fix: address Greptile review comments

- Sanitize custom_id to meet GCP label constraints (lowercase, alphanumeric, max 63 chars)
- Improve batch output detection heuristic with processed_time and candidates/status checks
- Move inline imports to module level
- Fix Content-Length header for transformed responses
- Add test for label sanitization

Made-with: Cursor

litellm/llms/vertex_ai/files/handler.py

@@ -17,9 +17,10 @@ from litellm.types.llms.openai import (
     HttpxBinaryResponseContent,
     OpenAIFileObject,
 )
+from litellm.litellm_core_utils.litellm_logging import Logging
 from litellm.types.llms.vertex_ai import VERTEX_CREDENTIALS_TYPES
 
-from .transformation import VertexAIJsonlFilesTransformation
+from .transformation import VertexAIFilesConfig, VertexAIJsonlFilesTransformation
 
 vertex_ai_files_transformation = VertexAIJsonlFilesTransformation()
 
@@ -197,9 +198,6 @@ class VertexAIFilesHandler(GCSBucketBase):
         )
 
         # Apply transformation to convert Vertex AI batch outputs to OpenAI format
-        from .transformation import VertexAIFilesConfig
-        from litellm.litellm_core_utils.litellm_logging import Logging
-        
         config = VertexAIFilesConfig()
         
         # Create a logging object for transformation

litellm/llms/vertex_ai/files/transformation.py

@@ -1,5 +1,6 @@
 import json
 import os
+import re
 import time
 from typing import Any, Dict, List, Optional, Tuple, Union
 
@@ -9,6 +10,7 @@ from openai.types.file_deleted import FileDeleted
 
 from litellm._uuid import uuid
 from litellm.files.utils import FilesAPIUtils
+from litellm.litellm_core_utils.litellm_logging import Logging
 from litellm.litellm_core_utils.prompt_templates.common_utils import extract_file_data
 from litellm.llms.base_llm.chat.transformation import BaseLLMException
 from litellm.llms.base_llm.files.transformation import (
@@ -32,12 +34,31 @@ from litellm.types.llms.openai import (
     PathLike,
 )
 from litellm.types.llms.vertex_ai import GcsBucketResponse
-from litellm.types.utils import ExtractedFileData, LlmProviders
+from litellm.types.utils import ExtractedFileData, LlmProviders, ModelResponse
 
 from ..common_utils import VertexAIError
 from ..vertex_llm_base import VertexBase
 
 
+def _sanitize_gcp_label_value(value: str) -> str:
+    """
+    Sanitize a string to meet GCP label value constraints.
+    
+    GCP label values must:
+    - Be lowercase
+    - Contain only letters, numbers, underscores, and hyphens
+    - Be max 63 characters
+    
+    Args:
+        value: The string to sanitize
+        
+    Returns:
+        A sanitized string that meets GCP label constraints
+    """
+    sanitized = re.sub(r"[^a-z0-9_-]", "_", value.lower())
+    return sanitized[:63]
+
+
 class VertexAIFilesConfig(VertexBase, BaseFilesConfig):
     """
     Config for VertexAI Files
@@ -254,7 +275,7 @@ class VertexAIFilesConfig(VertexBase, BaseFilesConfig):
             if custom_id:
                 if "labels" not in vertex_request_body:
                     vertex_request_body["labels"] = {}
-                vertex_request_body["labels"]["litellm_custom_id"] = str(custom_id)
+                vertex_request_body["labels"]["litellm_custom_id"] = _sanitize_gcp_label_value(str(custom_id))
             
             vertex_jsonl_content.append({"request": vertex_request_body})
         return vertex_jsonl_content
@@ -538,8 +559,20 @@ class VertexAIFilesConfig(VertexBase, BaseFilesConfig):
             # Try to parse the first line to see if it's Vertex AI batch output
             first_line = json.loads(lines[0])
             
-            # Check if it has Vertex AI batch output structure
-            if not ("response" in first_line and "request" in first_line):
+            # Check if it has Vertex AI batch output structure with discriminating fields
+            # Must have request, response, and processed_time
+            # Plus either candidates (success) or status (error)
+            has_base_structure = (
+                "response" in first_line
+                and "request" in first_line
+                and "processed_time" in first_line
+            )
+            has_success_or_error = (
+                "candidates" in first_line.get("response", {})
+                or "status" in first_line
+            )
+            
+            if not (has_base_structure and has_success_or_error):
                 # Not a Vertex AI batch output, return as-is
                 return content
             
@@ -573,10 +606,6 @@ class VertexAIFilesConfig(VertexBase, BaseFilesConfig):
         Transform a single Vertex AI batch output line to OpenAI format.
         Uses the existing VertexGeminiConfig transformation for the response.
         """
-        from litellm.types.utils import ModelResponse
-        import httpx
-        import time
-        
         # Extract custom_id from request labels
         custom_id = "unknown"
         request_data = vertex_output.get("request", {})
@@ -756,7 +785,7 @@ class VertexAIJsonlFilesTransformation(VertexGeminiConfig):
             if custom_id:
                 if "labels" not in vertex_request_body:
                     vertex_request_body["labels"] = {}
-                vertex_request_body["labels"]["litellm_custom_id"] = str(custom_id)
+                vertex_request_body["labels"]["litellm_custom_id"] = _sanitize_gcp_label_value(str(custom_id))
             
             vertex_jsonl_content.append({"request": vertex_request_body})
         return vertex_jsonl_content

tests/test_litellm/llms/vertex_ai/files/test_vertex_ai_files_transformation.py

@@ -526,3 +526,42 @@ class TestVertexBatchCustomIdLabels:
         # Step 4: Verify custom_id was preserved
         assert openai_output["custom_id"] == "my-custom-request-id"
         assert openai_output["response"]["status_code"] == 200
+
+    def test_custom_id_label_sanitization(self):
+        """Test that custom_id values are sanitized to meet GCP label constraints"""
+        from litellm.llms.vertex_ai.files.transformation import (
+            VertexAIJsonlFilesTransformation,
+            _sanitize_gcp_label_value,
+        )
+
+        transformation = VertexAIJsonlFilesTransformation()
+
+        # Test sanitization function
+        assert _sanitize_gcp_label_value("MyRequest-1") == "myrequest-1"
+        assert _sanitize_gcp_label_value("Request.With.Dots") == "request_with_dots"
+        assert _sanitize_gcp_label_value("Request With Spaces") == "request_with_spaces"
+        assert _sanitize_gcp_label_value("Request@#$%Special") == "request____special"
+        
+        # Test max length (63 chars)
+        long_id = "a" * 100
+        assert len(_sanitize_gcp_label_value(long_id)) == 63
+
+        # Test in actual transformation
+        openai_input = [
+            {
+                "custom_id": "MyRequest-1",
+                "method": "POST",
+                "url": "/v1/chat/completions",
+                "body": {
+                    "model": "gemini-1.5-flash-001",
+                    "messages": [{"role": "user", "content": "Hello"}],
+                }
+            }
+        ]
+
+        vertex_input = transformation._transform_openai_jsonl_content_to_vertex_ai_jsonl_content(
+            openai_input
+        )
+
+        # Verify label was sanitized
+        assert vertex_input[0]["request"]["labels"]["litellm_custom_id"] == "myrequest-1"