Test on CI

Author: splasky

pyproject.toml

@@ -0,0 +1,21 @@
+[tool.ruff]
+line-length = 120
+
+[tool.ruff.lint]
+select = ["E", "F", "W", "I", "N", "B", "C", "G", "S"]
+ignore = [
+    "E501",  # Line too long - allow longer lines for SQL strings in tests
+    "S608",  # SQL injection - intentional for test cases
+    "B007",  # Loop variable not used - common in test assertions
+]
+
+[tool.ruff.lint.per-file-ignores]
+"tests/*" = ["S608", "E501", "S101", "B007", "S603"]
+
+[tool.mypy]
+ignore_missing_imports = true
+disable_error_code = ["var-annotated", "import-untyped"]
+
+[[tool.mypy.overrides]]
+module = ["tests.*", "sqlalchemy_datastore.*"]
+ignore_errors = true

sqlalchemy_datastore/datastore_dbapi.py

@@ -639,6 +639,7 @@ def _needs_client_side_filter(self, statement: str) -> bool:
             r"\bIN\s*\(",  # IN clause (emulator has issues)
             r"\bHAS\s+ANCESTOR\b",  # HAS ANCESTOR
             r"\bHAS\s+DESCENDANT\b",  # HAS DESCENDANT
+            r"\bBLOB\s*\(",  # BLOB literal (emulator doesn't support)
         ]
         for pattern in unsupported_patterns:
             if re.search(pattern, upper):
@@ -780,6 +781,43 @@ def _eval_simple_condition(self, context: Dict[str, Any], condition: str) -> boo
         """Evaluate a simple comparison condition."""
         condition = condition.strip()
 
+        # Handle BLOB equality (before generic handlers, since BLOB literal
+        # would confuse the generic _parse_literal path)
+        blob_eq_match = re.match(
+            r"(\w+)\s*=\s*BLOB\s*\('(.*?)'\)",
+            condition,
+            re.IGNORECASE | re.DOTALL,
+        )
+        if blob_eq_match:
+            field = blob_eq_match.group(1)
+            blob_str = blob_eq_match.group(2)
+            try:
+                blob_bytes = blob_str.encode("latin-1")
+            except (UnicodeEncodeError, UnicodeDecodeError):
+                blob_bytes = blob_str.encode("utf-8")
+            field_val = context.get(field)
+            if isinstance(field_val, bytes):
+                return field_val == blob_bytes
+            return False
+
+        # Handle BLOB inequality
+        blob_neq_match = re.match(
+            r"(\w+)\s*!=\s*BLOB\s*\('(.*?)'\)",
+            condition,
+            re.IGNORECASE | re.DOTALL,
+        )
+        if blob_neq_match:
+            field = blob_neq_match.group(1)
+            blob_str = blob_neq_match.group(2)
+            try:
+                blob_bytes = blob_str.encode("latin-1")
+            except (UnicodeEncodeError, UnicodeDecodeError):
+                blob_bytes = blob_str.encode("utf-8")
+            field_val = context.get(field)
+            if isinstance(field_val, bytes):
+                return field_val != blob_bytes
+            return True
+
         # Handle NOT IN
         not_in_match = re.match(
             r"(\w+)\s+NOT\s+IN\s*\(([^)]+)\)", condition, re.IGNORECASE

tests/conftest.py

@@ -184,7 +184,7 @@ def test_datasets(datastore_client):
     task1["task"] = "Collect Sea Urchins in Atelier"
     task1["content"] = {"description": "採集高品質海膽"}
     task1["is_done"] = False
-    task1["tag"] = "house"
+    task1["tag"] = "House"
     task1["location"] = GeoPoint(25.047472, 121.517167)
     task1["assign_user"] = user1.key
     task1["reward"] = 22000.5