added submodule OSbot-Fast-API

DinisCruz · DinisCruz · commit dd5c9f1780b2 · 2024-10-11T13:20:08.000+01:00
simproved start-local-server.sh major change where LLMs_Chat_Completion now uses @DataClass (via GPT_Prompt_Simple) added prompt example to Swagger for /chat/completion
diff --git a/.gitmodules b/.gitmodules
@@ -0,0 +1,3 @@
+[submodule "modules/OSBot-Fast-API"]
+	path = modules/OSBot-Fast-API
+	url = git@github.com:owasp-sbot/OSBot-Fast-API.git
diff --git a/modules/OSBot-Fast-API b/modules/OSBot-Fast-API
@@ -0,0 +1 @@
+Subproject commit e3cd3b14434cf1ace4e0f632ddfef787099eade9
diff --git a/osbot_llms/backend/s3_minio/S3_DB__Chat_Threads.py b/osbot_llms/backend/s3_minio/S3_DB__Chat_Threads.py
@@ -42,7 +42,7 @@ def save_chat_completion__user_request(self, llm_chat_completion: LLMs__Chat_Com
                                       chat_thread_id      = chat_thread_id                   ,
                                       llm_request_id      = llm_request_id                   ,
                                       timestamp           =  timestamp_utc_now()             ,
-                                      llm_chat_completion =  llm_chat_completion.model_dump())
+                                      llm_chat_completion =  llm_chat_completion.json())
         s3_key                 = self.s3_key(chat_thread_id=chat_thread_id, llm_request_id=llm_request_id, request_type=request_type)
         s3_key_items           = s3_key.split('/')
         public_chat_id         = '/'.join(s3_key_items[2:6])
@@ -62,7 +62,7 @@ def save_chat_completion__user_response(self, llm_chat_completion: LLMs__Chat_Co
         chat_thread_id = llm_chat_completion.chat_thread_id
         llm_request_id = request_id
         s3_key         = self.s3_key(chat_thread_id=chat_thread_id, llm_request_id=llm_request_id, request_type=request_type)
-        data           = llm_chat_completion.model_dump()
+        data           = llm_chat_completion.json()
         metadata = {'request_id'    : request_id    ,
                     'request_type'  : request_type  }
         if self.s3_save_data(data=data, s3_key=s3_key, metadata=metadata):
diff --git a/osbot_llms/fast_api/routes/Routes__Chat.py b/osbot_llms/fast_api/routes/Routes__Chat.py
@@ -2,14 +2,16 @@
 import traceback
 
 from fastapi                                                                import Request
+from fastapi.params import Header, Body
+from osbot_utils.utils.Dev import pprint
 from starlette.responses                                                    import StreamingResponse
 from osbot_fast_api.api.Fast_API_Routes                                     import Fast_API_Routes
 from osbot_utils.context_managers.capture_duration                          import capture_duration
 from osbot_llms.OSBot_LLMs__Shared_Objects                                  import osbot_llms__shared_objects
 from osbot_llms.fast_api.routes.Routes__OpenAI                              import Routes__OpenAI
 from osbot_llms.llms.chats.LLM__Chat_Completion__Resolve_Engine             import LLM__Chat_Completion__Resolve_Engine
 from osbot_llms.llms.storage.Chats_Storage__S3_Minio                        import Chats_Storage__S3_Minio
-from osbot_llms.models.LLMs__Chat_Completion                                import LLMs__Chat_Completion
+from osbot_llms.models.LLMs__Chat_Completion import LLMs__Chat_Completion, SWAGGER_EXAMPLE__LLMs__Chat_Completion
 
 ROUTES_PATHS__CONFIG        = ['/config/status', '/config/version']
 HEADER_NAME__CHAT_ID        = 'osbot-llms-chat-id'
@@ -70,7 +72,7 @@ async def simulated_api_call():                         # Simulating the respons
             traceback.print_exc()
 
