ZacharyZhang-NY
71733d86d6
A review of the p2-1 record found seven places where the prose had drifted away from the reports it cites. None were product defects; all were record keeping, and record keeping is what a memory system is judged on. The quantization row mis-transcribed nine latencies, and the two worst made dimension truncation look twice as fast as the full scan when the report measures six percent. The extrapolation table scaled latency by atoms and bytes by vectors under one heading that named neither switch. The picture index stats existed only in prose, so the harness now reads the index shape off the artifact and times the encoder on a sample, which reproduces the hand-timed build within seven percent. The dense arms named a store that holds no vectors, so the report names the store the vectors came from and refuses to run where a dense leg would rank an empty index. The frozen baseline was migrated in place by the product CLI while the picture index was built, so its pinned mtime, digest and schema version are re-pinned and the step that moved it is named. The sqlite_vec import added two mypy errors; the fix is the missing stub declaration, not an inline ignore, and the count is back to the recorded baseline of twenty-four.
81 lines
3.2 KiB
Python
81 lines
3.2 KiB
Python
"""What this machine can actually perceive, computed once and read by status and doctor alike."""
|
|
|
|
from __future__ import annotations
|
|
|
|
from pathlib import Path
|
|
|
|
from .config import Settings
|
|
from .db import Database
|
|
from .status_schemas import DenseState, SenseState, SenseStatus, VisualState
|
|
from .vector_index import VectorIndex, expected_vectors
|
|
from .visual_index import VisualIndex
|
|
|
|
|
|
def sense_facts(settings: Settings, database: Database, visual: VisualIndex | None = None) -> SenseStatus:
|
|
"""Report the senses as facts, so the UI and the doctor cannot disagree about what is missing."""
|
|
index = visual or VisualIndex(settings, database)
|
|
claimed = _picture_claim(database)
|
|
on_disk = index.visual_atom_ids()
|
|
dense, covered, total = _dense_facts(settings, database)
|
|
return SenseStatus(
|
|
dense=dense,
|
|
dense_covered=covered,
|
|
dense_expected=total,
|
|
picture_atoms=claimed,
|
|
picture_bytes_missing=max(0, claimed - len(on_disk)),
|
|
visual=_visual_state(index, on_disk),
|
|
audio_atoms=_audio_count(database),
|
|
speech=_sense_state(settings.asr_model_dir),
|
|
captioning=_sense_state(settings.vlm_model_dir),
|
|
clip_path=_stated(settings.clip_model_dir),
|
|
asr_path=_stated(settings.asr_model_dir),
|
|
vlm_path=_stated(settings.vlm_model_dir),
|
|
)
|
|
|
|
|
|
def _dense_facts(settings: Settings, database: Database) -> tuple[DenseState, int, int]:
|
|
expected = expected_vectors(database)
|
|
if not settings.dense_embeddings_enabled:
|
|
return "off", 0, expected
|
|
facts = VectorIndex(settings, database).stats()
|
|
if not facts["built"]:
|
|
return "missing", 0, expected
|
|
# A wordless memory belongs to the picture index, so the denominator is what this index holds, not every row.
|
|
return ("current" if facts["pending"] == 0 else "behind"), int(facts["vectors"]), expected
|
|
|
|
|
|
def _visual_state(visual: VisualIndex, pictures: set[str]) -> VisualState:
|
|
"""A store of pictures the index has not encoded yet is stale: they are invisible to search."""
|
|
if not visual.available():
|
|
return "absent"
|
|
if not pictures:
|
|
return "ready"
|
|
loaded = visual.load()
|
|
return "ready" if loaded is not None and set(loaded[0]) == pictures else "stale"
|
|
|
|
|
|
def _picture_claim(database: Database) -> int:
|
|
"""Atoms that say they hold a picture, whether or not the bytes survived."""
|
|
with database.connect() as connection:
|
|
row = connection.execute(
|
|
"SELECT COUNT(*) FROM v_atoms_current WHERE image_ref IS NOT NULL AND image_ref != ''"
|
|
).fetchone()
|
|
return int(row[0])
|
|
|
|
|
|
def _sense_state(model_dir: Path | None) -> SenseState:
|
|
"""A sense with no configured snapshot is absent; one pointing at a deleted directory is missing."""
|
|
if model_dir is None:
|
|
return "absent"
|
|
return "ready" if model_dir.exists() else "missing"
|
|
|
|
|
|
def _stated(model_dir: Path | None) -> str | None:
|
|
return None if model_dir is None else str(model_dir)
|
|
|
|
|
|
def _audio_count(database: Database) -> int:
|
|
with database.connect() as connection:
|
|
row = connection.execute("SELECT COUNT(*) AS n FROM v_atoms_current WHERE modality = 'audio'").fetchone()
|
|
return int(row["n"])
|