top=len(self.members) != 1,
             )
             for name in names
         ]
@@ -91,6 +87,6 @@ class MentionedScanner(Scanner):
                         mention(member_id)
                     ) + message.content.count(alt_mention(member_id))
                     mentions[mention(message.author)].update_use(
-                        count, message.created_at
+                        count, message.created_at, member_id
                     )
         return impacted

src/scanners/mentions_scanner.py

@@ -9,7 +9,7 @@ from logs import ChannelLogs, MessageLog
 from .scanner import Scanner
 from data_types import Counter
 from utils import (
-    COMMON_HELP_ARGS,
+    generate_help,
     plural,
     precise,
     mention,
@@ -22,16 +22,15 @@ from utils import (
 class MentionsScanner(Scanner):
     @staticmethod
     def help() -> str:
-        return (
-            "```\n"
-            + "%mentions: Rank mentions by their usage\n"
-            + "arguments:\n"
-            + COMMON_HELP_ARGS
-            + "* <n> - top <n> mentions, default is 10\n"
-            + "* all - show role/channel/everyone/here mentions\n"
-            + "* everyone - include bots mentions\n"
-            + "Example: %mentions 10 #mychannel1 #mychannel2 @user\n"
-            + "```"
+        return generate_help(
+            "mentions",
+            "Rank mentions by their usage",
+            args=[
+                "<n> - top <n>, default is 10",
+                "all - show role/channel/everyone/here mentions",
+                "everyone - include bots mentions",
+            ],
+            example="10 #mychannel1 #mychannel2 @user",
         )
 
     def __init__(self):
@@ -43,7 +42,6 @@ class MentionsScanner(Scanner):
         )
 
     async def init(self, message: discord.Message, *args: str) -> bool:
-        # get max emotes to view
         self.top = 10
         for arg in args:
             if arg.isdigit():
@@ -68,7 +66,6 @@ class MentionsScanner(Scanner):
         names = [name for name in self.mentions]
         names.sort(key=lambda name: self.mentions[name].score(), reverse=True)
         names = names[: self.top]
-        # Get the total of all emotes used
         usage_count = Counter.total(self.mentions)
         res = [intro]
         res += [
@@ -76,6 +73,8 @@ class MentionsScanner(Scanner):
                 names.index(name),
                 name,
                 total_usage=usage_count,
+                transform=lambda id: f" by {mention(id)}",
+                top=len(self.members) != 1,
             )
             for name in names
         ]
@@ -106,24 +105,28 @@ class MentionsScanner(Scanner):
                 count = message.content.count(name) + message.content.count(
                     alt_mention(member_id)
                 )
-                mentions[name].update_use(count, message.created_at)
+                mentions[name].update_use(count, message.created_at, message.author)
             if all_mentions:
                 for role_id in message.role_mentions:
                     name = role_mention(role_id)
                     mentions[name].update_use(
-                        message.content.count(name), message.created_at
+                        message.content.count(name), message.created_at, message.author
                     )
                 for channel_id in message.channel_mentions:
                     name = channel_mention(channel_id)
                     mentions[name].update_use(
-                        message.content.count(name), message.created_at
+                        message.content.count(name), message.created_at, message.author
                     )
                 if "@everyone" in message.content:
                     mentions["@\u200beveryone"].update_use(
-                        message.content.count("@everyone"), message.created_at
+                        message.content.count("@everyone"),
+                        message.created_at,
+                        message.author,
                     )
                 if "@here" in message.content:
                     mentions["@\u200bhere"].update_use(
-                        message.content.count("@here"), message.created_at
+                        message.content.count("@here"),
+                        message.created_at,
+                        message.author,
                     )
         return impacted

src/scanners/messages_scanner.py

@@ -8,21 +8,17 @@ import discord
 from logs import ChannelLogs, MessageLog
 from .scanner import Scanner
 from data_types import Counter
-from utils import COMMON_HELP_ARGS, mention, channel_mention
+from utils import generate_help, mention, channel_mention
 
 
 class MessagesScanner(Scanner):
     @staticmethod
     def help() -> str:
-        return (
-            "```\n"
-            + "%msg: Rank users by their messages\n"
-            + "arguments:\n"
-            + COMMON_HELP_ARGS
-            + "* <n> - top <n>, default is 10\n"
-            + "* all/everyone - include bots\n"
-            + "Example: %msg 10 #channel\n"
-            + "```"
+        return generate_help(
+            "msg",
+            "Rank users by their messages",
+            args=["<n> - top <n>, default is 10", "all/everyone - include bots"],
+            example="10 #channel",
         )
 
     def __init__(self):
@@ -34,7 +30,6 @@ class MessagesScanner(Scanner):
         )
 
     async def init(self, message: discord.Message, *args: str) -> bool:
-        # get max emotes to view
         self.top = 10
         for arg in args:
             if arg.isdigit():
@@ -66,6 +61,7 @@ class MessagesScanner(Scanner):
                 total_usage=usage_count,
                 counted="message",
                 transform=lambda id: f" in {channel_mention(id)}",
+                top=self.channels != 1,
             )
             for name in names
         ]

src/scanners/presence_scanner.py

@@ -7,21 +7,13 @@ import discord
 from .scanner import Scanner
 from data_types import Presence
 from logs import ChannelLogs, MessageLog
-from utils import COMMON_HELP_ARGS
+from utils import generate_help
 
 
 class PresenceScanner(Scanner):
     @staticmethod
     def help() -> str:
-        return (
-            "```\n"
-            + "%pres: Show presence statistics\n"
-            + "arguments:\n"
-            + COMMON_HELP_ARGS
-            + "* all/everyone - include bots\n"
-            + "Example: %pres #mychannel1 @user\n"
-            + "```"
-        )
+        return generate_help("pres", "Show presence statistics")
 
     def __init__(self):
         super().__init__(

src/scanners/random_scanner.py

@@ -3,17 +3,28 @@ from typing import List
 # Custom libs
 
 from .history_scanner import HistoryScanner
+from utils import generate_help
 
 
 class RandomScanner(HistoryScanner):
     @staticmethod
     def help() -> str:
-        return super(RandomScanner, RandomScanner).help(
-            cmd="rand", text="Read a random message"
+        return generate_help(
+            "rand",
+            "Read a random message (add text to filter like %find)",
+            args=[
+                "image/gif - pull an image instead of a message",
+                "spoiler:allow/only - allow spoiler images",
+            ],
         )
 
     def __init__(self):
         super().__init__(help=RandomScanner.help())
 
