feat: P3 effective configuration, memory root, and revision-scoped records

harness/tests/test_effective_config_p3.py

@@ -0,0 +1,90 @@
+
+import pytest
+
+from tht.config import (
+    canonical_effective_config_json,
+    effective_config_fingerprint,
+    effective_config_identity,
+    effective_config_input_fingerprint,
+)
+
+
+def _write_cfg(tmp_path, raw):
+    import yaml as _yaml
+
+    from tht.config import load_config
+    path = tmp_path / "config.yaml"
+    path.write_text(_yaml.safe_dump(raw), encoding="utf-8")
+    cfg = load_config(path)
+    cfg._workspace_id = raw.get("workspace", {}).get("id", "psd")
+    cfg._config_source = f"workspace://{cfg._workspace_id}"
+    return cfg
+
+
+def _cfg(tmp_path, *, transport="thoth_rest", base_url="http://dwh.example.invalid", collection="psd", model="qwen3-embedding:0.6b"):
+    return {
+        "schemaVersion": 1,
+        "workspace": {"schema_version": 3, "id": "psd", "name": "PSD", "language": "it"},
+        "dwh": {
+            "type": transport,
+            "database": {"database": "warehouse", "schema": "dw"},
+            "endpoint": {"base_url": base_url, "api_key": "secret"},
+        } if transport == "thoth_rest" else {
+            "type": "postgres_direct",
+            "connection": {"host": "h", "port": 5432, "database": "warehouse", "schema": "dw", "user": "reader", "password": "secret"},
+        },
+        "vectors": {"type": "qdrant", "base_url": "http://qdrant:6333", "collection": collection, "collection_lifecycle": "self_heal"},
+        "embeddings": {"provider": "ollama_internal", "base_url": "http://embedding:11434", "model": model, "dimensions": 1024},
+        "roots": {"artifacts": str(tmp_path / "artifacts"), "indexes": str(tmp_path / "indexes")},
+        "paths": {"artifacts": str(tmp_path / "artifacts"), "indexes": str(tmp_path / "indexes"), "sessions": str(tmp_path / "sessions")},
+    }
+
+
+@pytest.fixture()
+def cfg(tmp_path):
+    return _write_cfg(tmp_path, _cfg(tmp_path))
+
+
+def test_canonical_json_is_deterministic_and_key_ordered(cfg):
+    doc = canonical_effective_config_json(cfg)
+    assert doc == canonical_effective_config_json(cfg)
+    keys = list(__import__("json").loads(doc).keys())
+    assert keys == ["schemaVersion", "dwh", "vector", "embedding", "roots"]
+
+
+def test_canonical_excludes_credentials_and_evidence(cfg):
+    doc = canonical_effective_config_json(cfg)
+    assert "secret" not in doc
+    assert "password" not in doc
+    assert "evidence" not in doc
+    assert "session_storage" not in doc
+    assert "runtime_identity" not in doc
+
+
+def test_identity_and_fingerprints_format(cfg):
+    ident = effective_config_identity("psd", cfg)
+    assert ident == "workspace://psd@v1:" + effective_config_fingerprint(cfg)[len("sha256:"):]
+    assert effective_config_input_fingerprint("psd", cfg).startswith("sha256:")
+    assert len(effective_config_fingerprint(cfg)) == len("sha256:") + 64
+
+
+def test_content_only_change_keeps_identity(tmp_path):
+    base = _write_cfg(tmp_path, _cfg(tmp_path))
+    changed = _cfg(tmp_path)
+    # A non-DWH-affecting change (collection lifecycle policy) must not alter the identity;
+    # the canonical document excludes it. Evidence changes are proven end-to-end by acceptance.
+    changed["vectors"]["collection_lifecycle"] = "require_existing"
+    changed_cfg = _write_cfg(tmp_path, changed)
+    assert effective_config_identity("psd", base) == effective_config_identity("psd", changed_cfg)
+
+
+def test_dwh_affecting_change_alters_identity(tmp_path):
+    base = _write_cfg(tmp_path, _cfg(tmp_path))
+    other = _write_cfg(tmp_path, _cfg(tmp_path, base_url="http://other.example.invalid"))
+    assert effective_config_identity("psd", base) != effective_config_identity("psd", other)
+
+
+def test_transport_change_alters_identity(tmp_path):
+    rest = _write_cfg(tmp_path, _cfg(tmp_path, transport="thoth_rest"))
+    direct = _write_cfg(tmp_path, _cfg(tmp_path, transport="postgres_direct"))
+    assert effective_config_identity("psd", rest) != effective_config_identity("psd", direct)

