Rank memory search in SQL instead of rebuilding BM25 in Python

The product path pulled every atom into Python and rebuilt a BM25Okapi
index per query: p50 20.16s over 165,973 atoms. It now runs one SQLite
statement with LIMIT/OFFSET.

- text_index.py: one normalizer for both sides of the index. CJK runs are
  split per character (unicode61 cannot segment them), camelCase and
  separators are split so "search atoms" finds SearchAtoms, latin tokens
  get a prefix match. Rows written before this existed hold raw text in
  atom_fts, so schema v3 rebuilds the index on upgrade.
- retrieval.py: filters (holder / scenario / source / modality / since /
  until / min_trust), a base64 cursor, and reciprocal rank fusion in place
  of the 0.55/0.35/0.10 blend whose graph term was always zero.
- An empty result is {status: "no_match"}, not a bare array.
- Go: search_atoms and compile_context rank through POST /api/search. The
  second Go tokenizer is deleted; two implementations could drift, which
  is what the MCP/HTTP parity requirement forbids.
- db.py passed 500 lines, so the row codec moved to row_codec.py.
- Tooling: installed the repo's own pyright. pyrightconfig.json had no
  exclude and swept the downloaded corpus (1171 errors); with excludes the
  count is 62, all pre-existing (evals 48, profiles 4, training 4,
  ingestion 3, runtime 2, cartridge_assets 1). backend/pyproject.toml
  carried a second [tool.pyright] that shadowed the root config; deleted.
- bench/scale_check.py measures the product search at corpus size. The
  100k-atom p50 number is not written until it is measured on an idle
  machine.

Verification: 112 pytest passed; go build/vet/test clean, gofmt clean;
ruff clean on every touched file; mypy clean on the 14 touched Python
files; repo pyright 62 errors, unchanged per-file ledger.
This commit is contained in:
ZacharyZhang-NY committed 2026-10-07 16:02:40 -04:00
1 parent 3d792417cf
commit 674c88e84e
29 files changed
+1097 -356

No files matched your search

+25 -7
View File
@@ -12,7 +12,9 @@ from __future__ import annotations
import sqlite3
SCHEMA_VERSION = 2
from .text_index import index_text
SCHEMA_VERSION = 3
ATOM_COLUMNS_V2 = """
atom_id TEXT PRIMARY KEY,
@@ -85,12 +87,14 @@ _BASE_TABLES = """
metadata_json TEXT NOT NULL
);
CREATE VIRTUAL TABLE IF NOT EXISTS atom_fts USING fts5(
atom_id UNINDEXED,
source,
text,
tokenize = 'unicode61'
);
CREATE VIRTUAL TABLE IF NOT EXISTS atom_fts USING fts5(
atom_id UNINDEXED,
source,
text,
tokenize = 'unicode61'
);
-- source and text hold text_index output, not the atom's own text: unicode61 sees one token
-- per CJK character and per identifier fragment, which is what makes those queries match.
"""
# A memory is usable now when nothing has closed it, it had begun by now, and its
@@ -178,6 +182,20 @@ def apply(connection: sqlite3.Connection) -> None:
_upgrade_v1_to_v2(connection)
else:
_create_v2(connection)
if version < 3:
_rebuild_fts(connection)
connection.executescript(VIEWS)
connection.execute(f"PRAGMA user_version = {SCHEMA_VERSION}")
def _rebuild_fts(connection: sqlite3.Connection) -> None:
"""Re-index every atom through text_index: rows indexed before it existed hold raw text, and
FTS5 cannot match 钥匙 inside a token that is a whole Chinese sentence."""
connection.execute("DROP TABLE IF EXISTS atom_fts")
connection.executescript(_BASE_TABLES)
rows = connection.execute("SELECT atom_id, source, text FROM atoms ORDER BY atom_id").fetchall()
connection.executemany(
"INSERT INTO atom_fts(atom_id, source, text) VALUES (?, ?, ?)",
[(row[0], index_text(row[1]), index_text(row[2])) for row in rows],
)