 
-    async def completion(self, llm_chat_completion: LLMs__Chat_Completion, request: Request):
+    async def completion(self, request: Request, llm_chat_completion: LLMs__Chat_Completion = SWAGGER_EXAMPLE__LLMs__Chat_Completion):
         request_id       = self.request_id(request)
         chat_save_result = self.chats_storage_s3_minio.save_user_request(llm_chat_completion, request_id)
 
diff --git a/osbot_llms/fast_api/routes/Routes__Info.py b/osbot_llms/fast_api/routes/Routes__Info.py
@@ -1,17 +1,14 @@
+from osbot_utils.base_classes.Type_Safe import Type_Safe
 from osbot_fast_api.api.Fast_API_Routes import Fast_API_Routes
 
-from osbot_llms.utils.Version import version__osbot_llms
-
+from osbot_llms.utils.Version           import version__osbot_llms
+from dataclasses                        import dataclass
 
 class Routes__Info(Fast_API_Routes):
     tag : str = 'info'
 
     def version(self):
         return {"version" : version__osbot_llms }
 
-    def ping(self):
-        return {"it_is" : "pong" }
-
     def setup_routes(self):
-        self.add_route_get(self.version)
-        self.add_route_get(self.ping   )
+        self.add_route_get(self.version)
diff --git a/osbot_llms/fast_api/routes/Routes__OpenAI.py b/osbot_llms/fast_api/routes/Routes__OpenAI.py
@@ -97,7 +97,7 @@ async def streamer():
                     gpt_response += answer
                     yield f"{answer}\n"
 
-            llm_chat_completion = LLMs__Chat_Completion(**gpt_prompt_with_system_and_history.model_dump())
+            llm_chat_completion = LLMs__Chat_Completion(**gpt_prompt_with_system_and_history.json())
             llm_chat_completion.llm_answer = gpt_response
 
             self.chats_storage_s3_minio.save_user_response(llm_chat_completion, request_id)
diff --git a/osbot_llms/models/GPT_Prompt_Simple.py b/osbot_llms/models/GPT_Prompt_Simple.py
@@ -1,16 +1,20 @@
+from dataclasses import dataclass, field
 from decimal import Decimal
-from typing                                     import Optional
+from typing import Optional, List
+
+from osbot_utils.base_classes.Type_Safe import Type_Safe
 from pydantic                                   import BaseModel
 DEFAULT_USER_PROMPT = 'Hi'
 DEFAULT_TEMPERATURE = 0.0
 DEFAULT_SEED        = 42
 
-class GPT_Prompt_Simple(BaseModel):
-    chat_thread_id: Optional[str]  = None
-    user_prompt   : str            = DEFAULT_USER_PROMPT
-    images        : list[str]      = []
-    temperature   : Decimal        = Decimal(DEFAULT_TEMPERATURE)
-    seed          : int            = DEFAULT_SEED
-    max_tokens    : Optional[int]  = None
-    user_data     : Optional[dict] = None
-    stream        : Optional[bool] = True
+@dataclass
+class GPT_Prompt_Simple(Type_Safe):
+    images        : list                = field(default_factory=list)
+    chat_thread_id: Optional[str]       = field(default='')
+    user_prompt   : str                 = DEFAULT_USER_PROMPT
+    temperature   : float               = float(DEFAULT_TEMPERATURE)
+    seed          : int                 = DEFAULT_SEED
+    max_tokens    : Optional[int]       = None
+    user_data     : Optional[dict]      = None
+    stream        : Optional[bool]      = True
diff --git a/osbot_llms/models/GPT_Prompt_With_System.py b/osbot_llms/models/GPT_Prompt_With_System.py
@@ -1,7 +1,8 @@
+from dataclasses import dataclass
 from typing                              import Optional
 
 from osbot_llms.models.GPT_Prompt_Simple import GPT_Prompt_Simple
 
-
+@dataclass
 class GPT_Prompt_With_System(GPT_Prompt_Simple):
     system_prompts: Optional[list[str]] = None
diff --git a/osbot_llms/models/GPT_Prompt_With_System_And_History.py b/osbot_llms/models/GPT_Prompt_With_System_And_History.py
@@ -1,8 +1,9 @@
+from dataclasses import dataclass
 from typing                                                import Optional
 
 from osbot_llms.models.GPT_History            import GPT_History
 from osbot_llms.models.GPT_Prompt_With_System import GPT_Prompt_With_System
 
