D79 Wave 2: bottom-up summaries + placement (#165) - #170
Merged
Conversation
Coverage reportClick to see where and how coverage changed
This report was generated by python-coverage-comment-action |
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Wave 2 of 3 for #165 — bottom-up section summaries and the placement hint, per
plan/designs/e0_files_design.md§4.1.What ships
REMEMBERSTACK_SUMMARY_MODEL(defaultz-ai/glm-4.7-flash): one-line summaries per section — leaves read their own direct blocks (oversized leaves shard at block grain), parents read their own blocks + child one-liners with balanced fan-in, the root reduction is one call with two fields (document summary + the D39 placement hint). Ceiling-driven reduction down to pairs; titles capped at every render site; a whitespace token estimate plus a hard character ceiling on every request; strict closed schemas with post-parse normalization (nomax_length/patternkeywords that strict providers could 400 on).Review trail
Codex implemented; Grok-4.5 (2 MAJOR / 5 MINOR) and Opus (6 MAJOR / 5 MINOR) reviewed independently — Opus reproduced three degradations with measurements (a 3000-word heading nulling a document with zero calls; 208KB of CJK passing a 58-"token" count; count-driven reduction giving up). The consolidated fix round (Codex) addressed every finding; Grok re-reviewed the fix diff: APPROVE, confirming all seven majors non-vacuously fixed. Final review and adjudication by Claude (the input-identity seed formulation and the §2.4 schema strip were the final-reviewer calls).
Gate
ruff clean, pyright 0 errors, 1003 passed including the Postgres-gated suites. No migration. Benchmarks and the tool catalog untouched.