markdown: Convert topic links generated by "#-mentions" to permalinks.

This commit converts the links generated by the markdown
of the "#-mention" of topics to permalinks -- the links containing
the "with" narrow operator, the operand being the last message
of the channel and topic of the mention.

Part of zulip#21505

zerver/lib/markdown/__init__.py

@@ -48,6 +48,7 @@
 from zerver.lib.markdown.fenced_code import FENCE_RE
 from zerver.lib.mention import (
     BEFORE_MENTION_ALLOWED_REGEX,
+    ChannelTopicInfo,
     FullNameInfo,
     MentionBackend,
     MentionData,
@@ -129,6 +130,7 @@ class DbData:
     active_realm_emoji: dict[str, EmojiInfo]
     sent_by_bot: bool
     stream_names: dict[str, int]
+    topic_info: dict[ChannelTopicInfo, int]
     translate_emoticons: bool
     user_upload_previews: dict[str, MarkdownImageMetadata]
 
@@ -2024,6 +2026,13 @@ def find_stream_id(self, name: str) -> int | None:
         stream_id = db_data.stream_names.get(name)
         return stream_id
 
+    def get_with_operand(self, channel_topic: ChannelTopicInfo) -> int | None:
+        db_data: DbData | None = self.zmd.zulip_db_data
+        if db_data is None:
+            return None
+        with_operand = db_data.topic_info.get(channel_topic)
+        return with_operand
+
     @override
     def handleMatch(  # type: ignore[override] # https://github.com/python/mypy/issues/10197
         self, m: Match[str], data: str
@@ -2039,7 +2048,13 @@ def handleMatch(  # type: ignore[override] # https://github.com/python/mypy/issu
         el.set("data-stream-id", str(stream_id))
         stream_url = encode_stream(stream_id, stream_name)
         topic_url = hash_util_encode(topic_name)
-        link = f"/#narrow/channel/{stream_url}/topic/{topic_url}"
+        channel_topic_object = ChannelTopicInfo(stream_name, topic_name)
+        with_operand = self.get_with_operand(channel_topic_object)
+        if with_operand is not None:
+            link = f"/#narrow/channel/{stream_url}/topic/{topic_url}/with/{with_operand}"
+        else:
+            link = f"/#narrow/channel/{stream_url}/topic/{topic_url}"
+
         el.set("href", link)
         text = f"#{stream_name} > {topic_name}"
         el.text = markdown.util.AtomicString(text)
@@ -2056,6 +2071,13 @@ def possible_linked_stream_names(content: str) -> set[str]:
     }
 
 
+def possible_linked_topics(content: str) -> set[ChannelTopicInfo]:
+    return {
+        ChannelTopicInfo(match.group("stream_name"), match.group("topic_name"))
+        for match in re.finditer(STREAM_TOPIC_LINK_REGEX, content, re.VERBOSE)
+    }
+
+
 class AlertWordNotificationProcessor(markdown.preprocessors.Preprocessor):
     allowed_before_punctuation = {" ", "\n", "(", '"', ".", ",", "'", ";", "[", "*", "`", ">"}
     allowed_after_punctuation = {
@@ -2670,6 +2692,9 @@ def do_convert(
         stream_names = possible_linked_stream_names(content)
         stream_name_info = mention_data.get_stream_name_map(stream_names)
 
+        linked_stream_topic_data = possible_linked_topics(content)
+        topic_info = mention_data.get_topic_info_map(linked_stream_topic_data)
+
         if content_has_emoji_syntax(content):
             active_realm_emoji = get_name_keyed_dict_for_active_realm_emoji(message_realm.id)
         else:
@@ -2683,6 +2708,7 @@ def do_convert(
             realm_url=message_realm.url,
             sent_by_bot=sent_by_bot,
             stream_names=stream_name_info,
+            topic_info=topic_info,
             translate_emoticons=translate_emoticons,
             user_upload_previews=user_upload_previews,
         )

zerver/lib/mention.py

@@ -6,6 +6,7 @@
 from django.conf import settings
 from django.db.models import Prefetch, Q
 
+from zerver.lib.topic import get_last_message_for_user_in_topic
 from zerver.lib.users import get_inaccessible_user_ids
 from zerver.models import NamedUserGroup, UserProfile
 from zerver.models.streams import get_linkable_streams
@@ -62,6 +63,19 @@ class PossibleMentions:
     message_has_stream_wildcards: bool
 