-
+@dataclass
 class GPT_Prompt_With_System_And_History(GPT_Prompt_With_System):
     histories      : Optional[list[GPT_History]] = None
diff --git a/osbot_llms/models/LLMs__Chat_Completion.py b/osbot_llms/models/LLMs__Chat_Completion.py
@@ -1,8 +1,17 @@
+from dataclasses import dataclass
 from typing import Optional
 
+from fastapi.params import Body
+
 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?',
+                                                                        system_prompts = ['use emojis in the answer' ],
+                                                                        #temperature    = 0.0                         ,
+                                                                        seed           = 42                           ,
+                                                                        stream         = False                        ))
 
+@dataclass
 class LLMs__Chat_Completion(GPT_Prompt_With_System_And_History):
     llm_platform: Optional[str] = None
     llm_provider: Optional[str] = None
diff --git a/server.py b/server.py
diff --git a/server__fastapi__llms.py b/server__fastapi__llms.py
@@ -0,0 +1,7 @@
+from osbot_llms.OSBot_LLMs__Server_Config import osbot_llms__server_config
+from osbot_llms.fast_api.Fast_API__LLMs import Fast_API__LLMs
+
+osbot_llms__server_config.s3_log_requests = True
+llm_fast_api = Fast_API__LLMs().setup()
+app          = llm_fast_api.app()
+
diff --git a/start-local-server.sh b/start-local-server.sh
@@ -1 +1 @@
-uvicorn server:app --reload
+uvicorn server__fastapi__llms:app --reload --host 0.0.0.0 --port 5010
diff --git a/tests/integration/backend/s3_minio/test_S3_DB__Chat_Threads.py b/tests/integration/backend/s3_minio/test_S3_DB__Chat_Threads.py
@@ -97,7 +97,7 @@ def test_save_chat_completion__user_request(self):
                                                             'seed'            : 42            ,
                                                             'stream'          : True          ,
                                                             'system_prompts'  : None          ,
-                                                            'temperature'     : '0'           ,
+                                                            'temperature'     : 0.0           ,
                                                             'user_data'       : None          ,
                                                             'user_prompt'     : user_prompt   },
                                    "llm_request_id"     : llm_request_id                       ,
diff --git a/tests/integration/fast_api/routes/test__api__Routes__Chat.py b/tests/integration/fast_api/routes/test__api__Routes__Chat.py
@@ -27,7 +27,7 @@ def test__completion__save_chat_completion__user_request(self):
                                 "selected_model"    : "llama3-70b-8192" }
         server_name         = osbot_llms__server_config.server_name
         llm_chat_completion = LLMs__Chat_Completion(user_prompt=user_prompt, chat_thread_id=chat_thread_id, user_data=user_data)
-        json_data           = json_load(llm_chat_completion.model_dump_json())
+        json_data           = llm_chat_completion.json()
         response            = self.client.post('/chat/completion', json=json_data)      # submit chat request
         request_id          = response.headers.get('fast-api-request-id')
         when_str            = self.s3_db_chat_threads.s3_key_generator.path__for_date_time__now_utc()
@@ -63,7 +63,7 @@ def test__view(self):
                                     'seed': 42,
                                     'stream': True,
                                     'system_prompts': None,
-                                    'temperature': '0',
+                                    'temperature': 0.0,
                                     'user_data': None,
                                     'user_prompt': user_prompt}
 