-    def get_results(self, intro: str) -> List[str]:
-        return self.history.to_string(type="random")
+    async def get_results(self, intro: str) -> List[str]:
+        if self.images_only:
+            return await self.history.to_string_image(
+                type="random", spoiler=self.spoiler, gif_only=self.gif_only
+            )
+        else:
+            return self.history.to_string(type="random")

src/scanners/reactions_scanner.py

@@ -8,20 +8,17 @@ import discord
 from logs import ChannelLogs, MessageLog
 from .scanner import Scanner
 from data_types import Counter
-from utils import COMMON_HELP_ARGS, mention, channel_mention
+from utils import generate_help, mention, channel_mention
 
 
 class ReactionsScanner(Scanner):
     @staticmethod
     def help() -> str:
-        return (
-            "```\n"
-            + "%react: Rank users by their reactions\n"
-            + "arguments:\n"
-            + COMMON_HELP_ARGS
-            + "* <n> - top <n>, default is 10\n"
-            + "Example: %react 10 #channel\n"
-            + "```"
+        return generate_help(
+            "react",
+            "Rank users by their reactions",
+            args=["<n> - top <n>, default is 10"],
+            example="10 #channel",
         )
 
     def __init__(self):
@@ -32,7 +29,6 @@ class ReactionsScanner(Scanner):
         )
 
     async def init(self, message: discord.Message, *args: str) -> bool:
-        # get max emotes to view
         self.top = 10
         for arg in args:
             if arg.isdigit():
@@ -62,6 +58,7 @@ class ReactionsScanner(Scanner):
                 total_usage=usage_count,
                 counted="reaction",
                 transform=lambda id: f" in {channel_mention(id)}",
+                top=self.channels != 1,
             )
             for name in names
         ]

src/scanners/scanner.py

@@ -4,12 +4,44 @@ from datetime import datetime
 import logging
 import re
 import discord
+import inspect
 
-from utils import no_duplicate, get_intro, delta
-from logs import GuildLogs, ChannelLogs, MessageLog, ALREADY_RUNNING, CANCELLED
+
+from utils import (
+    no_duplicate,
+    get_intro,
+    delta,
+    gdpr,
+    ISO8601_REGEX,
+    RELATIVE_REGEX,
+    parse_time,
+    command_cache,
+    FilterLevel,
+    SPLIT_TOKEN,
+)
+from logs import (
+    GuildLogs,
+    ChannelLogs,
+    MessageLog,
+    ALREADY_RUNNING,
+    CANCELLED,
+    NO_FILE,
+)
 
 
 class Scanner(ABC):
+    VALID_ARGS = [
+        "me",
+        "here",
+        "fast",
+        "fresh",
+        "mobile",
+        "mention",
+        "nsfw",
+        "nsfw:allow",
+        "nsfw:only",
+    ]
+
     def __init__(
         self,
         *,
@@ -17,12 +49,16 @@ class Scanner(ABC):
         valid_args: List[str] = [],
         help: str,
         intro_context: str,
+        all_args: bool = False,
     ):
         self.has_digit_args = has_digit_args
         self.valid_args = valid_args
+        self.all_args = all_args
         self.help = help
         self.intro_context = intro_context
 
+        self.other_args = []
+
         self.members = []
         self.raw_members = []
         self.full = False
@@ -32,137 +68,271 @@ class Scanner(ABC):
         self.chan_count = 0
 
     async def compute(
-        self, client: discord.client, message: discord.Message, *args: str
+        self,
+        client: discord.client,
+        message: discord.Message,
+        *args: str,
+        other_mentions: List[str] = [],
     ):
         args = list(args)
         guild = message.guild
-        logs = GuildLogs(guild)
+        progress = None
+        try:
+            with GuildLogs(guild) as logs:
+                # If "%cmd help" redirect to "%help cmd"
+                if len(args) > 1 and args[1] == "help":
+                    await client.bot.help(client, message, "help", args[0])
+                    return
 
