.gitea/workflows/ci.yml

@@ -163,7 +163,7 @@ jobs:
           python-version: '3.11'
 
       - name: Install test dependencies
-        run: pip install "pyspellchecker==0.9.0" pytest pytest-asyncio
+        run: pip install "pyspellchecker==0.9.0" "fastapi==0.115.6" pytest pytest-asyncio
         working-directory: ocr-service
 
       - name: Run OCR unit tests (no ML stack required)

ocr-service/main.py

@@ -27,6 +27,7 @@ from engines import kraken as kraken_engine
 from engines import surya as surya_engine
 from models import OcrBlock, OcrRequest
 from preprocessing import preprocess_page
+from utils import _validate_zip_entry
 
 TRAINING_TOKEN = os.environ.get("TRAINING_TOKEN", "")
 KRAKEN_MODEL_PATH = os.environ.get("KRAKEN_MODEL_PATH", "/app/models/german_kurrent.mlmodel")
@@ -291,14 +292,6 @@ def _check_training_token(x_training_token: str | None) -> None:
         raise HTTPException(status_code=403, detail="Invalid or missing X-Training-Token")
 
 
-def _validate_zip_entry(name: str, extract_dir: str) -> None:
-    """Reject ZIP Slip attacks: path traversal and absolute paths."""
-    if os.path.isabs(name) or name.startswith(".."):
-        raise HTTPException(status_code=400, detail=f"Unsafe ZIP entry: {name}")
-    resolved = os.path.realpath(os.path.join(extract_dir, name))
-    if not resolved.startswith(os.path.realpath(extract_dir)):
-        raise HTTPException(status_code=400, detail=f"ZIP Slip detected: {name}")
-
 
 def _rotate_backups(model_path: str, keep: int = 3) -> None:
     """Keep only the last `keep` timestamped backups of the model."""

ocr-service/test_tmpdir.py

@@ -6,12 +6,8 @@ import tempfile
 
 import pytest
 
-try:
-    from fastapi import HTTPException
-    from main import _validate_zip_entry
-    HAS_MAIN = True
-except ImportError:
-    HAS_MAIN = False
+from fastapi import HTTPException
+from utils import _validate_zip_entry
 
 _ENTRYPOINT = os.path.join(os.path.dirname(os.path.abspath(__file__)), "entrypoint.sh")
 
@@ -82,7 +78,6 @@ def test_tmpdir_is_inside_persistent_cache_volume():
     )
 
 
-@pytest.mark.skipif(not HAS_MAIN, reason="requires full ML stack (not available in CI)")
 def test_zipslip_still_anchors_under_custom_tmpdir(tmp_path):
     """_validate_zip_entry rejects path-traversal when extract_dir is under a custom TMPDIR.
 

ocr-service/utils.py

@@ -0,0 +1,14 @@
+"""Utility functions shared across the OCR service with no ML-stack imports."""
+
+import os
+
+from fastapi import HTTPException
+
+
+def _validate_zip_entry(name: str, extract_dir: str) -> None:
+    """Reject ZIP Slip attacks: path traversal and absolute paths."""
+    if os.path.isabs(name) or name.startswith(".."):
+        raise HTTPException(status_code=400, detail=f"Unsafe ZIP entry: {name}")
+    resolved = os.path.realpath(os.path.join(extract_dir, name))
+    if not resolved.startswith(os.path.realpath(extract_dir)):
+        raise HTTPException(status_code=400, detail=f"ZIP Slip detected: {name}")