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

2026-08-12 16:01:25 +02:00
parent beaba548c1
commit e23e526966
21 changed files with 1108 additions and 44 deletions
@@ -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"]