-        # If "%cmd help" redirect to "%help cmd"
-        if "help" in args:
-            await client.bot.help(client, message, "help", args[0])
-            return
-
-        # check args validity
-        str_channel_mentions = [str(channel.id) for channel in message.channel_mentions]
-        str_mentions = [str(member.id) for member in message.mentions]
-        for i, arg in enumerate(args[1:]):
-            if re.match(r"^<@!?\d+>$", arg):
-                arg = arg[3:-1] if "!" in arg else arg[2:-1]
-            elif re.match(r"^<#!?\d+>$", arg):
-                arg = arg[3:-1] if "!" in arg else arg[2:-1]
-            if (
-                arg not in self.valid_args + ["me", "here", "fast", "fresh"]
-                and (not arg.isdigit() or not self.has_digit_args)
-                and arg not in str_channel_mentions
-                and arg not in str_mentions
-            ):
-                await message.channel.send(
-                    f"Unrecognized argument: `{arg}`", reference=message
-                )
-                return
-
-        # Get selected channels or all of them if no channel arguments
-        self.channels = no_duplicate(message.channel_mentions)
-
-        # transform the "here" arg
-        if "here" in args:
-            self.channels += [message.channel]
-
-        self.full = len(self.channels) == 0
-        if self.full:
-            self.channels = guild.text_channels
-
-        # Get selected members
-        self.members = no_duplicate(message.mentions)
-        self.raw_members = no_duplicate(message.raw_mentions)
-
-        # transform the "me" arg
-        if "me" in args:
-            self.members += [message.author]
-            self.raw_members += [message.author.id]
-
-        if not await self.init(message, *args):
-            return
-
-        # Start computing data
-        async with message.channel.typing():
-            progress = await message.channel.send(
-                "```Starting analysis...```",
-                reference=message,
-                allowed_mentions=discord.AllowedMentions.none(),
-            )
-            total_msg, total_chan = await logs.load(
-                progress, self.channels, fast="fast" in args, fresh="fresh" in args
-            )
-            if total_msg == CANCELLED:
-                await message.channel.send(
-                    "Operation cancelled by user",
-                    reference=message,
-                )
-            elif total_msg == ALREADY_RUNNING:
-                await message.channel.send(
-                    "An analysis is already running on this server, please be patient.",
-                    reference=message,
-                )
-            else:
-                self.msg_count = 0
-                self.total_msg = 0
-                self.chan_count = 0
-                t0 = datetime.now()
-                for channel in self.channels:
-                    if channel.id in logs.channels:
-                        channel_logs = logs.channels[channel.id]
-                        count = sum(
-                            [
-                                self.compute_message(channel_logs, message_log)
-                                for message_log in channel_logs.messages
-                            ]
-                        )
-                        self.total_msg += len(channel_logs.messages)
-                        self.msg_count += count
-                        self.chan_count += 1 if count > 0 else 0
-                logging.info(f"scan {guild.id} > scanned in {delta(t0):,}ms")
-                if self.total_msg == 0:
-                    await message.channel.send(
-                        "There are no messages found matching the filters",
-                        reference=message,
-                    )
-                else:
-                    await progress.edit(content="```Computing results...```")
-                    # Display results
-                    t0 = datetime.now()
-                    results = self.get_results(
-                        get_intro(
-                            self.intro_context,
-                            self.full,
-                            self.channels,
-                            self.members,
-                            self.msg_count,
-                            self.chan_count,
-                        )
-                    )
-                    logging.info(f"scan {guild.id} > results in {delta(t0):,}ms")
-                    response = ""
-                    first = True
-                    for r in results:
-                        if len(response + "\n" + r) > 2000:
+                # check args validity
+                str_channel_mentions = [
+                    str(channel.id) for channel in message.channel_mentions
+                ]
+                str_mentions = [str(member.id) for member in message.mentions]
+                dates = []
+                for i, arg in enumerate(args[1:]):
+                    skip_check = False
+                    if self.all_args and (
+                        f"'{arg}'" in message.content or f'"{arg}"' in message.content
+                    ):
+                        self.other_args += [arg]
+                    elif re.match(r"^<@!?\d+>$", arg):
+                        arg = arg[3:-1] if "!" in arg else arg[2:-1]
+                    elif re.match(r"^<#!?\d+>$", arg):
+                        arg = arg[3:-1] if "!" in arg else arg[2:-1]
+                    elif re.match(ISO8601_REGEX, arg) or re.match(RELATIVE_REGEX, arg):
+                        dates += [parse_time(arg)]
+                        skip_check = True
+                        if len(dates) > 2:
                             await message.channel.send(
-                                response,
-                                reference=message if first else None,
-                                allowed_mentions=discord.AllowedMentions.none(),
+                                f"Too many date arguments: `{arg}`", reference=message
                             )
-                            first = False
-                            response = ""
-                        response += "\n" + r
-                    if len(response) > 0:
+                            return
+                    if (
+                        arg not in self.valid_args + Scanner.VALID_ARGS
+                        and (not arg.isdigit() or not self.has_digit_args)
+                        and arg not in str_channel_mentions
+                        and arg not in str_mentions
+                        and arg not in other_mentions
+                        and not skip_check
+                        and len(arg) > 0
+                    ):
+                        if self.all_args:
+                            self.other_args += [arg]
+                        else:
+                            await message.channel.send(
+                                f"Unrecognized argument: `{arg}`", reference=message
+                            )
+                            return
+
+                for arg in self.other_args:
+                    args.remove(arg)
+
+                self.start_date = None if len(dates) < 1 else min(dates)
+                self.stop_date = None if len(dates) < 2 else max(dates)
+
+                if self.start_date is not None and self.start_date > datetime.now():
+                    await message.channel.send(
+                        f"Start date is after today", reference=message
+                    )
+                    return
+
+                # Get selected channels or all of them if no channel arguments
+                self.channels = no_duplicate(message.channel_mentions)
+
+                # transform the "here" arg
+                if "here" in args:
+                    self.channels += [message.channel]
+
+                self.full = len(self.channels) == 0
+                if self.full:
+                    self.channels = guild.text_channels
+
+                # Get selected members
+                self.members = no_duplicate(message.mentions)
+                self.raw_members = no_duplicate(message.raw_mentions)
+
+                # transform the "me" arg
+                if "me" in args:
+                    self.members += [message.author]
+                    self.raw_members += [message.author.id]
+
+                self.mention_users = "mention" in args or "mobile" in args
+
+                # nsfw filter
+                if "nsfw" in args or "nsfw:allow" in args:
+                    self.nsfw = FilterLevel.ALLOW
+                elif "nsfw:only" in args:
+                    self.nsfw = FilterLevel.ONLY
+                else:
+                    self.nsfw = FilterLevel.NONE
+
+                # fix nsfw filter if channel specified
+                if not self.full and any(channel.nsfw for channel in self.channels):
+                    self.nsfw = FilterLevel.ALLOW
+                elif all(channel.nsfw for channel in self.channels):
+                    self.nsfw = FilterLevel.ONLY
+
+                # filter nsfw channels
+                if self.nsfw == FilterLevel.NONE:
+                    self.channels = list(
+                        filter(lambda channel: not channel.nsfw, self.channels)
+                    )
+                elif self.nsfw == FilterLevel.ONLY:
+                    self.channels = list(
+                        filter(lambda channel: channel.nsfw, self.channels)
+                    )
+
+                if not await self.init(message, *args):
+                    return
+
+                # Start computing data
+                async with message.channel.typing():
+                    progress = await message.channel.send(
+                        "```Starting analysis...```",
+                        reference=message,
+                        allowed_mentions=discord.AllowedMentions.none(),
+                    )
+                    total_msg, total_chan = await logs.load(
+                        progress,
+                        self.channels,
+                        self.start_date,
+                        self.stop_date,
+                        fast="fast" in args,
+                        fresh="fresh" in args,
+                    )
+                    if total_msg == CANCELLED:
                         await message.channel.send(
-                            response,
-                            reference=message if first else None,
-                            allowed_mentions=discord.AllowedMentions.none(),
+                            "Operation cancelled by user",
+                            reference=message,
                         )
-            # Delete custom progress message
-            await progress.delete()
+                    elif total_msg == ALREADY_RUNNING:
+                        await message.channel.send(
+                            "An analysis is already running on this server, please be patient.",
+                            reference=message,
+                        )
+                    elif total_msg == NO_FILE:
+                        await message.channel.send(gdpr.TEXT)
+                    else:
+                        if self.start_date is not None and len(logs.channels) > 0:
+                            self.start_date = max(
+                                self.start_date,
+                                min(
+                                    [
+                                        logs.channels[channel.id].start_date
+                                        for channel in self.channels
+                                        if channel.id in logs.channels
+                                        and logs.channels[channel.id].start_date
+                                        is not None
+                                    ]
+                                ),
+                            )
+                            if self.stop_date is None:
+                                self.stop_date = datetime.utcnow()
+
+                        self.msg_count = 0
+                        self.total_msg = 0
+                        self.chan_count = 0
+                        t0 = datetime.now()
+                        for channel in self.channels:
+                            if channel.id in logs.channels:
+                                channel_logs = logs.channels[channel.id]
+                                count = sum(
+                                    [
+                                        self.compute_message(channel_logs, message_log)
+                                        for message_log in channel_logs.messages
+                                        if (
+                                            self.start_date is None
+                                            or message_log.created_at >= self.start_date
+                                        )
+                                        and (
+                                            self.stop_date is None
+                                            or message_log.created_at <= self.stop_date
+                                        )
+                                    ]
+                                )
+                                self.total_msg += len(channel_logs.messages)
+                                self.msg_count += count
+                                self.chan_count += 1 if count > 0 else 0
+                        logging.info(f"scan {guild.id} > scanned in {delta(t0):,}ms")
+                        if self.msg_count == 0:
+                            await message.channel.send(
+                                "There are no messages found matching the filters",
+                                reference=message,
+                            )
+                        else:
+                            await progress.edit(content="```Computing results...```")
+                            # Display results
+                            t0 = datetime.now()
+                            intro = get_intro(
+                                self.intro_context,
+                                self.full,
+                                self.channels,
+                                self.members,
+                                self.msg_count,
+                                self.chan_count,
+                                self.start_date,
+                                self.stop_date,
+                            )
+                            if inspect.iscoroutinefunction(self.get_results):
+                                results = await self.get_results(intro)
+                            else:
+                                results = self.get_results(intro)
+                            logging.info(
+                                f"scan {guild.id} > results in {delta(t0):,}ms"
+                            )
+                            response = ""
+                            first = True
+                            allowed_mentions = (
+                                discord.AllowedMentions.all()
+                                if self.mention_users
+                                else discord.AllowedMentions.none()
+                            )
+                            file = None
+                            for r in results:
+                                if r:
+                                    if isinstance(r, discord.File):
+                                        file = r
+                                    elif isinstance(r, int) and r == SPLIT_TOKEN:
+                                        await message.channel.send(
+                                            response,
+                                            reference=message if first else None,
+                                            allowed_mentions=allowed_mentions,
+                                            file=file,
+                                        )
+                                        first = False
+                                        file = None
+                                        response = ""
+                                    elif isinstance(r, str):
+                                        if len(response + "\n" + r) > 2000:
+                                            await message.channel.send(
+                                                response,
+                                                reference=message if first else None,
+                                                allowed_mentions=allowed_mentions,
+                                                file=file,
+                                            )
+                                            first = False
+                                            file = None
+                                            response = ""
+                                        response += "\n" + r
+                            if len(response) > 0 or file is not None:
+                                await message.channel.send(
+                                    response,
+                                    reference=message if first else None,
+                                    allowed_mentions=allowed_mentions,
+                                    file=file,
+                                )
+                            command_cache.cache(self, message, args)
+                # Delete custom progress message
+                await progress.delete()
+        except Exception as error:
+            logging.exception(error)
+            await message.channel.send(
+                "An unexpected error happened while computing your command, we're sorry for the inconvenience.",
+                reference=message,
+            )
+            if progress is not None:
+                await progress.delete()
 
