feat: add debug-chat-completion-delay30s model

inference/providers/debug/chat_completion.py

@@ -53,6 +53,15 @@ async def chat_completion(
     ):
         if provider_model_id == "debug-error" and messages[-1].content != "Only say your name":
             raise_http_error(ErrorCode.PROVIDER_ERROR, "Debug error for test")
+
+        if provider_model_id == "debug-chat-completion-delay":
+            # content format should be "#number#some other content"
+            message = messages[-1].content
+            second = int(message.split("#")[1]) if message.startswith("#") else 30
+            import asyncio
+
+            await asyncio.sleep(second)
+
         input_tokens = estimate_input_tokens(
             [message.model_dump() for message in messages],
             [function.model_dump() for function in functions] if functions else None,
@@ -99,6 +108,14 @@ async def chat_completion_stream(
         if provider_model_id == "debug-error":
             raise_http_error(ErrorCode.PROVIDER_ERROR, "Debug error for test")
 
+        if provider_model_id == "debug-chat-completion-delay":
+            # content format should be "#number#some other content"
+            message = messages[-1].content
+            second = int(message.split("#")[1]) if message.startswith("#") else 30
+            import asyncio
+
+            await asyncio.sleep(second)
+
         if provider_model_id == "debug-tool-call-hallucinations" and messages[-1].role == ChatCompletionRole.user:
             output_message = create_tool_call_hallucination_message()
             finish_reason = ChatCompletionFinishReason.function_calls

inference/providers/debug/resources/i18n/en.yml

@@ -5,6 +5,9 @@ debug_api_key_description: "debug"
 debug_chat_completion_name: "Debug Chat Completion"
 debug_chat_completion_description: "Debug chat completion model for testing purposes."
 
+debug_chat_completion_delay_name: "Debug Chat Completion Delay"
+debug_chat_completion_delay_description: "Chat completion model with a delay for testing response latency and behavior."
+
 debug_error_name: "Debug Error"
 debug_error_description: "Debug error model for testing purposes."
 

inference/providers/debug/resources/models/debug-chat-completion-delay.yml

@@ -0,0 +1,26 @@
+model_schema_id: debug/debug-chat-completion-delay
+provider_model_id: debug-chat-completion-delay
+type: chat_completion
+name: "i18n:debug_chat_completion_delay_name"
+description: "i18n:debug_chat_completion_delay_description"
+default_endpoint_url:
+
+
+properties:
+  function_call: false
+  streaming: true
+  input_token_limit: 8096
+  output_token_limit: 8096
+
+config_schemas:
+  - config_id: temperature
+  - config_id: top_p
+  - config_id: max_tokens
+  - config_id: stop
+  - config_id: top_k
+
+pricing:
+  input_token: 0
+  output_token: 0
+  unit: 1000
+  currency: USD

inference/providers/debug/resources/models/debug-error.yml

@@ -7,7 +7,7 @@ default_endpoint_url:
 
 
 properties:
-  function_call: false
+  function_call: true
   streaming: true
   input_token_limit: 8096
   output_token_limit: 8096

inference/test/test_chat_completion.py

@@ -97,7 +97,7 @@ async def test_chat_completion_by_normal_function_call(self, test_data):
         message = test_data["message"]
         function_call = test_data["function_call"]
 
-        if not function_call or "azure" in model_schema_id or "openrouter" in model_schema_id:
+        if not function_call or "azure" in model_schema_id or "openrouter" or "debug-error" in model_schema_id:
             pytest.skip("Skip the test case without function call.")
         configs = {
             "temperature": 0.5,
@@ -287,6 +287,7 @@ async def test_chat_completion_by_stream_and_function_call(self, test_data):
             or not stream
             or "azure" in model_schema_id
             or "openrouter" in model_schema_id
+            or "debug-error" in model_schema_id
             or "togetherai" in model_schema_id
         ):
             pytest.skip("Skip the test case without function call or stream.")
@@ -384,7 +385,7 @@ async def test_chat_completion_by_function_call_and_length(self, test_data):
         if (
             not function_call
             or "google_gemini" in model_schema_id
-            or "debug-tool-call-hallucinations" in model_schema_id
+            or "debug" in model_schema_id
             or "sensetime" in model_schema_id
             or "openrouter" in model_schema_id
         ):
@@ -445,7 +446,7 @@ async def test_chat_completion_by_stream_and_function_call_and_length(self, test
             not function_call
             or not stream
             or "azure" in model_schema_id
-            or "debug-tool-call-hallucinations" in model_schema_id
+            or "debug" in model_schema_id
             or "mistralai" in model_schema_id
             or "google_gemini" in model_schema_id
             or "sensetime" in model_schema_id