Several Smaller Updates (#121)

* Remove User Data Tracking For Join Requests * Ignore some frequent exceptions * Improve Reply-Search * Make use of `post_shutdown` * Add `/traceback` TagHint and link to alternative pastebin services
python-telegram-bot · Oct 15, 2023 · 5ba66cb · 5ba66cb
1 parent c318ec8
commit 5ba66cb
Show file tree

Hide file tree

Showing 5 changed files with 66 additions and 92 deletions.
diff --git a/components/callbacks.py b/components/callbacks.py
@@ -216,52 +216,54 @@ async def reply_search(update: Update, context: ContextTypes.DEFAULT_TYPE) -> No
     If the message is a reply, the bot will reply to the referenced message directly.
     """
     message = cast(Message, update.effective_message)
+    if not message.text:
+        return
+
     last = 0.0
     github_matches: List[Tuple[int, Tuple[str, str, str, str, str]]] = []
     found_entries: List[Tuple[int, BaseEntry]] = []
-
     no_entity_text = get_text_not_in_entities(message).strip()
 
     search = cast(Search, context.bot_data["search"])
     github = search.github
 
     # Parse exact matches for GitHub threads & ptbcontrib found_entries first
-    if not (no_entity_text.startswith("!search") or no_entity_text.endswith("!search")):
-        for match in GITHUB_PATTERN.finditer(no_entity_text):
-            logging.debug(match.groupdict())
-            owner, repo, number, sha, ptbcontrib = (
-                cast(str, match.groupdict()[x])
-                for x in ("owner", "repo", "number", "sha", "ptbcontrib")
-            )
-            if number or sha or ptbcontrib:
-                github_matches.append((match.start(), (owner, repo, number, sha, ptbcontrib)))
-
-        for gh_match in github_matches:
-            last = keep_typing(
-                last,
-                cast(Chat, update.effective_chat),
-                ChatAction.TYPING,
-                application=context.application,
-            )
-            owner, repo, number, sha, ptbcontrib = gh_match[1]
-            owner = owner or DEFAULT_REPO_OWNER
-            repo = repo or DEFAULT_REPO_NAME
-            if number:
-                issue = await github.get_thread(int(number), owner, repo)
-                if issue is not None:
-                    found_entries.append((gh_match[0], issue))
-            elif sha:
-                commit = await github.get_commit(sha, owner, repo)
-                if commit is not None:
-                    found_entries.append((gh_match[0], commit))
-            elif ptbcontrib:
-                contrib = github.ptb_contribs.get(ptbcontrib)
-                if contrib:
-                    found_entries.append((gh_match[0], contrib))
-
-    else:
-        # Parse fuzzy search next
-        for match in ENCLOSED_REGEX.finditer(no_entity_text):
+    for match in GITHUB_PATTERN.finditer(no_entity_text):
+        logging.debug(match.groupdict())
+        owner, repo, number, sha, ptbcontrib = (
+            cast(str, match.groupdict()[x])
+            for x in ("owner", "repo", "number", "sha", "ptbcontrib")
+        )
+        if number or sha or ptbcontrib:
+            github_matches.append((match.start(), (owner, repo, number, sha, ptbcontrib)))
+
+    for gh_match in github_matches:
+        last = keep_typing(
+            last,
+            cast(Chat, update.effective_chat),
+            ChatAction.TYPING,
+            application=context.application,
+        )
+        owner, repo, number, sha, ptbcontrib = gh_match[1]
+        owner = owner or DEFAULT_REPO_OWNER
+        repo = repo or DEFAULT_REPO_NAME
+        if number:
+            issue = await github.get_thread(int(number), owner, repo)
+            if issue is not None:
+                found_entries.append((gh_match[0], issue))
+        elif sha:
+            commit = await github.get_commit(sha, owner, repo)
+            if commit is not None:
+                found_entries.append((gh_match[0], commit))
+        elif ptbcontrib:
+            contrib = github.ptb_contribs.get(ptbcontrib)
+            if contrib:
+                found_entries.append((gh_match[0], contrib))
+
+    # Parse fuzzy search next, if requested. Here we use message.text instead of no_entity_text
+    # to avoid tlds like .bot and .app to mess things up for us
+    if message.text.startswith("!search") or message.text.endswith("!search"):
+        for match in ENCLOSED_REGEX.finditer(message.text):
             last = keep_typing(
                 last,
                 cast(Chat, update.effective_chat),

diff --git a/components/errorhandler.py b/components/errorhandler.py
@@ -2,7 +2,7 @@
 import json
 import logging
 import traceback
-from typing import Optional, cast
+from typing import cast
 
 from telegram import Update
 from telegram.error import BadRequest
@@ -34,20 +34,6 @@ async def error_handler(update: object, context: CallbackContext) -> None:
     )
     message_2 = f"<pre>{html.escape(tb_string)}</pre>"
 