     @abstractmethod
     async def init(self, message: discord.Message, *args: str) -> bool:

src/scanners/words_scanner.py

@@ -0,0 +1,122 @@
+from typing import Dict, List
+from collections import defaultdict
+import discord
+import re
+
+# Custom libs
+
+from logs import ChannelLogs, MessageLog
+from .scanner import Scanner
+from data_types import Counter
+from utils import generate_help, plural, precise, mention
+
+
+class WordsScanner(Scanner):
+    @staticmethod
+    def help() -> str:
+        return generate_help(
+            "words",
+            "(BETA) Rank words by their usage",
+            args=[
+                "<n> - words containings <n> or more letters, default is 3",
+                "<n2> - top <n2> words, default is 10",
+                "all/everyone - include bots",
+            ],
+            example="5 10 #mychannel1 #mychannel2 @user",
+        )
+
+    def __init__(self):
+        super().__init__(
+            has_digit_args=True,
+            valid_args=["all", "everyone"],
+            help=WordsScanner.help(),
+            intro_context="Words ({}+ letters)",
+        )
+
+    async def init(self, message: discord.Message, *args: str) -> bool:
+        self.letters = None
+        self.top = None
+        for arg in args:
+            if arg.isdigit():
+                if self.letters is None:
+                    self.letters = int(arg)
+                elif self.top is None:
+                    self.top = int(arg)
+        if self.letters is None:
+            self.letters = 3
+        if self.top is None:
+            self.top = 10
+        self.words = defaultdict(Counter)
+        self.all_messages = "all" in args or "everyone" in args
+        return True
+
+    def compute_message(self, channel: ChannelLogs, message: MessageLog):
+        return WordsScanner.analyse_message(
+            message,
+            self.words,
+            self.raw_members,
+            all_messages=self.all_messages,
+            letters_threshold=self.letters,
+        )
+
+    def get_results(self, intro: str) -> List[str]:
+        words = [word for word in self.words]
+        words.sort(key=lambda word: self.words[word].score(), reverse=True)
+        words = words[: self.top]
+        usage_count = Counter.total(self.words)
+        res = [intro.format(self.letters)]
+        res += [
+            self.words[word].to_string(
+                words.index(word),
+                f"`{word}`",
+                total_usage=usage_count,
+                transform=lambda id: f" by {mention(id)}",
+                top=len(self.members) != 1,
+            )
+            for word in words
+        ]
+        res += [
+            f"Total: {plural(usage_count,'time')} ({precise(usage_count/self.msg_count)}/msg)"
+        ]
+        return res
+
+    special_cases = ["'s", "s"]
+
+    @staticmethod
+    def analyse_message(
+        message: MessageLog,
+        words: Dict[str, Counter],
+        raw_members: List[int],
+        *,
+        all_messages: bool,
+        letters_threshold: int,
+    ) -> bool:
+        impacted = False
+        # If author is included in the selection (empty list is all)
+        if (
+            (not message.bot or all_messages)
+            and len(raw_members) == 0
+            or message.author in raw_members
+        ):
+            impacted = True
+            content = message.content
+            content = re.sub(r"```.+```", "", content, flags=re.DOTALL)
+            content = re.sub(r"`.+`", "", content, flags=re.DOTALL)
+            content = re.sub(r"\w+:\/\/[^ ]+", "", content)
+            for word in re.split("[^\w\-':]", content):
+                m = re.match(
+                    r"(?!^:\w+:$)^[^\w]*((?![\d_])\w[\w\-']*(?![\d_])\w)[^\w]*$", word
+                )
+                if m:
+                    word = m[1].lower()
+                    if len(word) >= letters_threshold:
+                        for case in WordsScanner.special_cases:
+                            if word.endswith(case) and word[: -len(case)] in words:
+                                word = word[: -len(case)]
+                                break
+                            if word + case in words:
+                                words[word] = words[word + case]
+                                del words[word + case]
+                                break
+                        words[word].update_use(1, message.created_at, message.author)
+        return impacted

