perf(persistence): Speed up hydrating a page of cached messages

[VelikovPetar]

Submit a pull request

Linear: FLU-

Github Issue: #

CLA

• I have signed the Stream CLA (required).
• The code changes follow best practices
• Code changes are tested (add some information if not applicable)

Description of the pull request

Two behavior-preserving optimizations to message-page hydration in MessageDao and PinnedMessageDao (the read path behind getMessagesByCid, thread loads, and getMessageById). Both cut redundant SQLite work when turning a page of cached rows into Message objects.

1. Derive own reactions in memory instead of a second query.
Hydration previously fired two reaction queries per page: one for all reactions (getReactionsForMessages) and a second for the current user's subset (getReactionsForMessagesByUserId). The current user's reactions are already contained in the first result, so we now filter them in memory (_ownReactionsFrom) and drop the second query entirely. Reaction.userId maps directly from the reactions.userId column, so the in-memory filter is exactly equivalent to the old SQL predicate, and iteration order (createdAt ASC) is preserved.

2. Don't re-fetch quoted messages already in the page.
Quoted-message previews were always fetched and re-hydrated via a separate query, even when the quoted message was already one of the rows on the current page. _resolveQuotedMessages now rebuilds in-page quotes from the maps already loaded for the page (reactions, poll, user, location) and only queries the DB for quotes that fall outside the page. In-page reuse is gated on fetchSharedLocation == true because a quote must always carry its shared location — when the page didn't load locations, those quotes fall through to the DB branch (which always fetches locations), exactly matching the previous behavior. Quotes remain hydrated a single level deep and never carry a draft.

Correctness

• No behavior change: own-reactions filtering, single-level quote hydration, no-draft-on-quotes, and "a quote always keeps its shared location even when the parent page opts out" are all preserved.
• getReactionsForMessagesByUserId is left in place (still public, still directly tested); it's just no longer part of the hydration path.

Testing

• Existing message_dao / pinned_message_dao suites pass (93 tests), including the quote/reaction/location edge cases (hydrates sharedLocation for quoted message even when fetchSharedLocation=false on the parent, hydrates quotes to a single level only, does not hydrate drafts for quoted messages, hydrates multiple latest and own reactions).
• dart analyze is clean on both changed files.
• Micro-benchmark (file-backed SQLite, 400 iterations, old 2-query path vs. new 1-query + filter path) confirmed the reaction change is never slower and the gain grows with page size: ~1.22x on a typical page (30 msg × 6 reactors), and ~0.9 ms saved per page on a large one (100 msg × 12 reactors). The win widens further on slower mobile storage and reduces contention on the single main-isolate SQLite connection.

Screenshots / Videos

No UI changes.

Summary by CodeRabbit

• Performance
  • Improved cached message-page hydration by reducing redundant reaction and quoted-message lookups.
  • Faster loading for message and pinned-message lists, including quoted-message previews.
  • Preserved shared-location details while avoiding unnecessary refetching of quoted messages already in the loaded page.

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro

Run ID: aecefcff-b577-4b0c-b415-98ce94271792

📥 Commits

Reviewing files that changed from the base of the PR and between 5f7f33c and eb15e86.

📒 Files selected for processing (3)

• packages/stream_chat_persistence/CHANGELOG.md
• packages/stream_chat_persistence/lib/src/dao/message_dao.dart
• packages/stream_chat_persistence/lib/src/dao/pinned_message_dao.dart

---

📝 Walkthrough

Walkthrough

Cached message and pinned-message hydration now fetches reactions once per batch, derives current-user reactions in memory, and reuses loaded page messages when resolving quoted previews. Out-of-page quotes continue to load with shared locations.

Changes

Message hydration optimization

Layer / File(s)	Summary
Batch reaction and quote hydration packages/stream_chat_persistence/lib/src/dao/message_dao.dart	MessageDao derives own reactions from fetched reactions and resolves quoted messages through page reuse or out-of-page loading.
Pinned message hydration packages/stream_chat_persistence/lib/src/dao/pinned_message_dao.dart	PinnedMessageDao applies the same reaction and quoted-message hydration strategy.
Performance changelog packages/stream_chat_persistence/CHANGELOG.md	Adds a note describing the faster cached hydration behavior.

Estimated code review effort: 3 (Moderate) | ~25 minutes

Possibly related PRs

• GetStream/stream-chat-flutter#2681: Refactors the same DAO hydration pipeline with batched reaction fetching and cursor/page hydration changes.
• GetStream/stream-chat-flutter#2731: Applies related reaction reuse and quoted-message hydration changes in both message DAOs.

Suggested reviewers: xsahil03x, renefloor

🚥 Pre-merge checks | ✅ 5

✅ Passed checks (5 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title clearly summarizes the main performance change to cached message-page hydration in persistence.
Docstring Coverage	✅ Passed	No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch feature/optimize_reactions_and_quoted_messages_hydration

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 71.28%. Comparing base (5f7f33c) to head (eb15e86).

Additional details and impacted files

@@           Coverage Diff           @@
##           master    #2820   +/-   ##
=======================================
  Coverage   71.28%   71.28%           
=======================================
  Files         430      430           
  Lines       26934    26967   +33     
=======================================
+ Hits        19199    19223   +24     
- Misses       7735     7744    +9     

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.