diff --git a/tests/integration/fast_api/routes/test__int__Routes__Chat.py b/tests/integration/fast_api/routes/test__int__Routes__Chat.py
@@ -58,7 +58,7 @@ def test_handle_other_llms__groq(self):
 
     def test_completion(self):
         self.llm_chat_completion.stream = True
-        streaming_response              = invoke_async_function(self.routes_chat.completion(self.llm_chat_completion, self.request))
+        streaming_response              = invoke_async_function(self.routes_chat.completion(self.request, self.llm_chat_completion))
         items                           = invoke_async_function(self.collect_body_iterator(streaming_response.body_iterator))
         answer                          = ''.join(items)
 
@@ -67,20 +67,20 @@ def test_completion(self):
         assert '42'                           in answer
 
         self.llm_chat_completion.stream = False
-        response_str                    = invoke_async_function(self.routes_chat.completion(self.llm_chat_completion, self.request))
+        response_str                    = invoke_async_function(self.routes_chat.completion(self.request, self.llm_chat_completion))
         assert type(response_str) is str
         assert '42' in response_str
 
     def test_completion__openai(self):
         user_data = dict(selected_platform = 'Groq (Free)'     ,
                          selected_provider = '1. Meta'         ,
                          selected_model    = 'llama3-70b-8192')
-        self.llm_chat_completion.llm_platform = None
-        self.llm_chat_completion.llm_provider = None
-        self.llm_chat_completion.llm_model    = None
+        # self.llm_chat_completion.llm_platform = None
+        # self.llm_chat_completion.llm_provider = None
+        # self.llm_chat_completion.llm_model    = None
         self.llm_chat_completion.user_data    = user_data
         self.llm_chat_completion.stream       = True
-        streaming_response                    = invoke_async_function(self.routes_chat.completion(self.llm_chat_completion, self.request))
+        streaming_response                    = invoke_async_function(self.routes_chat.completion(self.request, self.llm_chat_completion))
         items                                 = invoke_async_function(self.collect_body_iterator(streaming_response.body_iterator))
         answer                                = ''.join(items)
 
diff --git a/tests/integration/fast_api/test_Fast_API__Chat_Threads.py b/tests/integration/fast_api/test_Fast_API__Chat_Threads.py
@@ -37,7 +37,7 @@ def test__prompt_with_system__stream__save_chat_completion__user_request(self):
 
         llm_chat_completion = LLMs__Chat_Completion(user_prompt=user_prompt, chat_thread_id=chat_thread_id,
                                                     user_data=user_data)
-        json_data           = json_load(llm_chat_completion.model_dump_json())
+        json_data           = llm_chat_completion.json()
         response            = self.client.post('chat/completion', json=json_data)
 
         assert response.status_code == 200
diff --git a/tests/integration/fast_api/test_Fast_API__via_Http.py b/tests/integration/fast_api/test_Fast_API__via_Http.py
@@ -2,6 +2,10 @@
 from fastapi                                        import FastAPI
 from osbot_fast_api.utils.Fast_API_Server           import Fast_API_Server
 from osbot_utils.context_managers.print_duration    import print_duration
+from osbot_utils.testing.Stdout import Stdout
+from osbot_utils.utils.Dev import pprint
+
+from osbot_llms.models.LLMs__Chat_Completion import LLMs__Chat_Completion
 from osbot_llms.utils.Version                       import version__osbot_llms
 from tests.llm_fast_api__for_tests                  import llm_fast_api, llm_fast_api__app
 
