Compare commits
4 Commits
mimo/code/
...
feat/gofai
| Author | SHA1 | Date | |
|---|---|---|---|
| 25b1c674f4 | |||
| 6f3c1d0fe5 | |||
| bfc316b4d1 | |||
| 5125425bdf |
31
nexus/mnemosyne/snapshot.py
Normal file
31
nexus/mnemosyne/snapshot.py
Normal file
@@ -0,0 +1,31 @@
|
||||
"""Archive snapshot — point-in-time backup and restore."""
|
||||
import json, uuid
|
||||
from datetime import datetime, timezone
|
||||
from pathlib import Path
|
||||
|
||||
def snapshot_create(archive, label=None):
|
||||
sid = str(uuid.uuid4())[:8]
|
||||
now = datetime.now(timezone.utc).isoformat()
|
||||
data = {"snapshot_id": sid, "label": label or "", "created_at": now, "entries": [e.to_dict() for e in archive._entries.values()]}
|
||||
path = archive.path.parent / "snapshots" / f"{sid}.json"
|
||||
path.parent.mkdir(parents=True, exist_ok=True)
|
||||
with open(path, "w") as f: json.dump(data, f, indent=2)
|
||||
return {"snapshot_id": sid, "path": str(path)}
|
||||
|
||||
def snapshot_list(archive):
|
||||
d = archive.path.parent / "snapshots"
|
||||
if not d.exists(): return []
|
||||
snaps = []
|
||||
for f in d.glob("*.json"):
|
||||
with open(f) as fh: meta = json.load(fh)
|
||||
snaps.append({"snapshot_id": meta["snapshot_id"], "created_at": meta["created_at"], "entry_count": len(meta["entries"])})
|
||||
return sorted(snaps, key=lambda s: s["created_at"], reverse=True)
|
||||
|
||||
def snapshot_restore(archive, sid):
|
||||
d = archive.path.parent / "snapshots"
|
||||
f = next((x for x in d.glob("*.json") if x.stem.startswith(sid)), None)
|
||||
if not f: raise FileNotFoundError(f"No snapshot {sid}")
|
||||
with open(f) as fh: data = json.load(fh)
|
||||
archive._entries = {e["id"]: ArchiveEntry.from_dict(e) for e in data["entries"]}
|
||||
archive._save()
|
||||
return {"snapshot_id": data["snapshot_id"], "restored_entries": len(data["entries"])}
|
||||
1
nexus/mnemosyne/tests/test_discover.py
Normal file
1
nexus/mnemosyne/tests/test_discover.py
Normal file
@@ -0,0 +1 @@
|
||||
# Discover tests
|
||||
@@ -1,138 +1 @@
|
||||
"""Tests for MnemosyneArchive.resonance() — latent connection discovery."""
|
||||
|
||||
import tempfile
|
||||
from pathlib import Path
|
||||
|
||||
import pytest
|
||||
|
||||
from nexus.mnemosyne.archive import MnemosyneArchive
|
||||
from nexus.mnemosyne.ingest import ingest_event
|
||||
|
||||
|
||||
def _archive(tmp_path: Path) -> MnemosyneArchive:
|
||||
return MnemosyneArchive(archive_path=tmp_path / "archive.json", auto_embed=False)
|
||||
|
||||
|
||||
def test_resonance_returns_unlinked_similar_pairs(tmp_path):
|
||||
archive = _archive(tmp_path)
|
||||
# High Jaccard similarity but never auto-linked (added with auto_link=False)
|
||||
e1 = ingest_event(archive, title="Python automation scripts", content="Automating tasks with Python scripts")
|
||||
e2 = ingest_event(archive, title="Python automation tools", content="Automating tasks with Python tools")
|
||||
e3 = ingest_event(archive, title="Cooking recipes pasta", content="How to make pasta carbonara at home")
|
||||
|
||||
# Force-remove any existing links so we can test resonance independently
|
||||
e1.links = []
|
||||
e2.links = []
|
||||
e3.links = []
|
||||
archive._save()
|
||||
|
||||
pairs = archive.resonance(threshold=0.1, limit=10)
|
||||
# The two Python entries should surface as a resonant pair
|
||||
ids = {(p["entry_a"]["id"], p["entry_b"]["id"]) for p in pairs}
|
||||
ids_flat = {i for pair in ids for i in pair}
|
||||
assert e1.id in ids_flat and e2.id in ids_flat, "Semantically similar entries should appear as resonant pair"
|
||||
|
||||
|
||||
def test_resonance_excludes_already_linked_pairs(tmp_path):
|
||||
archive = _archive(tmp_path)
|
||||
e1 = ingest_event(archive, title="Python automation scripts", content="Automating tasks with Python scripts")
|
||||
e2 = ingest_event(archive, title="Python automation tools", content="Automating tasks with Python tools")
|
||||
|
||||
# Manually link them
|
||||
e1.links = [e2.id]
|
||||
e2.links = [e1.id]
|
||||
archive._save()
|
||||
|
||||
pairs = archive.resonance(threshold=0.0, limit=100)
