Productize section heading contextualization - #501
Open
GiulioDER wants to merge 1 commit into
Open
Conversation
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
2 participants
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Summary
Add consistent selection of the existing �ge-small-context-section-v1 profile across CLI, MCP, and generation paths.
The implementation preserves raw evidence text, chunk identity, offsets, provenance, trust metadata, and citations while adding deterministic heading context only to embedding passages. Profile identity and cache keys prevent stale or aliased vectors.
Included
Validation
The focused local suite passed 237 tests with 53 skipped on 2026-08-24. Five PostgreSQL lifecycle tests were unavailable because the configured local test database endpoint was not running. VPS2 profile preflight passed, and the shadow regeneration is running in an isolated staging database.
The default embedding profile remains unchanged. Production promotion still requires independent calibration, parity checks, and the preregistered retrieval evaluation.