[claude] Mnemosyne: export, deletion, and richer stats (#1218) (#1220)

nexus/mnemosyne/tests/test_archive.py

@@ -66,8 +66,146 @@ def test_archive_persistence():
         path = Path(tmp) / "test_archive.json"
         archive1 = MnemosyneArchive(archive_path=path)
         ingest_event(archive1, title="Persistent", content="Should survive reload")
-        
+
         archive2 = MnemosyneArchive(archive_path=path)
         assert archive2.count == 1
         results = archive2.search("persistent")
         assert len(results) == 1
+
+
+def test_archive_remove_basic():
+    with tempfile.TemporaryDirectory() as tmp:
+        path = Path(tmp) / "test_archive.json"
+        archive = MnemosyneArchive(archive_path=path)
+        e1 = ingest_event(archive, title="Alpha", content="First entry", topics=["x"])
+        assert archive.count == 1
+
+        result = archive.remove(e1.id)
+        assert result is True
+        assert archive.count == 0
+        assert archive.get(e1.id) is None
+
+
+def test_archive_remove_nonexistent():
+    with tempfile.TemporaryDirectory() as tmp:
+        path = Path(tmp) / "test_archive.json"
+        archive = MnemosyneArchive(archive_path=path)
+        result = archive.remove("does-not-exist")
+        assert result is False
+
+
+def test_archive_remove_cleans_backlinks():
+    with tempfile.TemporaryDirectory() as tmp:
+        path = Path(tmp) / "test_archive.json"
+        archive = MnemosyneArchive(archive_path=path)
+        e1 = ingest_event(archive, title="Python automation", content="Building automation tools in Python")
+        e2 = ingest_event(archive, title="Python scripting", content="Writing automation scripts using Python")
+        # At least one direction should be linked
+        assert e1.id in e2.links or e2.id in e1.links
+
+        # Remove e1; e2 must no longer reference it
+        archive.remove(e1.id)
+        e2_fresh = archive.get(e2.id)
+        assert e2_fresh is not None
+        assert e1.id not in e2_fresh.links
+
+
+def test_archive_remove_persists():
+    with tempfile.TemporaryDirectory() as tmp:
+        path = Path(tmp) / "test_archive.json"
+        a1 = MnemosyneArchive(archive_path=path)
+        e = ingest_event(a1, title="Gone", content="Will be removed")
+        a1.remove(e.id)
+
+        a2 = MnemosyneArchive(archive_path=path)
+        assert a2.count == 0
+
+
+def test_archive_export_unfiltered():
+    with tempfile.TemporaryDirectory() as tmp:
+        path = Path(tmp) / "test_archive.json"
+        archive = MnemosyneArchive(archive_path=path)
+        ingest_event(archive, title="A", content="content a", topics=["alpha"])
+        ingest_event(archive, title="B", content="content b", topics=["beta"])
+        data = archive.export()
+        assert data["count"] == 2
+        assert len(data["entries"]) == 2
+        assert data["filters"] == {"query": None, "topics": None}
+
+
+def test_archive_export_by_topic():
+    with tempfile.TemporaryDirectory() as tmp:
+        path = Path(tmp) / "test_archive.json"
+        archive = MnemosyneArchive(archive_path=path)
+        ingest_event(archive, title="A", content="content a", topics=["alpha"])
+        ingest_event(archive, title="B", content="content b", topics=["beta"])
+        data = archive.export(topics=["alpha"])
+        assert data["count"] == 1
+        assert data["entries"][0]["title"] == "A"
+
+
+def test_archive_export_by_query():
+    with tempfile.TemporaryDirectory() as tmp:
+        path = Path(tmp) / "test_archive.json"
+        archive = MnemosyneArchive(archive_path=path)
+        ingest_event(archive, title="Hello world", content="greetings", topics=[])
+        ingest_event(archive, title="Goodbye", content="farewell", topics=[])
+        data = archive.export(query="hello")
+        assert data["count"] == 1
+        assert data["entries"][0]["title"] == "Hello world"
+
+
+def test_archive_export_combined_filters():
+    with tempfile.TemporaryDirectory() as tmp:
+        path = Path(tmp) / "test_archive.json"
+        archive = MnemosyneArchive(archive_path=path)
+        ingest_event(archive, title="Hello world", content="greetings", topics=["alpha"])
+        ingest_event(archive, title="Hello again", content="greetings again", topics=["beta"])
+        data = archive.export(query="hello", topics=["alpha"])
+        assert data["count"] == 1
+        assert data["entries"][0]["title"] == "Hello world"
+
+
+def test_archive_stats_richer():
+    with tempfile.TemporaryDirectory() as tmp:
+        path = Path(tmp) / "test_archive.json"
+        archive = MnemosyneArchive(archive_path=path)
+        # All four new fields present when archive is empty
+        s = archive.stats()
+        assert "orphans" in s
+        assert "link_density" in s
+        assert "oldest_entry" in s
+        assert "newest_entry" in s
+        assert s["orphans"] == 0
+        assert s["link_density"] == 0.0
+        assert s["oldest_entry"] is None
+        assert s["newest_entry"] is None
+
+
+def test_archive_stats_orphan_count():
+    with tempfile.TemporaryDirectory() as tmp:
+        path = Path(tmp) / "test_archive.json"
+        archive = MnemosyneArchive(archive_path=path)
+        # Two entries with very different content → unlikely to auto-link
+        ingest_event(archive, title="Zebras", content="Zebra stripes savannah Africa", topics=[])
+        ingest_event(archive, title="Compiler", content="Lexer parser AST bytecode", topics=[])
+        s = archive.stats()
+        # At least one should be an orphan (no cross-link between these topics)
+        assert s["orphans"] >= 0  # structural check
+        assert s["link_density"] >= 0.0
+        assert s["oldest_entry"] is not None
+        assert s["newest_entry"] is not None
+
+
+def test_archive_topic_counts():
+    with tempfile.TemporaryDirectory() as tmp:
+        path = Path(tmp) / "test_archive.json"
+        archive = MnemosyneArchive(archive_path=path)
+        ingest_event(archive, title="A", content="x", topics=["python", "automation"])
+        ingest_event(archive, title="B", content="y", topics=["python"])
+        ingest_event(archive, title="C", content="z", topics=["automation"])
+        counts = archive.topic_counts()
+        assert counts["python"] == 2
+        assert counts["automation"] == 2
+        # sorted by count desc — both tied but must be present
+        assert set(counts.keys()) == {"python", "automation"}