Collection: adding an error message column (#328)

* altering the table and routes code to add error column

Author: nishika26

backend/app/alembic/versions/5a59c6c29a82_add_error_message_column_in_collections_.py

@@ -0,0 +1,28 @@
+"""add error message column in collections table
+
+Revision ID: 5a59c6c29a82
+Revises: e9dd35eff62c
+Create Date: 2025-08-11 15:40:40.127161
+
+"""
+from alembic import op
+import sqlalchemy as sa
+import sqlmodel.sql.sqltypes
+
+
+# revision identifiers, used by Alembic.
+revision = "5a59c6c29a82"
+down_revision = "e9dd35eff62c"
+branch_labels = None
+depends_on = None
+
+
+def upgrade():
+    op.add_column(
+        "collection",
+        sa.Column("error_message", sqlmodel.sql.sqltypes.AutoString(), nullable=True),
+    )
+
+
+def downgrade():
+    op.drop_column("collection", "error_message")

backend/app/api/routes/collections.py

@@ -1,6 +1,9 @@
 import inspect
 import logging
 import time
+import json
+import ast
+import re
 from uuid import UUID, uuid4
 from typing import Any, List, Optional
 from dataclasses import dataclass, field, fields, asdict, replace
@@ -13,6 +16,7 @@
 
 from app.api.deps import CurrentUser, SessionDep, CurrentUserOrgProject
 from app.core.cloud import AmazonCloudStorage
+from app.api.routes.responses import handle_openai_error
 from app.core.util import now, post_callback
 from app.crud import (
     DocumentCrud,
@@ -28,6 +32,32 @@
 router = APIRouter(prefix="/collections", tags=["collections"])
 
 
+def extract_error_message(err: Exception) -> str:
+    err_str = str(err).strip()
+
+    body = re.sub(r"^Error code:\s*\d+\s*-\s*", "", err_str)
+    message = None
+    try:
+        payload = json.loads(body)
+        if isinstance(payload, dict):
+            message = payload.get("error", {}).get("message")
+    except Exception:
+        pass
+
+    if message is None:
+        try:
+            payload = ast.literal_eval(body)
+            if isinstance(payload, dict):
+                message = payload.get("error", {}).get("message")
+        except Exception:
+            pass
+
+    if not message:
+        message = body
+
+    return message.strip()[:1000]
+
+
 @dataclass
 class ResponsePayload:
     status: str
@@ -246,6 +276,9 @@ def do_create_collection(
             collection = collection_crud.read_one(UUID(payload.key))
             collection.status = CollectionStatus.failed
             collection.updated_at = now()
+            message = extract_error_message(err)
+            collection.error_message = message
+
             collection_crud._update(collection)
         except Exception as suberr:
             logger.warning(
@@ -283,7 +316,6 @@ def create_collection(
     collection_crud = CollectionCrud(session, current_user.id)
     collection_crud.create(collection)
 
-    # 2. Launch background task
     background_tasks.add_task(
         do_create_collection, session, current_user, request, payload, client
     )

backend/app/crud/document.py

@@ -81,13 +81,15 @@ def read_each(self, doc_ids: List[UUID]):
         )
         results = self.session.exec(statement).all()
 
-        (m, n) = map(len, (results, doc_ids))
-        if m != n:
+        (retrieved_count, requested_count) = map(len, (results, doc_ids))
+        if retrieved_count != requested_count:
             try:
-                raise ValueError(f"Requested {n} retrieved {m}")
+                raise ValueError(
+                    f"Requested atleast {requested_count} document retrieved {retrieved_count}"
+                )
             except ValueError as err:
                 logger.error(
-                    f"[DocumentCrud.read_each] Mismatch in retrieved documents | {{'owner_id': {self.owner_id}, 'requested_count': {n}, 'retrieved_count': {m}}}",
+                    f"[DocumentCrud.read_each] Mismatch in retrieved documents | {{'owner_id': {self.owner_id}, 'requested_count': {requested_count}, 'retrieved_count': {retrieved_count}}}",
                     exc_info=True,
                 )
                 raise

backend/app/models/collection.py

@@ -43,6 +43,7 @@ class Collection(SQLModel, table=True):
     llm_service_name: Optional[str] = Field(default=None, nullable=True)
 
     status: CollectionStatus = Field(default=CollectionStatus.processing)
+    error_message: Optional[str] = Field(default=None, nullable=True)
 
     created_at: datetime = Field(default_factory=now)
     updated_at: datetime = Field(default_factory=now)