added better defaults

Author: DinisCruz

osbot_llms/fast_api/routes/Routes__Chat.py

@@ -16,6 +16,9 @@
 ROUTES_PATHS__CONFIG        = ['/config/status', '/config/version']
 HEADER_NAME__CHAT_ID        = 'osbot-llms-chat-id'
 HEADER_NAME__CHAT_THREAD_ID = 'osbot-llms-thread-id'
+HEADER_NAME__CHAT_PLATFORM  = 'osbot-llms-platform'
+HEADER_NAME__CHAT_PROVIDER  = 'osbot-llms-provider'
+HEADER_NAME__CHAT_MODEL     = 'osbot-llms-model'
 
 class Routes__Chat(Fast_API_Routes):
     tag                     : str = 'chat'
@@ -78,6 +81,11 @@ async def completion(self, request: Request, llm_chat_completion: LLMs__Chat_Com
 
         routes_open_ai   = Routes__OpenAI()
         user_data        = llm_chat_completion.user_data
+        if user_data is None:
+            user_data = dict(selected_platform = llm_chat_completion.llm_platform ,
+                             selected_provider = llm_chat_completion.llm_provider ,
+                             selected_model    = llm_chat_completion.llm_model    )
+            llm_chat_completion.user_data = user_data
 
         # for now use the code in routes_open_ai.prompt_with_system__stream which is already working for OpenAI
         if user_data and 'selected_platform' in user_data and user_data.get('selected_platform') != 'OpenAI (Paid)':
@@ -86,6 +94,9 @@ async def completion(self, request: Request, llm_chat_completion: LLMs__Chat_Com
                 pass
                 response.headers.append(HEADER_NAME__CHAT_ID       , chat_save_result.get('public_chat_id'        ,''))
                 response.headers.append(HEADER_NAME__CHAT_THREAD_ID, chat_save_result.get('public_chat_thread__id',''))
+                response.headers.append(HEADER_NAME__CHAT_PLATFORM , user_data.get('selected_platform'            ,''))
+                response.headers.append(HEADER_NAME__CHAT_PROVIDER , user_data.get('selected_provider'            ,''))
+                response.headers.append(HEADER_NAME__CHAT_MODEL    , user_data.get('selected_model'               ,''))
             return response
         else:
             stream = llm_chat_completion.stream

osbot_llms/models/LLMs__Chat_Completion.py

@@ -5,11 +5,14 @@
 
 from osbot_llms.models.GPT_Prompt_With_System_And_History import GPT_Prompt_With_System_And_History
 
-SWAGGER_EXAMPLE__LLMs__Chat_Completion  = Body(..., example=dict(user_prompt    ='Good morning, what is 44-2?',
+SWAGGER_EXAMPLE__LLMs__Chat_Completion  = Body(..., example=dict(user_prompt    ='Good morning, what is 44-2?' ,
                                                                         system_prompts = ['use emojis in the answer' ],
                                                                         #temperature    = 0.0                         ,
-                                                                        seed           = 42                           ,
-                                                                        stream         = False                        ))
+                                                                        seed            = 42                          ,
+                                                                        stream          = False                       ,
+                                                                        llm_platform    = "Groq (Free)"               ,
+                                                                        llm_provider    = "1. Meta"                   ,
+                                                                        llm_model       = "llama-3.1-70b-versatile"   ))
 
 @dataclass
 class LLMs__Chat_Completion(GPT_Prompt_With_System_And_History):