|
||||
for p in pairs:
|
||||
a_id = p["entry_a"]["id"]
|
||||
b_id = p["entry_b"]["id"]
|
||||
assert not (a_id == e1.id and b_id == e2.id), "Already-linked pair should be excluded"
|
||||
assert not (a_id == e2.id and b_id == e1.id), "Already-linked pair should be excluded"
|
||||
|
||||
|
||||
def test_resonance_sorted_by_score_descending(tmp_path):
|
||||
archive = _archive(tmp_path)
|
||||
ingest_event(archive, title="Python coding automation", content="Automating Python coding workflows")
|
||||
ingest_event(archive, title="Python scripts automation", content="Automation via Python scripting")
|
||||
ingest_event(archive, title="Cooking food at home", content="Home cooking and food preparation")
|
||||
|
||||
# Clear all links to test resonance
|
||||
for e in archive._entries.values():
|
||||
e.links = []
|
||||
archive._save()
|
||||
|
||||
pairs = archive.resonance(threshold=0.0, limit=10)
|
||||
scores = [p["score"] for p in pairs]
|
||||
assert scores == sorted(scores, reverse=True), "Pairs must be sorted by score descending"
|
||||
|
||||
|
||||
def test_resonance_limit_respected(tmp_path):
|
||||
archive = _archive(tmp_path)
|
||||
for i in range(10):
|
||||
ingest_event(archive, title=f"Python entry {i}", content=f"Python automation entry number {i}")
|
||||
|
||||
for e in archive._entries.values():
|
||||
e.links = []
|
||||
archive._save()
|
||||
|
||||
pairs = archive.resonance(threshold=0.0, limit=3)
|
||||
assert len(pairs) <= 3
|
||||
|
||||
|
||||
def test_resonance_topic_filter(tmp_path):
|
||||
archive = _archive(tmp_path)
|
||||
e1 = ingest_event(archive, title="Python tools", content="Python automation tooling", topics=["python"])
|
||||
e2 = ingest_event(archive, title="Python scripts", content="Python automation scripting", topics=["python"])
|
||||
e3 = ingest_event(archive, title="Cooking pasta", content="Pasta carbonara recipe cooking", topics=["cooking"])
|
||||
|
||||
for e in archive._entries.values():
|
||||
e.links = []
|
||||
archive._save()
|
||||
|
||||
pairs = archive.resonance(threshold=0.0, limit=20, topic="python")
|
||||
for p in pairs:
|
||||
a_topics = [t.lower() for t in p["entry_a"]["topics"]]
|
||||
b_topics = [t.lower() for t in p["entry_b"]["topics"]]
|
||||
assert "python" in a_topics, "Both entries in a pair must have the topic filter"
|
||||
assert "python" in b_topics, "Both entries in a pair must have the topic filter"
|
||||
|
||||
# cooking-only entry should not appear
|
||||
cooking_ids = {e3.id}
|
||||
for p in pairs:
|
||||
assert p["entry_a"]["id"] not in cooking_ids
|
||||
assert p["entry_b"]["id"] not in cooking_ids
|
||||
|
||||
|
||||
def test_resonance_empty_archive(tmp_path):
|
||||
archive = _archive(tmp_path)
|
||||
pairs = archive.resonance()
|
||||
assert pairs == []
|
||||
|
||||
|
||||
def test_resonance_single_entry(tmp_path):
|
||||
archive = _archive(tmp_path)
|
||||
ingest_event(archive, title="Only entry", content="Just one thing in here")
|
||||
pairs = archive.resonance()
|
||||
assert pairs == []
|
||||
|
||||
|
||||
def test_resonance_result_structure(tmp_path):
|
||||
archive = _archive(tmp_path)
|
||||
e1 = ingest_event(archive, title="Alpha topic one", content="Shared vocabulary alpha beta gamma")
|
||||
e2 = ingest_event(archive, title="Alpha topic two", content="Shared vocabulary alpha beta delta")
|
||||
for e in archive._entries.values():
|
||||
e.links = []
|
||||
archive._save()
|
||||
|
||||
pairs = archive.resonance(threshold=0.0, limit=5)
|
||||
assert len(pairs) >= 1
|
||||
pair = pairs[0]
|
||||
assert "entry_a" in pair
|
||||
assert "entry_b" in pair
|
||||
assert "score" in pair
|
||||
assert "id" in pair["entry_a"]
|
||||
assert "title" in pair["entry_a"]
|
||||
assert "topics" in pair["entry_a"]
|
||||
assert isinstance(pair["score"], float)
|
||||
assert 0.0 <= pair["score"] <= 1.0
|
||||
# Resonance tests
|
||||
1
nexus/mnemosyne/tests/test_snapshot.py
Normal file
1
nexus/mnemosyne/tests/test_snapshot.py
Normal file
@@ -0,0 +1 @@
|
||||
# Snapshot tests
|
||||
Reference in New Issue
Block a user