feat: index semantic records in qdrant

harness/tests/test_search_pack.py

@@ -1,5 +1,5 @@
 import json
-from datetime import datetime
+from datetime import UTC, datetime
 from types import SimpleNamespace
 
 from typer.testing import CliRunner
@@ -7,8 +7,8 @@ from typer.testing import CliRunner
 from tht.cli import app
 from tht.config import load_config
 from tht.jobs.dwh_pipeline import DwhPreprocessPipeline, config_dwh_binding
-from tht.ports.vector import VectorReadUnavailable
 from tht.mschema.models import ColumnPhysical, PhysicalSchema, TablePhysical
+from tht.ports.vector import VectorReadUnavailable
 from tht.vectorstore.embeddings import EmbeddingsError
 
 
@@ -22,7 +22,11 @@ class _FakeEmbedder:
 
 
 class _FakeSearcher:
+    def __init__(self):
+        self.calls = []
+
     def search(self, vec, top_n, kinds=None):
+        self.calls.append({"top_n": top_n, "kinds": kinds})
         if kinds == ["solved_question"]:
             return [SimpleNamespace(
                 kind="memory", ref="s-1", id="m1", title="q solved",
@@ -47,7 +51,7 @@ class _FakeSearcher:
 
 def _workspace(tmp_path, with_session=None):
     physical = PhysicalSchema(
-        database="d", schema="s", introspected_at=datetime(2026, 1, 1),
+        database="d", schema="s", introspected_at=datetime(2026, 1, 1, tzinfo=UTC),
         tables={"fact_ablazione": TablePhysical(
             comment="Ablazioni", columns={"cod_paz": ColumnPhysical(type="bigint")})},
     )
@@ -55,7 +59,7 @@ def _workspace(tmp_path, with_session=None):
     cfg.write_text(
         "database: {database: d, schema: s, user: u, password: p, transport: direct}\n"
         "vector_db: {database: v, schema: public, user: u, password: p}\n"
-        "embeddings: {base_url: 'http://localhost:11434', model: nomic-embed-text, dim: 8}\n"
+        "embeddings: {base_url: 'http://localhost:11434', model: qwen3-embedding:0.6b, dim: 1024}\n"
         f"paths: {{artifacts: {tmp_path/'artifacts'}, indexes: {tmp_path/'i'}, "
         f"sessions: {tmp_path/'sessions'}}}\n"
     )
@@ -91,10 +95,15 @@ def _patch(monkeypatch, embedder, searcher):
 def test_pack_single_embed_and_sections(tmp_path, monkeypatch):
     cfg = _workspace(tmp_path)
     emb = _FakeEmbedder()
-    _patch(monkeypatch, emb, _FakeSearcher())
+    searcher = _FakeSearcher()
+    _patch(monkeypatch, emb, searcher)
     res = CliRunner().invoke(app, ["search", "pack", "quanti pazienti", "-c", str(cfg)])
     assert res.exit_code == 0, res.output
     assert emb.calls == 1  # UN solo embedding per le tre ricerche
+    assert [call["kinds"] for call in searcher.calls] == [
+        ["schema_table", "schema_column"],
+        ["solved_question"],
+    ]
     assert "fact_ablazione" in res.output and "Ablazioni" in res.output
     # Evidence is fail-closed until an ACTIVE corpus exists; legacy vector rows
     # must not leak into a new search pack.