fix(managed_files.py): support for DELETE endpoint for files

litellm/proxy/hooks/managed_files.py

@@ -8,7 +8,7 @@ from abc import ABC, abstractmethod
 from datetime import datetime
 from typing import TYPE_CHECKING, Any, Dict, List, Literal, Optional, Union, cast
 
-from litellm import verbose_logger
+from litellm import Router, verbose_logger
 from litellm.caching.caching import DualCache
 from litellm.integrations.custom_logger import CustomLogger
 from litellm.litellm_core_utils.prompt_templates.common_utils import extract_file_data
@@ -59,7 +59,7 @@ class BaseFileEndpoints(ABC):
         pass
 
 
-class _PROXY_LiteLLMManagedFiles(CustomLogger, BaseFileEndpoints):
+class _PROXY_LiteLLMManagedFiles(CustomLogger):
     # Class variables or attributes
     def __init__(
         self, internal_usage_cache: InternalUsageCache, prisma_client: PrismaClient
@@ -92,6 +92,25 @@ class _PROXY_LiteLLMManagedFiles(CustomLogger, BaseFileEndpoints):
             litellm_parent_otel_span=litellm_parent_otel_span,
         )
 
+    async def delete_unified_file_id(
+        self, file_id: str, litellm_parent_otel_span: Optional[Span] = None
+    ) -> OpenAIFileObject:
+        key = f"litellm_proxy/{file_id}"
+        ## get old value
+        old_value = await self.internal_usage_cache.async_get_cache(
+            key=key,
+            litellm_parent_otel_span=litellm_parent_otel_span,
+        )
+        if old_value is None or not isinstance(old_value, OpenAIFileObject):
+            raise Exception(f"LiteLLM Managed File object with id={file_id} not found")
+        ## delete old value
+        await self.internal_usage_cache.async_set_cache(
+            key=key,
+            value=None,
+            litellm_parent_otel_span=litellm_parent_otel_span,
+        )
+        return old_value
+
     async def async_pre_call_hook(
         self,
         user_api_key_dict: UserAPIKeyAuth,
@@ -322,6 +341,24 @@ class _PROXY_LiteLLMManagedFiles(CustomLogger, BaseFileEndpoints):
         return []
 
     async def afile_delete(
-        self, custom_llm_provider: str, file_id: str, **data: Dict
+        self,
+        file_id: str,
+        litellm_parent_otel_span: Optional[Span],
+        llm_router: Router,
+        **data: Dict,
     ) -> OpenAIFileObject:
-        raise NotImplementedError("afile_delete not implemented")
+        model_file_id_mapping = await self.get_model_file_id_mapping(
+            [file_id], litellm_parent_otel_span
+        )
+        specific_model_file_id_mapping = model_file_id_mapping.get(file_id)
+        if specific_model_file_id_mapping:
+            for model_id, file_id in specific_model_file_id_mapping.items():
+                await llm_router.afile_delete(model=model_id, file_id=file_id, **data)  # type: ignore
+
+        stored_file_object = await self.delete_unified_file_id(
+            file_id, litellm_parent_otel_span
+        )
+        if stored_file_object:
+            return stored_file_object
+        else:
+            raise Exception(f"LiteLLM Managed File object with id={file_id} not found")

litellm/proxy/openai_files_endpoints/files_endpoints.py

@@ -565,7 +565,8 @@ async def get_file(
                     code=500,
                 )
             response = await managed_files_obj.afile_retrieve(
-                file_id=file_id, litellm_parent_otel_span=user_api_key_dict.parent_otel_span, **data  # type: ignore
+                file_id=file_id,
+                litellm_parent_otel_span=user_api_key_dict.parent_otel_span,
             )
         else:
             response = await litellm.afile_retrieve(
@@ -663,6 +664,7 @@ async def delete_file(
     from litellm.proxy.proxy_server import (
         add_litellm_data_to_request,
         general_settings,
+        llm_router,
         proxy_config,
         proxy_logging_obj,
         version,
@@ -685,10 +687,41 @@ async def delete_file(
             proxy_config=proxy_config,
         )
 
-        response = await litellm.afile_delete(
-            custom_llm_provider=custom_llm_provider, file_id=file_id, **data  # type: ignore
+        ## check if file_id is a litellm managed file
+        is_base64_unified_file_id = (
+            _PROXY_LiteLLMManagedFiles._is_base64_encoded_unified_file_id(file_id)
         )
 
+        if is_base64_unified_file_id:
+            managed_files_obj = cast(
+                Optional[_PROXY_LiteLLMManagedFiles],
+                proxy_logging_obj.get_proxy_hook("managed_files"),
+            )
+            if managed_files_obj is None:
+                raise ProxyException(
+                    message="Managed files hook not found",
+                    type="None",
+                    param="None",
+                    code=500,
+                )
+            if llm_router is None:
+                raise ProxyException(
+                    message="LLM Router not found",
+                    type="None",
+                    param="None",
+                    code=500,
+                )
+            response = await managed_files_obj.afile_delete(
+                file_id=file_id,
+                litellm_parent_otel_span=user_api_key_dict.parent_otel_span,
+                llm_router=llm_router,
+                **data,
+            )
+        else:
+            response = await litellm.afile_delete(
+                custom_llm_provider=custom_llm_provider, file_id=file_id, **data  # type: ignore
+            )
+
         ### ALERTING ###
         asyncio.create_task(
             proxy_logging_obj.update_request_status(

litellm/router.py

@@ -729,6 +729,9 @@ class Router:
         self.aresponses = self.factory_function(
             litellm.aresponses, call_type="aresponses"
         )
+        self.afile_delete = self.factory_function(
+            litellm.afile_delete, call_type="afile_delete"
+        )
         self.responses = self.factory_function(litellm.responses, call_type="responses")
 
     def validate_fallbacks(self, fallback_param: Optional[List]):
@@ -2435,6 +2438,8 @@ class Router:
             model_name = data["model"]
             self.total_calls[model_name] += 1
 
+            ### get custom
+
             response = original_function(
                 **{
                     **data,
@@ -2514,9 +2519,15 @@ class Router:
             # Perform pre-call checks for routing strategy
             self.routing_strategy_pre_call_checks(deployment=deployment)
 
+            try:
+                _, custom_llm_provider, _, _ = get_llm_provider(model=data["model"])
+            except Exception:
+                custom_llm_provider = None
+
             response = original_function(
                 **{
                     **data,
+                    "custom_llm_provider": custom_llm_provider,
                     "caching": self.cache_responses,
                     **kwargs,
                 }
@@ -3058,6 +3069,7 @@ class Router:
             "anthropic_messages",
             "aresponses",
             "responses",
+            "afile_delete",
         ] = "assistants",
     ):
         """
@@ -3102,7 +3114,7 @@ class Router:
                 return await self._pass_through_moderation_endpoint_factory(
                     original_function=original_function, **kwargs
                 )
-            elif call_type in ("anthropic_messages", "aresponses"):
+            elif call_type in ("anthropic_messages", "aresponses", "afile_delete"):
                 return await self._ageneric_api_call_with_fallbacks(
                     original_function=original_function,
                     **kwargs,