 
+@dataclass(frozen=True)
+class ChannelTopicInfo:
+    channel_name: str
+    topic_name: str
+
+
+@dataclass
+class ChannelInfo:
+    channel_id: int
+    recipient_id: int
+    history_public_to_subscribers: bool
+
+
 class MentionBackend:
     # Be careful about reuse: MentionBackend contains caches which are
     # designed to only have the lifespan of a sender user (typically a
@@ -73,7 +87,8 @@ class MentionBackend:
     def __init__(self, realm_id: int) -> None:
         self.realm_id = realm_id
         self.user_cache: dict[tuple[int, str], FullNameInfo] = {}
-        self.stream_cache: dict[str, int] = {}
+        self.stream_cache: dict[str, ChannelInfo] = {}
+        self.topic_cache: dict[ChannelTopicInfo, int] = {}
 
     def get_full_name_info_list(
         self, user_filters: list[UserFilter], message_sender: UserProfile | None
@@ -145,7 +160,7 @@ def get_stream_name_map(self, stream_names: set[str]) -> dict[str, int]:
 
         for stream_name in stream_names:
             if stream_name in self.stream_cache:
-                result[stream_name] = self.stream_cache[stream_name]
+                result[stream_name] = self.stream_cache[stream_name].channel_id
             else:
                 unseen_stream_names.append(stream_name)
 
@@ -162,15 +177,58 @@ def get_stream_name_map(self, stream_names: set[str]) -> dict[str, int]:
                 .values(
                     "id",
                     "name",
+                    "recipient_id",
+                    "history_public_to_subscribers",
                 )
             )
 
             for row in rows:
-                self.stream_cache[row["name"]] = row["id"]
+                self.stream_cache[row["name"]] = ChannelInfo(
+                    row["id"], row["recipient_id"], row["history_public_to_subscribers"]
+                )
                 result[row["name"]] = row["id"]
 
         return result
 
+    def get_topic_info_map(
+        self, channel_topic: set[ChannelTopicInfo], message_sender: UserProfile | None
+    ) -> dict[ChannelTopicInfo, int]:
+        if not channel_topic:
+            return {}
+
+        result: dict[ChannelTopicInfo, int] = {}
+        unseen_channel_topic: list[ChannelTopicInfo] = []
+
+        for channel_topic_object in channel_topic:
+            if channel_topic_object in self.topic_cache:
+                result[channel_topic_object] = self.topic_cache[channel_topic_object]
+            else:
+                unseen_channel_topic.append(channel_topic_object)
+
+        for channel_topic_object in unseen_channel_topic:
+            channel_info = self.stream_cache.get(channel_topic_object.channel_name)
+
+            assert channel_info is not None
+            recipient_id = channel_info.recipient_id
+            topic_name = channel_topic_object.topic_name
+            history_public_to_subscribers = channel_info.history_public_to_subscribers
+
+            topic_latest_message = get_last_message_for_user_in_topic(
+                self.realm_id,
+                message_sender,
+                recipient_id,
+                topic_name,
+                history_public_to_subscribers,
+            )
+
+            if topic_latest_message is None:
+                continue
+
+            self.topic_cache[channel_topic_object] = topic_latest_message
+            result[channel_topic_object] = topic_latest_message
+
+        return result
+
 
 def user_mention_matches_topic_wildcard(mention: str) -> bool:
     return mention in topic_wildcards
@@ -248,6 +306,7 @@ def __init__(
     ) -> None:
         self.mention_backend = mention_backend
         realm_id = mention_backend.realm_id
+        self.message_sender = message_sender
         mentions = possible_mentions(content)
         possible_mentions_info = get_possible_mentions_info(
             mention_backend, mentions.mention_texts, message_sender
@@ -312,6 +371,11 @@ def get_group_members(self, user_group_id: int) -> list[int]:
     def get_stream_name_map(self, stream_names: set[str]) -> dict[str, int]:
         return self.mention_backend.get_stream_name_map(stream_names)
 
+    def get_topic_info_map(
+        self, channel_topic_names: set[ChannelTopicInfo]
+    ) -> dict[ChannelTopicInfo, int]:
+        return self.mention_backend.get_topic_info_map(channel_topic_names, self.message_sender)
+
 
 def silent_mention_syntax_for_user(user_profile: UserProfile) -> str:
     return f"@_**{user_profile.full_name}|{user_profile.id}**"

zerver/lib/topic.py

@@ -76,6 +76,34 @@ def messages_for_topic(
     )
 
 
+def get_last_message_for_user_in_topic(
+    realm_id: int,
+    user_profile: UserProfile | None,
+    recipient_id: int,
+    topic_name: str,
+    history_public_to_subscribers: bool,
+) -> int | None:
+    if history_public_to_subscribers:
+        return (
+            messages_for_topic(realm_id, recipient_id, topic_name)
+            .values_list("id", flat=True)
+            .last()
+        )
+
+    elif user_profile is not None:
+        return (
+            UserMessage.objects.filter(
+                user_profile=user_profile,
+                message__recipient_id=recipient_id,
+                message__subject__iexact=topic_name,
+            )
+            .values_list("message_id", flat=True)
+            .last()
+        )
+
+    return None
+
+
 def save_message_for_edit_use_case(message: Message) -> None:
     message.save(
         update_fields=[