src/utils/command_cache.py

@@ -0,0 +1,45 @@
+from typing import List
+import logging
+import discord
+
+from scanners import Scanner
+
+command_cache = {}
+
+
+def cache(scanner: Scanner, message: discord.Message, args: List[str]):
+    id = message.channel.id
+    command_cache[id] = (
+        type(scanner),
+        list(args),
+        [str(channel.id) for channel in message.channel_mentions]
+        + [str(member.id) for member in message.mentions],
+    )
+
+
+async def repeat(
+    client: discord.client,
+    message: discord.Message,
+    *args: str,
+    add_args: List[str] = [],
+):
+    if len(args) > 1 and args[1] == "help":
+        await client.bot.help(client, message, "help", args[0])
+        return
+    id = message.channel.id
+    if id not in command_cache:
+        await message.channel.send(
+            "No command to repeat on this channel (type %help for more info)",
+            reference=message,
+        )
+        return
+    (
+        scannerType,
+        original_args,
+        original_mentions,
+    ) = command_cache[id]
+    args = original_args + add_args + list(args[1:]) + ["fast"]
+    logging.info(f"repeating {args}")
+    await scannerType().compute(
+        client, message, *args, other_mentions=original_mentions
+    )

src/utils/gdpr.py

@@ -0,0 +1,65 @@
+import discord
+
+from logs import GuildLogs
+
+
+HELP = """```
+%gdpr: Displays GDPR information
+arguments:
+* agree - agree to GDPR
+* revoke - remove this server's data
+```"""
+
+TEXT = """
+__**About Analyst-bot's data usage**__
+**TL;DR**
+Analyst-bot collects text message information. It does not share collected data with any third-party and data is retained 18 months or until the bot is leaving the guild/server.
+**Data collection**
+Analyst-bot collects a Discord guild/server's history when asked to.
+This includes:
+- Visible text channel names
+- Visible text messages: date and time of creation and edition,  author,  content,  reactions and other available metadata (pinned, tts, etc.)
+This does __not__ includes:
+- Voice channels and not visible channels
+- Not visible text messages
+- Visible text messages' embedded content, images and other attachments
+**Data processing**
+Any data collected is only processed in order to produce a one-time report sent to the user immediately. No temporary data are retained.
+**Data storage and retain policy**
+Analyst-bot stores the collected data in files that are accessible by the software and its administrator only.
+Any collected data are retained maximum 18 months until deletion or when the bot is leaving a guild/server.
+**Data sharing**
+Analyst-bot does not share the data collected with any third-party.
+**Right to retract**
+If you want to have your data removed, you can use the `%gdpr revoke` command or remove this bot from your guild/server.
+**Terms agreement**
+By agreeing to these terms, you ensure having the legal age if you are in a country that does have one and you also ensure having the consent of every member involved.
+
+*If you want more information, please contact the creator of this bot: <https://github.com/Klemek/discord-analyst>.*
+
+Type `%gdpr agree` to agree to these terms, `%gdpr revoke` to remove this guild/server's collected data or `%gdpr` to see this message again.
+"""
+
+AGREE_TEXT = "Thanks for agreeing for these terms, you can now run analysis on this guild/server."
+
+REVOKE_TEXT = "This guild/server's data has been deleted. To run new analysis you must agree to the terms again."
+
+
+async def process(client: discord.client, message: discord.Message, *args: str):
+    args = list(args)
+    if len(args) == 1:
+        await message.channel.send(TEXT)
+    elif args[1] == "help":
+        await client.bot.help(client, message, "help", args[0])
+    elif len(args) > 2:
+        await message.channel.send(f"Too many arguments", reference=message)
+    elif args[1] in ["agree", "accept"]:
+        GuildLogs.init_log(message.channel.guild)
+        await message.channel.send(AGREE_TEXT, reference=message)
+    elif args[1] in ["revoke", "cancel", "remove", "delete"]:
+        GuildLogs.remove_log(message.channel.guild)
+        await message.channel.send(REVOKE_TEXT, reference=message)
+    else:
+        await message.channel.send(
+            f"Unrecognized argument: `{args[1]}`", reference=message
+        )

src/utils/utils.py

@@ -1,19 +1,47 @@
-from typing import List, Dict, Union, Optional, Any
+from enum import IntEnum
+from typing import Callable, List, Dict, Union, Optional, Any
 import os
 import logging
 import discord
 import math
-from datetime import datetime
+from datetime import datetime, timedelta
+import re
+import time
+import dateutil.parser
+from dateutil.relativedelta import relativedelta
 