-    user_id: Optional[int] = None
-    message_3 = ""
-    if isinstance(update, Update) and update.effective_user:
-        user_id = update.effective_user.id
-    if context.job:
-        user_id = context.job.user_id
-    if user_id:
-        data_str = html.escape(
-            json.dumps(context.application.user_data[user_id], indent=2, ensure_ascii=False)
-        )
-        message_3 = (
-            "<code>user_data</code> associated with this exception:\n\n" f"<pre>{data_str}</pre>"
-        )
-
     # Finally, send the messages
     # We send update and traceback in two parts to reduce the chance of hitting max length
     try:
@@ -61,11 +47,6 @@ async def error_handler(update: object, context: CallbackContext) -> None:
                 f"Hey.\nThe error <code>{html.escape(str(context.error))}</code> happened."
                 f" The traceback is too long to send, but it was written to the log."
             )
-            sent_message = await context.bot.send_message(
-                chat_id=ERROR_CHANNEL_CHAT_ID, text=message
-            )
+            await context.bot.send_message(chat_id=ERROR_CHANNEL_CHAT_ID, text=message)
         else:
             raise exc
-    finally:
-        if message_3:
-            await sent_message.reply_html(message_3)
diff --git a/components/joinrequests.py b/components/joinrequests.py
@@ -1,5 +1,5 @@
 import datetime
-from typing import Any, Dict, Tuple, Union, cast
+from typing import Tuple, Union, cast
 
 from telegram import (
     CallbackQuery,
@@ -30,19 +30,14 @@ def get_dtm_str() -> str:
 async def approve_user(
     user: Union[int, User], chat_id: int, group_name: str, context: ContextTypes.DEFAULT_TYPE
 ) -> None:
-    user_data = cast(Dict[Any, Any], context.user_data)
     try:
         if isinstance(user, User):
             await user.approve_join_request(chat_id=chat_id)
         else:
             await context.bot.approve_chat_join_request(user_id=user, chat_id=chat_id)
-        user_data.setdefault(int(chat_id), {}).setdefault("approved", []).append(get_dtm_str())
     except BadRequest as exc:
         user_mention = f"{user.username} - {user.id}" if isinstance(user, User) else str(user)
         error_message = f"{exc} - {user_mention} - {group_name}"
-        user_data.setdefault(int(chat_id), {}).setdefault("approve failed", []).append(
-            f"{get_dtm_str()}: {exc}"
-        )
         raise BadRequest(error_message) from exc
     except Forbidden as exc:
         if "user is deactivated" not in exc.message:
@@ -52,19 +47,14 @@ async def approve_user(
 async def decline_user(
     user: Union[int, User], chat_id: int, group_name: str, context: ContextTypes.DEFAULT_TYPE
 ) -> None:
-    user_data = cast(Dict[Any, Any], context.user_data)
     try:
         if isinstance(user, User):
             await user.decline_join_request(chat_id=chat_id)
         else:
             await context.bot.decline_chat_join_request(user_id=user, chat_id=chat_id)
-        user_data.setdefault(int(chat_id), {}).setdefault("declined", []).append(get_dtm_str())
     except BadRequest as exc:
         user_mention = f"{user.username} - {user.id}" if isinstance(user, User) else str(user)
         error_message = f"{exc} - {user_mention} - {group_name}"
-        user_data.setdefault(int(chat_id), {}).setdefault("declined failed", []).append(
-            f"{get_dtm_str()}: {exc}"
-        )
         raise BadRequest(error_message) from exc
     except Forbidden as exc:
         if "user is deactivated" not in exc.message:
@@ -81,16 +71,9 @@ async def join_request_callback(update: Update, context: ContextTypes.DEFAULT_TY
         # No need to ping the user again if we already did
         return
 
-    user_data = cast(Dict[Any, Any], context.user_data)
     on_topic = join_request.chat.username == ONTOPIC_USERNAME
     group_mention = ONTOPIC_CHAT_ID if on_topic else OFFTOPIC_CHAT_ID
 