osbot_llms/test__chained_fast_api.py

@@ -0,0 +1,64 @@
+from unittest import TestCase
+
+import httpx
+from fastapi                                import FastAPI, Request
+from osbot_utils.helpers.Random_Guid        import Random_Guid
+from osbot_utils.utils.Http                 import url_join_safe
+from osbot_fast_api.utils.Fast_API_Server   import Fast_API_Server
+from starlette.responses                    import StreamingResponse
+
+
+class test__chained_fast_api(TestCase):
+
+    def test_check__fast_api_2(self):
+
+        # FastAPI_2 (Backend Service):
+        app_2 = FastAPI()
+
+        async def generate_stream_data(api_key):
+            for i in range(3):
+                yield f"Data chunk {i}: {api_key}\n"
+
+        @app_2.post("/chat/completions")
+        async def chat_completions(request: Request):
+            request_body = await request.json()
+            api_key      = request_body.get("api_key", '(no key)')
+            return StreamingResponse(generate_stream_data(api_key), media_type="text/plain")
+
+
+        # FastAPI_1 (Intermediate Service)
+        app_1 = FastAPI()
+
+        async def proxy_stream(url: str, json_payload: dict, headers: dict):
+            async with httpx.AsyncClient() as client:
+                async with client.stream("POST", url, json=json_payload, headers=headers) as response:
+                    async for chunk in response.aiter_text():
+                        yield chunk
+
+
+        @app_1.post("/chat/completions")
+        async def chat_completions(request: Request):
+            request_body = await request.json()
+            api_key      = f"YOUR_API_KEY : {Random_Guid()}"
+            backend_url  = request_body.get("backend_url")
+            request_body["api_key"] = api_key
+
+            response  = StreamingResponse(proxy_stream(backend_url, request_body, headers={"accept": "text/plain"}),media_type="text/plain")
+            response.headers['api_key'] = api_key
+            return response
+
+        with Fast_API_Server(app=app_2) as fast_api_2:
+            response_1 = fast_api_2.requests_post("/chat/completions", data={})
+            assert response_1.status_code == 200
+            assert response_1.text        == 'Data chunk 0: (no key)\nData chunk 1: (no key)\nData chunk 2: (no key)\n'
+
+            post_data  = dict(user_prompt = '40_2',
+                              backend_url = url_join_safe(fast_api_2.url(), '/chat/completions'))
+            with Fast_API_Server(app=app_1) as fast_api_1:
+                response_2 = fast_api_1.requests_post("/chat/completions", data=post_data)
+                assert response_2.status_code == 200
+                api_key = response_2.headers.get('api_key')
+                assert 'YOUR_API_KEY' in api_key
+                assert response_2.text == (f'Data chunk 0: {api_key}\n'
+                                           f'Data chunk 1: {api_key}\n'
+                                           f'Data chunk 2: {api_key}\n')

tests/integration/fast_api/routes/test__api__Routes__Chat.py

@@ -4,7 +4,8 @@
 from osbot_llms.OSBot_LLMs__Server_Config                       import osbot_llms__server_config
 from osbot_llms.OSBot_LLMs__Shared_Objects                      import osbot_llms__shared_objects
 from osbot_llms.backend.s3_minio.S3_DB__Chat_Threads            import CHAT__REQUEST_TYPE__USER_RESPONSE
-from osbot_llms.fast_api.routes.Routes__Chat                    import HEADER_NAME__CHAT_THREAD_ID, HEADER_NAME__CHAT_ID
+from osbot_llms.fast_api.routes.Routes__Chat                    import HEADER_NAME__CHAT_THREAD_ID, HEADER_NAME__CHAT_ID, \
+    HEADER_NAME__CHAT_PLATFORM, HEADER_NAME__CHAT_PROVIDER, HEADER_NAME__CHAT_MODEL
 from osbot_llms.models.LLMs__Chat_Completion                    import LLMs__Chat_Completion
 from osbot_llms.testing.TestCase__S3_Minio__Temp_Chat_Threads   import TestCase__S3_Minio__Temp_Chat_Threads
 from tests.llm_fast_api__for_tests                              import llm_fast_api__client
@@ -36,7 +37,9 @@ def test__completion__save_chat_completion__user_request(self):
         s3_key              = f'{s3_folder}{request_type}.json.gz'
         file_data           = self.s3_db_chat_threads.s3_file_data(s3_key)
 
-        assert list_set(dict(response.headers))                   == [ 'content-type', 'fast-api-request-id', HEADER_NAME__CHAT_ID, HEADER_NAME__CHAT_THREAD_ID]
+        assert list_set(dict(response.headers))                   == sorted([ 'content-type', 'fast-api-request-id',
+                                                                              HEADER_NAME__CHAT_ID, HEADER_NAME__CHAT_THREAD_ID,
+                                                                              HEADER_NAME__CHAT_PLATFORM, HEADER_NAME__CHAT_PROVIDER, HEADER_NAME__CHAT_MODEL])
         assert self.s3_db_chat_threads.s3_folder_files(s3_folder) == [ 'user-request.json.gz', 'user-response.json.gz']
         assert list_set(file_data)                                == [ 'chat_thread_id', 'histories', 'images', 'llm_answer',
                                                                        'llm_model', 'llm_platform', 'llm_provider', 'max_tokens',

tests/integration/fast_api/test_Fast_API__via_Http.py

@@ -38,8 +38,14 @@ def test_version(self):
 
     def test__chat__completion(self):
         with self.llm_fast_api as _:
-            llm_chat_completion = LLMs__Chat_Completion(user_prompt='51-9')
+            kwargs = dict(llm_platform = "Groq (Free)"               ,
+                          llm_provider = "1. Meta"                   ,
+                          llm_model    = "llama-3.1-70b-versatile"   ,
+                          user_prompt  = '51-9'                     )
+            llm_chat_completion = LLMs__Chat_Completion(**kwargs)
             response = self.fast_api_server.requests_post('/chat/completion', data=llm_chat_completion)
+            #pprint(dict(response.headers))
+            #pprint(response.text)
             assert '42' in response.text