 # OTHER
 
-COMMON_HELP_ARGS = (
-    ""
-    + "* @member/me - filter for one or more member\n"
-    + "* #channel/here - filter for one or more channel\n"
-    + "* fast - only read cache\n"
-    + "* fresh - does not read cache (long)\n"
-)
+COMMON_HELP_ARGS = [
+    "@member/me - filter for one or more member",
+    "#channel/here - filter for one or more channel",
+    "<date1> - filter after <date1>",
+    "<date2> - filter before <date2>",
+    "fast - only read cache",
+    "fresh - does not read cache (long)",
+    "nsfw:allow/only - allow messages from nsfw channels",
+    "mobile/mention - mentions users (fix @invalid-user bug)",
+]
+
+
+def generate_help(
+    cmd: str,
+    info: str,
+    *,
+    args=["all/everyone - include bots"],
+    example="#mychannel1 @user",
+    replace_args=[],
+):
+    arg_list = "* " + "\n* ".join(
+        args + replace_args + COMMON_HELP_ARGS[len(replace_args) :]
+    )
+    return f"""```
+%{cmd}: {info}
+arguments:
+{arg_list}
+(Sample dates: 2020 / 2021-11 / 2021-06-28 / 2020-06-28T23:00 / today / week / 8days / 1y)
+Example: %{cmd} {example}
+```"""
 
 
 def delta(t0: datetime):
@@ -24,6 +52,35 @@ def deltas(t0: datetime):
     return (datetime.now() - t0).total_seconds()
 
 
+class FilterLevel(IntEnum):
+    NONE = 0
+    ALLOW = 1
+    ONLY = 2
+
+
+SPLIT_TOKEN = 1152317803
+
+
+# FILE
+
+IMAGE_FORMAT = [".png", ".jpg", ".jpeg", ".bmp"]
+EMBED_IMAGES = ["image"]
+
+GIF_FORMAT = [".gif", ".gifv"]
+EMBED_GIF = ["gifv"]
+
+
+def is_extension(filepath: str, ext_list: List[str]) -> bool:
+    filename, file_extension = os.path.splitext(filepath.lower())
+    return file_extension in ext_list
+
+
+def get_resource_path(filename: str) -> str:
+    return os.path.realpath(
+        os.path.join(os.path.dirname(__file__), "..", "resources", filename)
+    )
+
+
 # DISCORD API
 
 
@@ -55,22 +112,50 @@ def message_link(message: discord.Message) -> str:
     return f"https://discord.com/channels/{message.channel.guild.id}/{message.channel.id}/{message.id}"
 
 
+def escape_text(text: str) -> str:
+    return discord.utils.escape_markdown(discord.utils.escape_mentions(text))
+
+
 class FakeMessage:
     def __init__(self, id: int):
         self.id = id
 
 
-# FILE
+def has_image(message: discord.Message) -> bool:
+    for attachment in message.attachments:
+        if is_extension(attachment.filename, GIF_FORMAT + IMAGE_FORMAT):
+            return True
+    for embed in message.embeds:
+        if embed.type in (EMBED_IMAGES + EMBED_GIF):
+            return True
+    return False
 
 
-def is_extension(filepath: str, ext_list: List[str]) -> bool:
-    filename, file_extension = os.path.splitext(filepath.lower())
-    return file_extension in ext_list
+def is_image_spoiler(message: discord.Message) -> bool:
+    if len(message.attachments) > 0:
+        return message.attachments[0].is_spoiler()
+    elif len(message.embeds) > 0:
+        return re.match(r"\|\|[^|]*http[^|]\|\|", message.content.lower()) is not None
+    else:
+        return False
 
 
-def get_resource_path(filename: str) -> str:
-    return os.path.realpath(
-        os.path.join(os.path.dirname(__file__), "..", "resources", filename)
+def is_image_gif(message: discord.Message) -> bool:
+    if len(message.attachments) > 0:
+        return is_extension(message.attachments[0].filename, GIF_FORMAT)
+    elif len(message.embeds) > 0:
+        return message.embeds[0].type in EMBED_GIF
+    else:
+        return False
+
+
+def should_allow_spoiler(message: discord.Message, spoiler: FilterLevel) -> bool:
+    is_spoiler = is_image_spoiler(message)
+    return (
+        not is_spoiler
+        and spoiler <= FilterLevel.ONLY
+        or is_spoiler
+        and spoiler >= FilterLevel.ALLOW
     )
 
 
@@ -92,14 +177,37 @@ def no_duplicate(seq: list) -> list:
 # DICTS
 
 
-def top_key(d: Dict[Union[str, int], int]) -> Union[str, int]:
-    return sorted(d, key=lambda k: d[k])[-1]
+def top_key(
+    d: Dict[Union[str, int], int], key: Optional[Callable] = None, reverse: bool = False
+) -> Union[str, int]:
+    if len(d) == 0:
+        return None
+    if key is None:
+        key = lambda k: d[k]
+    return sorted(d, key=key, reverse=reverse)[-1]
 
 
 def val_sum(d: Dict[Any, int]) -> int:
+    if len(d) == 0:
+        return 0
     return sum(d.values())
 
 
+def serialize(
+    obj: Any, *, not_serialized: List[str] = [], dates: List[str] = []
+) -> Dict:
+    output = dict(obj.__dict__)
+    for key in not_serialized:
+        output.pop(key, None)
+    for key in dates:
+        if output[key]:
+            try:
+                output[key] = getattr(obj, key).isoformat()
+            except AttributeError:
+                pass
+    return output
+
+
 # MESSAGE FORMATTING
 
 
@@ -135,38 +243,87 @@ def precise(p: float, *, precision: int = 2) -> str:
 
