user2595/codex-py
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

fix: audit remediation Wave 3 — M-1, MED/LOW sweep, ingest robustness + D-1 close-out #14

Merged
user2595 merged 15 commits from fix/audit-wave-3 into main 2026-06-16 04:54:00 +00:00
Conversation 0 Commits 15 Files Changed 25 +8 -2
2 changed files with 8 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files
Showing only changes of commit 45c3b16dd7 - Show all commits

3
codex/cli.py
View File

@@ -102,7 +102,8 @@ def search_paper(
cite_boost: bool = typer.Option( cite_boost: bool = typer.Option(
False, False,
"--cite-boost", "--cite-boost",
help="Weight results by citation PageRank (F-15). Graceful when corpus < 5 papers.", help="Re-rank the top results by citation PageRank — a within-page "
"tie-breaker, not a hard re-ranking (F-15). Graceful when corpus < 5 papers.",
), ),
) -> None: ) -> None:
"""Semantic similarity search over paper abstracts.""" """Semantic similarity search over paper abstracts."""

7
codex/embed.py
View File

@@ -1,4 +1,4 @@
"""Hybrid dense + sparse embeddings via BGE-M3 (ADR-0002). """Dense (+ optional sparse) embeddings via BGE-M3 (ADR-0002).
Uses :class:`FlagEmbedding.BGEM3FlagModel` for encoding because the Uses :class:`FlagEmbedding.BGEM3FlagModel` for encoding because the
``return_dense`` / ``return_sparse`` kwargs are part of the FlagEmbedding ``return_dense`` / ``return_sparse`` kwargs are part of the FlagEmbedding
@@ -7,6 +7,11 @@ vanilla ``SentenceTransformer`` load of ``BAAI/bge-m3`` (same weights,
same model); sparse output is a list of ``{token_id: weight}`` dicts per same model); sparse output is a list of ``{token_id: weight}`` dicts per
text. text.
Status: only the **dense** path is wired into ingest/search today; the search
"hybrid" is dense + Postgres FTS. :meth:`Embedder.encode_sparse` / :meth:`encode`
are provided for a future sparse-retrieval layer but are not yet consumed by the
pipeline (audit C-14).
Notes for callers Notes for callers
----------------- -----------------
* Empty input is handled explicitly — no model call is issued. * Empty input is handled explicitly — no model call is issued.