Configuration used: CodeRabbit UI

Review profile: CHILL

Plan: Pro

app/migrated_routes/product.py (3)

128-139: SQLite unique violations still return 500 instead of 409.

The exception handling only detects psycopg2.UniqueViolation. Under SQLite (used in tests and local development), ie.orig is a sqlite3.IntegrityError, causing the code to raise and leak a 500 error instead of returning 409. Import sqlite3 and broaden the check:

+import sqlite3
+
 from flask.views import MethodView

Then update the exception handling:

         except IntegrityError as ie:
             db.session.rollback()
             if (
-                isinstance(ie.orig, UniqueViolation)
+                (isinstance(ie.orig, UniqueViolation)
+                or (isinstance(ie.orig, sqlite3.IntegrityError)
+                    and "UNIQUE constraint" in str(ie.orig)))
                 and ie.orig.diag.constraint_name
                 == ProductCollection._NAME_UNIQUE_CONSTRAINT.name
             ):
                 abort(409, message="Product with this name already exists")

Note: For SQLite, you'll need to parse the constraint name from the error message or use a different approach since ie.orig.diag is psycopg2-specific.

---

236-250: Same SQLite handling gap exists in PUT.

This exception block has the same issue as the POST handler: SQLite IntegrityError instances fall through to raise, returning 500 instead of 409. Apply the same fix pattern as suggested for lines 128-139.

---

221-224: Walrus operator still blocks empty name updates.

Line 221 uses if name := data.get("name") which skips the assignment when name is an empty string. While validation prevents empty names, the check should use key presence for consistency. Line 223 correctly uses "description" in data, which allows clearing the description field.

Apply this diff to make name handling consistent:

-        if name := data.get("name"):
-            product.name = name
+        if "name" in data:
+            product.name = data["name"]

app/schemas.py (1)

84-84: Consider accepting **_ if kwargs is intentionally unused.

The kwargs parameter is unused but required by the @pre_load decorator signature. If you don't need it, replace with **_ to signal intentional non-use:

-    def strip_strings(self, data, **kwargs):
+    def strip_strings(self, data, **_):

app/migrated_routes/product.py (2)

31-42: Constraint discovery could fail silently.

_get_name_unique_constraint() uses next() without a default, so if the constraint isn't found (e.g., schema mismatch, migration issue), it raises StopIteration during class definition, preventing the entire module from loading. Consider adding error handling:

     @staticmethod
     def _get_name_unique_constraint():
         name_col = Product.__table__.c.name
-        return next(
-            con
-            for con in Product.__table__.constraints
-            if isinstance(con, UniqueConstraint)
-            and len(con.columns) == 1
-            and con.columns.contains_column(name_col)
-        )
+        try:
+            return next(
+                con
+                for con in Product.__table__.constraints
+                if isinstance(con, UniqueConstraint)
+                and len(con.columns) == 1
+                and con.columns.contains_column(name_col)
+            )
+        except StopIteration:
+            raise RuntimeError("Product.name unique constraint not found in schema")

---

233-233: .extend() can create duplicates if called repeatedly.

Line 233 uses product.subcategories.extend(subcategories) which appends to the existing relationships. If a client sends a PUT request with the same subcategory IDs multiple times, or if subcategories already exist, this creates duplicate associations (until the unique constraint at commit catches them). Consider replacing the collection or checking for existing associations first:

-                product.subcategories.extend(subcategories)
+                # Replace existing subcategories or merge carefully
+                existing_ids = {sc.id for sc in product.subcategories}
+                new_subcategories = [sc for sc in subcategories if sc.id not in existing_ids]
+                product.subcategories.extend(new_subcategories)

Or clarify in documentation that PUT subcategories is additive, not a replacement operation.

Configuration used: CodeRabbit UI

Review profile: CHILL

Plan: Pro