 # DATE FORMATTING
 
+ISO8601_REGEX = r"^([\+-]?\d{4}(?!\d{2}\b))((-?)((0[1-9]|1[0-2])(\3([12]\d|0[1-9]|3[01]))?|W([0-4]\d|5[0-2])(-?[1-7])?|(00[1-9]|0[1-9]\d|[12]\d{2}|3([0-5]\d|6[1-6])))([T\s]((([01]\d|2[0-3])((:?)[0-5]\d)?|24\:?00)([\.,]\d+(?!:))?)?(\17[0-5]\d([\.,]\d+)?)?([zZ]|([\+-])([01]\d|2[0-3]):?([0-5]\d)?)?)?)?$"
+ISO8601_FULL = "0000-01-01T00:00:00"
+
+
+def parse_iso_datetime(str_date: str) -> datetime:
+    if re.match(
+        "^\d{4}(-\d{2}(-\d{2}(T\d{2}(:\d{2}(:\d{2}(:\d{2})?)?)?)?)?)?$", str_date
+    ):
+        str_date = str_date + "0000-01-01T00:00:00"[len(str_date) :]
+    return dateutil.parser.parse(str_date)
+
+
+RELATIVE_REGEX = r"(yesterday|today|\d*hours?|\d+h(ours?)?|\d*days?|\d+d(ays?)?|\d*weeks?|\d+w(eeks?)?|\d*months?|\d+m(onths?)?|\d*years?|\d+y(ears?)?)"
+
+
+def parse_relative_time(src: str) -> datetime:
+    today = datetime.utcnow().date()
+    today = datetime(today.year, today.month, today.day)
+    if src == "today":
+        return today
+    elif src == "yesterday":
+        return today - relativedelta(days=1)
+    else:
+        m = re.match("(\d*)(\w+)", src)
+        delta = None
+        value = int(m[1]) if m[1] else 1
+        unit = m[2][0]
+        if unit == "h":
+            delta = relativedelta(hours=value)
+        elif unit == "d":
+            delta = relativedelta(days=value)
+        elif unit == "w":
+            delta = relativedelta(weeks=value)
+        elif unit == "m":
+            delta = relativedelta(months=value)
+        elif unit == "y":
+            delta = relativedelta(years=value)
+        return datetime.utcnow() - delta
+
+
+def parse_time(src: str) -> datetime:
+    if re.match(RELATIVE_REGEX, src):
+        return parse_relative_time(src)
+    else:
+        return parse_iso_datetime(src)
+
 
 def str_date(date: datetime) -> str:
-    return date.strftime("%d %b. %Y")  # 12 Jun. 2018
+    return f"<t:{int(time.mktime(date.timetuple()))}:D>"
 
 
 def str_datetime(date: datetime) -> str:
-    return date.strftime("%H:%M, %d %b. %Y")  # 12:05, 12 Jun. 2018
+    return f"<t:{int(time.mktime(date.timetuple()))}:f>"
 
 
-def from_now(src: Optional[datetime]) -> str:
-    if src is None:
-        return "never"
-    delay = datetime.utcnow() - src
+def str_delta(delay: timedelta) -> str:
     seconds = delay.seconds
     minutes = seconds // 60
     hours = minutes // 60
     if delay.days < 1:
         if hours < 1:
             if minutes == 0:
-                return "now"
+                return "no time"
             elif minutes == 1:
-                return "a minute ago"
+                return "a minute"
             else:
-                return f"{minutes} minutes ago"
+                return f"{minutes} minutes"
         elif hours == 1:
-            return "an hour ago"
+            return "an hour"
         else:
-            return f"{hours} hours ago"
+            return f"{hours} hours"
     elif delay.days == 1:
-        return "yesterday"
+        return "one day"
     else:
-        return f"{delay.days:,} days ago"
+        return f"{delay.days:,} days"
+
+
+def from_now(src: Optional[datetime]) -> str:
+    if src is None:
+        return "never"
+    return f"<t:{int(time.mktime(src.timetuple()))}:R>"
 
 
 # APP SPECIFIC
@@ -179,46 +336,48 @@ def get_intro(
     members: List[discord.Member],
     nmm: int,  # number of messages impacted
     nc: int,  # number of impacted channels
+    start_datetime: datetime,
+    stop_datetime: datetime,
 ) -> str:
     """
     Get the introduction sentence of the response
     """
+    time_text = ""
+    if start_datetime is not None:
+        stop_datetime = datetime.now() if stop_datetime is None else stop_datetime
+        time_text = f" (in {str_delta(stop_datetime - start_datetime)})"
     # Show all data (members, channels) when it's less than 5 units
     if len(members) == 0:
         # Full scan of the server
         if full:
-            return f"{subject} in this server ({nc} channels, {nmm:,} messages):"
+            return f"{subject} in this server ({nc} channels, {nmm:,} messages){time_text}:"
         elif len(channels) < 5:
-            return f"{aggregate([c.mention for c in channels])} {subject.lower()} in {nmm:,} messages:"
+            return f"{aggregate([c.mention for c in channels])} {subject.lower()} in {nmm:,} messages{time_text}:"
         else:
-            return (
-                f"These {len(channels)} channels {subject.lower()} in {nmm:,} messages:"
-            )
+            return f"These {len(channels)} channels {subject.lower()} in {nmm:,} messages{time_text}:"
     elif len(members) < 5:
         if full:
-            return f"{aggregate([m.mention for m in members])} {subject.lower()} in {nmm:,} messages:"
+            return f"{aggregate([m.mention for m in members])} {subject.lower()} in {nmm:,} messages{time_text}:"
         elif len(channels) < 5:
             return (
                 f"{aggregate([m.mention for m in members])} on {aggregate([c.mention for c in channels])} "
-                f"{subject.lower()} in {nmm:,} messages:"
+                f"{subject.lower()} in {nmm:,} messages{time_text}:"
             )
         else:
             return (
                 f"{aggregate([m.mention for m in members])} on these {len(channels)} channels "
-                f"{subject.lower()} in {nmm:,} messages:"
+                f"{subject.lower()} in {nmm:,} messages{time_text}:"
             )
     else:
         if full:
-            return (
-                f"These {len(members)} members {subject.lower()} in {nmm:,} messages:"
-            )
+            return f"These {len(members)} members {subject.lower()} in {nmm:,} messages{time_text}:"
         elif len(channels) < 5:
             return (
                 f"These {len(members)} members on {aggregate([c.mention for c in channels])} "
-                f"{subject.lower()} in {nmm:,} messages:"
+                f"{subject.lower()} in {nmm:,} messages{time_text}:"
             )
         else:
             return (
                 f"These {len(members)} members on these {len(channels)} channels "
-                f"{subject.lower()} in {nmm:,} messages:"
+                f"{subject.lower()} in {nmm:,} messages{time_text}:"
             )

