refactor(ai): OCR sidecar canonical naming cleanup — typhoon→np-dms, remove hardcoded keys, asyncio.to_thread, ADR-040/041

tests/unit/ocr-sidecar/test_api_key_validation.py

@@ -0,0 +1,42 @@
+# File: tests/unit/ocr-sidecar/test_api_key_validation.py
+# Change Log:
+# - 2026-06-20: Added ADR-040 API key startup and request validation tests.
+
+import importlib
+import os
+import sys
+from pathlib import Path
+
+import pytest
+from fastapi.testclient import TestClient
+
+from test_path_traversal import SIDECAR_DIR, install_import_stubs, load_app
+
+
+def test_sidecar_fails_fast_when_api_key_missing(monkeypatch: pytest.MonkeyPatch, tmp_path: Path) -> None:
+    install_import_stubs()
+    monkeypatch.delenv("OCR_SIDECAR_API_KEY", raising=False)
+    monkeypatch.setenv("OCR_SIDECAR_UPLOAD_BASE", str(tmp_path))
+    if str(SIDECAR_DIR) not in sys.path:
+        sys.path.insert(0, str(SIDECAR_DIR))
+    sys.modules.pop("app", None)
+    with pytest.raises(RuntimeError, match="OCR_SIDECAR_API_KEY is required"):
+        importlib.import_module("app")
+
+
+def test_sidecar_rejects_invalid_api_key(tmp_path: Path) -> None:
+    app_module = load_app(tmp_path)
+    client = TestClient(app_module.app)
+    response = client.post(
+        "/embed",
+        json={"text": "hello"},
+        headers={"X-API-Key": "wrong-key"},
+    )
+    assert response.status_code == 401
+
+
+def test_sidecar_rejects_missing_api_key(tmp_path: Path) -> None:
+    app_module = load_app(tmp_path)
+    client = TestClient(app_module.app)
+    response = client.post("/embed", json={"text": "hello"})
+    assert response.status_code == 401

tests/unit/ocr-sidecar/test_path_traversal.py

@@ -0,0 +1,114 @@
+# File: tests/unit/ocr-sidecar/test_path_traversal.py
+# Change Log:
+# - 2026-06-20: Added ADR-040 path traversal tests for OCR sidecar.
+
+import importlib
+import os
+import sys
+import types
+from pathlib import Path
+from unittest.mock import patch
+
+from fastapi.testclient import TestClient
+
+SIDECAR_DIR = Path(__file__).resolve().parents[3] / "specs" / "04-Infrastructure-OPS" / "04-00-docker-compose" / "Desk-5439" / "ocr-sidecar"
+
+
+def install_import_stubs() -> None:
+    """ติดตั้ง stub สำหรับ dependency หนักเพื่อให้ unit test import app ได้เร็ว"""
+    fitz_module = types.ModuleType("fitz")
+    fitz_module.Document = object
+    fitz_module.open = lambda *args, **kwargs: None
+    sys.modules["fitz"] = fitz_module
+    typhoon_module = types.ModuleType("typhoon_ocr")
+    typhoon_module.prepare_ocr_messages = lambda *args, **kwargs: [{"content": []}]
+    sys.modules["typhoon_ocr"] = typhoon_module
+    flag_module = types.ModuleType("FlagEmbedding")
+    flag_module.BGEM3FlagModel = lambda *args, **kwargs: None
+    flag_module.FlagReranker = lambda *args, **kwargs: None
+    sys.modules["FlagEmbedding"] = flag_module
+    pil_module = types.ModuleType("PIL")
+    pil_image_module = types.ModuleType("PIL.Image")
+    pil_module.Image = pil_image_module
+    sys.modules["PIL"] = pil_module
+    sys.modules["PIL.Image"] = pil_image_module
+    pythainlp_module = types.ModuleType("pythainlp")
+    tokenize_module = types.ModuleType("pythainlp.tokenize")
+    tokenize_module.word_tokenize = lambda text, **kwargs: text.split()
+    util_module = types.ModuleType("pythainlp.util")
+    util_module.normalize = lambda text: text
+    sys.modules["pythainlp"] = pythainlp_module
+    sys.modules["pythainlp.tokenize"] = tokenize_module
+    sys.modules["pythainlp.util"] = util_module
+
+
+def load_app(upload_base: Path):
+    install_import_stubs()
+    os.environ["OCR_SIDECAR_API_KEY"] = "test-key"
+    os.environ["OCR_SIDECAR_UPLOAD_BASE"] = str(upload_base)
+    if str(SIDECAR_DIR) not in sys.path:
+        sys.path.insert(0, str(SIDECAR_DIR))
+    sys.modules.pop("app", None)
+    return importlib.import_module("app")
+
+
+class FakePage:
+    def get_text(self) -> str:
+        return "A" * 120
+
+
+class FakeDocument:
+    name = "fake.pdf"
+
+    def __len__(self) -> int:
+        return 1
+
+    def __getitem__(self, index: int) -> FakePage:
+        return FakePage()
+
+
+def test_ocr_rejects_parent_traversal_outside_upload_base(tmp_path: Path) -> None:
+    upload_base = tmp_path / "uploads"
+    upload_base.mkdir()
+    app_module = load_app(upload_base)
+    client = TestClient(app_module.app)
+    outside_path = upload_base / ".." / "outside.pdf"
+    response = client.post(
+        "/ocr",
+        json={"pdfPath": str(outside_path)},
+        headers={"X-API-Key": "test-key"},
+    )
+    assert response.status_code == 403
+
+
+def test_ocr_rejects_prefix_sibling_path(tmp_path: Path) -> None:
+    upload_base = tmp_path / "uploads"
+    sibling = tmp_path / "uploads_evil"
+    upload_base.mkdir()
+    sibling.mkdir()
+    app_module = load_app(upload_base)
+    client = TestClient(app_module.app)
+    response = client.post(
+        "/ocr",
+        json={"pdfPath": str(sibling / "document.pdf")},
+        headers={"X-API-Key": "test-key"},
+    )
+    assert response.status_code == 403
+
+
+def test_ocr_accepts_canonical_path_inside_upload_base(tmp_path: Path) -> None:
+    upload_base = tmp_path / "uploads"
+    upload_base.mkdir()
+    pdf_path = upload_base / "document.pdf"
+    pdf_path.write_bytes(b"%PDF-1.4\n")
+    app_module = load_app(upload_base)
+    client = TestClient(app_module.app)
+    with patch.object(app_module.fitz, "open", return_value=FakeDocument()):
+        response = client.post(
+            "/ocr",
+            json={"pdfPath": str(pdf_path)},
+            headers={"X-API-Key": "test-key"},
+        )
+    assert response.status_code == 200
+    assert response.json()["engineUsed"] == "fast-path"
+