harness/tests/test_p3_dwh_binding.py

@@ -0,0 +1,62 @@
+
+import pytest
+
+
+def _cfg(tmp_path, **overrides):
+    base = {
+        "dwh": {
+            "type": "thoth_rest",
+            "database": {"database": "warehouse", "schema": "dw"},
+            "endpoint": {"base_url": "http://dwh.example.invalid", "api_key": "secret"},
+        },
+        "vectors": {"type": "qdrant", "base_url": "http://qdrant:6333", "collection": "psd", "collection_lifecycle": "self_heal"},
+        "embeddings": {"provider": "ollama_internal", "base_url": "http://embedding:11434", "model": "qwen3-embedding:0.6b", "dimensions": 1024},
+        "roots": {"artifacts": str(tmp_path / "artifacts"), "indexes": str(tmp_path / "indexes")},
+        "paths": {"artifacts": str(tmp_path / "artifacts"), "indexes": str(tmp_path / "indexes"), "sessions": str(tmp_path / "sessions")},
+    }
+    base.update(overrides)
+    return base
+
+
+def _write_cfg(tmp_path, raw):
+    import yaml as _yaml
+
+    from tht.config import load_config
+    path = tmp_path / "config.yaml"
+    path.write_text(_yaml.safe_dump(raw), encoding="utf-8")
+    cfg = load_config(path)
+    cfg._workspace_id = "psd"
+    cfg._config_source = "workspace://psd"
+    return cfg
+
+
+@pytest.fixture()
+def binding(tmp_path):
+    from tht.jobs.dwh_pipeline import config_dwh_binding
+    return config_dwh_binding(_write_cfg(tmp_path, _cfg(tmp_path)))
+
+
+def test_binding_has_versioned_fingerprints(binding):
+    assert binding["workspace_id"] == "psd"
+    assert binding["config_fingerprint"].startswith("sha256:")
+    assert binding["input_fingerprint"].startswith("sha256:")
+    assert len(binding["config_fingerprint"]) == 71
+
+
+def test_content_only_change_keeps_binding(tmp_path):
+    from tht.jobs.dwh_pipeline import config_dwh_binding
+    cfg1 = _write_cfg(tmp_path, _cfg(tmp_path))
+    import copy
+    raw = copy.deepcopy(_cfg(tmp_path))
+    raw["vectors"]["collection_lifecycle"] = "require_existing"
+    cfg3 = _write_cfg(tmp_path, raw)
+    assert config_dwh_binding(cfg1)["config_fingerprint"] == config_dwh_binding(cfg3)["config_fingerprint"]
+
+
+def test_endpoint_change_changes_binding(tmp_path):
+    from tht.jobs.dwh_pipeline import config_dwh_binding
+    cfg1 = _write_cfg(tmp_path, _cfg(tmp_path))
+    raw = _cfg(tmp_path)
+    raw["dwh"] = {"type": "thoth_rest", "database": {"database": "warehouse", "schema": "dw"}, "endpoint": {"base_url": "http://other.example.invalid", "api_key": "secret"}}
+    cfg2 = _write_cfg(tmp_path, raw)
+    assert config_dwh_binding(cfg1)["config_fingerprint"] != config_dwh_binding(cfg2)["config_fingerprint"]

harness/tests/test_qdrant_vector_store.py

@@ -347,7 +347,8 @@ def test_upsert_serializes_qdrant_point_payloads(record, semantic_kind):
     store.upsert("memory" if semantic_kind == "memory" else "evidence" if semantic_kind == "evidence" else "schema_records", [record])
 
     point = next(iter(fake.points.values()))
-    assert point["id"] == point_id("demo", semantic_kind, record.record.id)
+    expected_revision = "a" * 40 if semantic_kind in ("schema_table", "schema_column", "evidence") else None
+    assert point["id"] == point_id("demo", semantic_kind, record.record.id, expected_revision)
     assert point["vector"] == record.embedding
     assert point["payload"]["workspace_id"] == "demo"
     assert point["payload"]["workspace_revision"] == "a" * 40