test-requirements.txt

@@ -0,0 +1,3 @@
+pytest~=6.2.3
+pytest-cov
+coveralls

tests/integration/scanners/test_first_scanner.py

@@ -0,0 +1,90 @@
+from unittest import TestCase
+from unittest.mock import MagicMock
+from src.scanners import FirstScanner
+from datetime import datetime, timedelta
+
+from tests.utils import AsyncTestCase, fake_message
+
+
+class TestFirstScanner(AsyncTestCase):
+    def test_help(self):
+        self.assertGreater(len(FirstScanner.help()), 0)
+        self.assertIn("%first", FirstScanner.help())
+
+    def test_empty_no_messages(self):
+        scanner = FirstScanner()
+
+        command_msg = MagicMock()
+        self._await(scanner.init(command_msg, []))
+
+        results = self._await(scanner.get_results(""))
+        self.assertListEqual(["There was no messages matching your filters"], results)
+
+    def test_empty_filtered(self):
+        scanner = FirstScanner()
+        scanner.raw_members = [1]
+
+        self._await(scanner.init(fake_message(), []))
+
+        messages = [fake_message(author=2), fake_message(author=3)]
+
+        for msg in messages:
+            scanner.compute_message(msg.channel, msg)
+
+        results = self._await(scanner.get_results(""))
+        self.assertListEqual(["There was no messages matching your filters"], results)
+
+    def test_normal(self):
+        scanner = FirstScanner()
+
+        self._await(scanner.init(fake_message(), []))
+
+        messages = [
+            fake_message(id=1, created_at=timedelta(days=-2)),
+            fake_message(id=2, created_at=timedelta(days=-3)),
+            fake_message(id=3, created_at=timedelta(days=-1)),
+        ]
+
+        for msg in messages:
+            scanner.compute_message(msg.channel, msg)
+
+        results = self._await(scanner.get_results(""))
+
+        expected = messages[1]
+        self.assertListEqual(
+            [
+                "First message out of 3",
+                f"{expected.created_at.strftime('%H:%M, %d %b. %Y')} (2 days ago) <@1> said:",
+                f"> {expected.content}",
+                "<https://discord.com/channels/1/1/2>",
+            ],
+            results,
+        )
+
+    def test_filtered(self):
+        scanner = FirstScanner()
+        scanner.raw_members = [1]
+
+        self._await(scanner.init(fake_message(), []))
+
+        messages = [
+            fake_message(id=1, author=1, created_at=timedelta(days=-2)),
+            fake_message(id=2, author=2, created_at=timedelta(days=-3)),
+            fake_message(id=3, author=1, created_at=timedelta(days=-1)),
+        ]
+
+        for msg in messages:
+            scanner.compute_message(msg.channel, msg)
+
+        results = self._await(scanner.get_results(""))
+
+        expected = messages[0]
+        self.assertListEqual(
+            [
+                "First message out of 2",
+                f"{expected.created_at.strftime('%H:%M, %d %b. %Y')} (yesterday) <@1> said:",
+                f"> {expected.content}",
+                "<https://discord.com/channels/1/1/1>",
+            ],
+            results,
+        )

tests/utils.py

@@ -0,0 +1,99 @@
+from typing import List, Optional, Dict, Union
+from unittest import TestCase
+import asyncio
+from datetime import datetime, timedelta
+from unittest.mock import MagicMock
+import random
+import string
+
+
+class AsyncTestCase(TestCase):
+    def setUp(self):
+        self.loop = asyncio.new_event_loop()
+        asyncio.set_event_loop(None)
+
+    def tearDown(self):
+        self.loop.close()
+
+    def _await(self, fn):
+        return self.loop.run_until_complete(fn)
+
+
+RANDOM_TEXT_CHARS = string.ascii_letters + string.digits + string.punctuation
+
+
+def random_text(min_len: int = 3, max_len: int = 45):
+    return "".join(
+        random.choice(RANDOM_TEXT_CHARS)
+        for _ in range(random.randrange(min_len, max_len))
+    )
+
+
+def fake_guild(id: int = 1):
+    return MagicMock(id=id)
+
+
+def fake_channel(id: int = 1, name: str = "fake-channel"):
+    return MagicMock(id=id, name=name, guild=fake_guild())
+
+
+def fake_message(
+    id: int = 1,
+    channel_id: int = 1,
+    channel_name: str = "fake-channel",
+    created_at: Optional[Union[datetime, timedelta]] = None,
+    edited_at: Optional[datetime] = None,
+    author: int = 1,
+    pinned: bool = False,
+    mention_everyone: bool = False,
+    tts: bool = False,
+    bot: bool = False,
+    content: Optional[str] = None,
+    mentions: Optional[List[int]] = None,
+    reference: Optional[int] = None,
+    role_mentions: Optional[List[int]] = None,
+    channel_mentions: Optional[List[int]] = None,
+    image: bool = False,
+    attachment: bool = False,
+    embed: bool = False,
+    reactions: Optional[Dict[str, List[int]]] = None,
+):
+    if created_at is None:
+        created_at = datetime.now() + timedelta(hours=random.randrange(-30 * 24, 0))
+    elif isinstance(created_at, timedelta):
+        created_at = datetime.now() + created_at
+    if isinstance(edited_at, timedelta):
+        edited_at = datetime.now() + edited_at
+    if content is None:
+        content = random_text()
+    if mentions is None:
+        mentions = []
+    if role_mentions is None:
+        role_mentions = []
+    if channel_mentions is None:
+        channel_mentions = []
+    if reactions is None:
+        reactions = {}
+    return MagicMock(
+        id=id,
+        channel=fake_channel(channel_id, channel_name),
+        created_at=created_at,
+        edited_at=edited_at,
+        author=author,
+        pinned=pinned,
+        mention_everyone=mention_everyone,
+        tts=tts,
+        bot=bot,
+        content=content,
+        mentions=mentions,
+        raw_mentions=mentions,
+        reference=reference,
+        role_mentions=role_mentions,
+        raw_role_mentions=role_mentions,
+        channel_mentions=channel_mentions,
+        raw_channel_mentions=channel_mentions,
+        image=image,
+        attachment=attachment,
+        embed=embed,
+        reactions=reactions,
+    )