tests/unit/ocr-sidecar/test_residency_wiring.py

@@ -0,0 +1,81 @@
+# File: tests/unit/ocr-sidecar/test_residency_wiring.py
+# Change Log:
+# - 2026-06-20: Added ADR-040 residency wiring tests for process_ocr.
+# - 2026-06-20: Updated for async process_ocr (Phase 6 — async I/O refactor).
+
+import asyncio
+from pathlib import Path
+from types import SimpleNamespace
+from unittest.mock import MagicMock, patch
+
+import pytest
+
+from test_path_traversal import load_app
+
+
+class FakeAsyncResponse:
+    """จำลอง httpx.AsyncClient response สำหรับ async process_ocr"""
+
+    def raise_for_status(self) -> None:
+        return None
+
+    def json(self) -> dict:
+        return {"choices": [{"message": {"content": "{\"natural_text\": \"ok\"}"}}]}
+
+
+class FakeAsyncClient:
+    """จำลอง httpx.AsyncClient สำหรับ async process_ocr"""
+
+    def __init__(self, *args, **kwargs) -> None:
+        self.payload = None
+        FakeAsyncClient.last_payload = None
+
+    async def post(self, url: str, json: dict, headers: dict) -> FakeAsyncResponse:
+        self.payload = json
+        FakeAsyncClient.last_payload = json
+        return FakeAsyncResponse()
+
+    async def aclose(self) -> None:
+        pass
+
+
+FakeAsyncClient.last_payload = None
+
+
+def test_process_ocr_uses_calculated_residency_keep_alive(tmp_path: Path) -> None:
+    """T019: process_ocr ต้องเรียก calculate_ocr_residency และใช้ค่า keep_alive ที่คำนวณได้"""
+    app_module = load_app(tmp_path)
+    decision = SimpleNamespace(keep_alive_seconds=120, reason="headroom-sufficient", vram_headroom_mb=9000.0)
+    fake_client = FakeAsyncClient()
+    with patch.object(app_module, "calculate_ocr_residency", return_value=decision) as calculate, \
+         patch.object(app_module, "prepare_ocr_messages", return_value=[{"content": []}]), \
+         patch.object(app_module, "ollama_client", fake_client):
+        result = asyncio.run(app_module.process_ocr("/tmp/test.pdf", page_num=1))
+    assert result == "ok"
+    calculate.assert_called_once_with(app_module.OCR_ACTIVE_PROFILE)
+    assert FakeAsyncClient.last_payload["keep_alive"] == 120
+
+
+def test_process_ocr_rejects_backend_keep_alive_override(tmp_path: Path) -> None:
+    """T021: process_ocr ต้องปฏิเสธ keep_alive จาก backend"""
+    app_module = load_app(tmp_path)
+
+    async def run_test():
+        with pytest.raises(ValueError, match="keep_alive must be calculated"):
+            await app_module.process_ocr("/tmp/test.pdf", options_override={"keep_alive": 0})
+
+    asyncio.run(run_test())
+
+
+def test_ocr_endpoint_rejects_keep_alive_override(tmp_path: Path) -> None:
+    """T021: /ocr endpoint ต้องปฏิเสธ keep_alive ใน request body"""
+    app_module = load_app(tmp_path)
+    from fastapi.testclient import TestClient
+
+    client = TestClient(app_module.app)
+    response = client.post(
+        "/ocr",
+        json={"pdfPath": str(tmp_path / "document.pdf"), "keep_alive": 0},
+        headers={"X-API-Key": "test-key"},
+    )
+    assert response.status_code == 400