-    user_data.setdefault(int(chat_id), {}).setdefault("received join request", []).append(
-        get_dtm_str()
-    )
-    user_data[int(chat_id)]["user_chat_id"] = user_chat_id
-    user_data[int(chat_id)]["user_id"] = user.id
-
     text = (
         f"Hi, {user.mention_html()}! I'm {context.bot.bot.mention_html()}, the "
         f"guardian of the group {group_mention}, that you requested to join.\n\nBefore you can "
@@ -114,7 +97,6 @@ async def join_request_callback(update: Update, context: ContextTypes.DEFAULT_TY
         # If the user blocked the bot, let's give the admins a chance to handle that
         # TG also notifies the user and forwards the message once the user unblocks the bot, but
         # forwarding it still doesn't hurt ...
-        user_data.setdefault("blocked bot", []).append(get_dtm_str())
         text = (
             f"User {user.mention_html()} with id {user.id} requested to join the group "
             f"{join_request.chat.username} but has blocked me. Please manually handle this."
@@ -133,23 +115,16 @@ async def join_request_callback(update: Update, context: ContextTypes.DEFAULT_TY
 
 
 async def join_request_buttons(update: Update, context: ContextTypes.DEFAULT_TYPE) -> None:
-    user_data = cast(Dict[Any, Any], context.user_data)
     callback_query = cast(CallbackQuery, update.callback_query)
     user = cast(User, update.effective_user)
     _, press, chat_id = cast(str, callback_query.data).split()
     if press == "2":
-        user_data.setdefault(int(chat_id), {}).setdefault("pressed button 2", []).append(
-            get_dtm_str()
-        )
         jobs = cast(JobQueue, context.job_queue).get_jobs_by_name(
             f"JOIN_TIMEOUT {chat_id} {user.id}"
         )
         if jobs:
             for job in jobs:
                 job.schedule_removal()
-                user_data.setdefault(int(chat_id), {}).setdefault(
-                    "removed join timeout", []
-                ).append(get_dtm_str())
 
         try:
             await approve_user(
@@ -162,9 +137,6 @@ async def join_request_buttons(update: Update, context: ContextTypes.DEFAULT_TYP
 
         reply_markup = None
     else:
-        user_data.setdefault(int(chat_id), {}).setdefault("pressed button 1", []).append(
-            get_dtm_str()
-        )
         reply_markup = InlineKeyboardMarkup.from_button(
             InlineKeyboardButton(
                 text="⚠️ Tap again to confirm",
@@ -193,3 +165,11 @@ async def join_request_timeout_job(context: ContextTypes.DEFAULT_TYPE) -> None:
     except Forbidden as exc:
         if "user is deactivated" not in exc.message:
             raise exc
+    except BadRequest as exc:
+        # These apparently happen frequently, e.g. when user clear the chat
+        if exc.message not in [
+            "Message to edit not found",
+            "Can't access the chat",
+            "Chat not found",
+        ]:
+            raise exc
diff --git a/components/taghints.py b/components/taghints.py
@@ -120,8 +120,9 @@
     "pastebin": {
         "message": (
             "{query} Please post code or tracebacks using a pastebin rather than via plain text "
-            "or a picture. https://pastebin.com/ is quite popular, but there are many "
-            "alternatives out there. Of course, for very short snippets, text is fine. Please at "
+            "or a picture. https://pastebin.com/ is quite popular, but there are "
+            "<a href='https://github.com/lorien/awesome-pastebin'>many alternatives</a> "
+            "out there. Of course, for very short snippets, text is fine. Please at "
             "least format it as monospace in that case."
         ),
         "help": "Ask users not to post code as text or images.",
@@ -283,6 +284,18 @@
         "help": "Tell users not to use AI/LLM generated answers",
         "group_command": True,
     },
+    "traceback": {
+        "message": (
+            "{query} Please show the <i>full</i> traceback via a pastebin. Make sure to include "
+            "everything from the first <code>Traceback (most recent call last):</code> until the "
+            "last error message. https://pastebin.com/ is a popular pastebin service, but there "
+            "are <a href='https://github.com/lorien/awesome-pastebin'>many alternatives</a> out "
+            "there."
+        ),
+        "default": "Hey.",
+        "help": "Ask for the full traceback",
+        "group_command": True,
+    },
 }
 
 

diff --git a/rules_bot.py b/rules_bot.py
@@ -1,4 +1,3 @@
-import asyncio
 import configparser
 import logging
 import os
@@ -111,6 +110,7 @@ def main() -> None:
         .token(config["KEYS"]["bot_api"])
         .defaults(defaults)
         .post_init(post_init)
+        .post_shutdown(post_shutdown)
         .job_queue(RulesJobQueue())
         .build()
     )
@@ -207,8 +207,6 @@ def main() -> None:
     application.add_error_handler(error_handler)
 
     application.run_polling(allowed_updates=Update.ALL_TYPES, close_loop=False)
-    # Can be used in AppBuilder.post_shutdown once #3126 is released
-    asyncio.get_event_loop().run_until_complete(post_shutdown(application))
 
 
 if __name__ == "__main__":