@@ -30,4 +34,13 @@ def test_version(self):
         with self.llm_fast_api as _:
             version__fast_api    = _.version__fast_api_server()
             assert self.fast_api_server.requests_get('/config/version').json() == {'version': version__fast_api   }
-            assert self.fast_api_server.requests_get('/info/version'  ).json() == {'version': version__osbot_llms }
+            assert self.fast_api_server.requests_get('/info/version'  ).json() == {'version': version__osbot_llms }
+
+    def test__chat__completion(self):
+        with self.llm_fast_api as _:
+            llm_chat_completion = LLMs__Chat_Completion(user_prompt='51-9')
+            response = self.fast_api_server.requests_post('/chat/completion', data=llm_chat_completion)
+            assert '42' in response.text
+
+
+
diff --git a/tests/integration/llms/chats/storage/test_Chats_Storage__S3_Minio.py b/tests/integration/llms/chats/storage/test_Chats_Storage__S3_Minio.py
@@ -43,7 +43,7 @@ def test_save_user_request(self):
             assert s3_file_contents.chat_thread_id                  == chat_thread_id
             assert s3_file_contents.llm_request_id                  == request_id
             assert s3_file_contents.llm_chat_completion.images      == []
-            assert s3_file_contents.llm_chat_completion.temperature == '0'
+            assert s3_file_contents.llm_chat_completion.temperature == 0.0
             assert s3_file_contents.llm_chat_completion.seed        == 42
             assert s3_file_contents.llm_chat_completion.user_prompt == user_prompt
 
diff --git a/tests/integration/llms/chats/test_LLM__Chat_Completion__Resolve_Engine.py b/tests/integration/llms/chats/test_LLM__Chat_Completion__Resolve_Engine.py
@@ -18,7 +18,7 @@ def test_map_provider(self):
 
         llm_chat_completion = LLMs__Chat_Completion(**kwargs)
         response = self.llm_resolve_engine.map_provider(llm_chat_completion)
-        assert response.json() == { 'llm_chat_completion': { 'chat_thread_id': None,
+        assert response.json() == { 'llm_chat_completion': { 'chat_thread_id': '',
                                                              'histories': None,
                                                              'images': [],
                                                              'llm_answer': None,
@@ -29,7 +29,7 @@ def test_map_provider(self):
                                                              'seed': 42,
                                                              'stream': True,
                                                              'system_prompts': None,
-                                                             'temperature': Decimal('0'),
+                                                             'temperature': 0.0,
                                                              'user_data': { 'selected_model': 'phi3',
                                                                             'selected_platform': 'Ollama (Local)',
                                                                             'selected_provider': 'Microsoft'},
diff --git a/tests/integration/llms/chats/test_LLM__Platform_Engine__Mistral.py b/tests/integration/llms/chats/test_LLM__Platform_Engine__Mistral.py
@@ -15,7 +15,7 @@ def setUpClass(cls):
         cls.llm_model              = MISTRAL__DEFAULT_MODEL
         cls.system_prompt          = 'just reply with the answer in one word'
         cls.user_prompt            = '1+2'
-        cls.llm_chat_completion    = LLMs__Chat_Completion(user_prompt=cls.user_prompt, system_prompt=cls.system_prompt)
+        cls.llm_chat_completion    = LLMs__Chat_Completion(user_prompt=cls.user_prompt, system_prompts=[cls.system_prompt])
         cls.kwargs                 = dict(llm_platform=cls.llm_platform, llm_provider=cls.llm_provider, llm_model=cls.llm_model, llm_chat_completion=cls.llm_chat_completion)
         cls.llm_engine_groq        = LLM__Platform_Engine__Mistral(**cls.kwargs)
 
diff --git a/tests/integration/llms/test_API_Perplexity.py b/tests/integration/llms/test_API_Perplexity.py
@@ -5,7 +5,7 @@
 from osbot_llms.models.GPT_History import GPT_History
 
 
-#@pytest.mark.skip("Re-enable when adding offical support to TCB to the Perplexity API")
+#@pytest.mark.skip("Re-enable when adding ofical support to TCB to the Perplexity API")
 class test_API_Perplexity(TestCase):
 
     def setUp(self):
@@ -33,15 +33,14 @@ def test_ask_using_messages__async_mode(self):
     def test_ask_using_messages__sync_mode(self):
         model             = "mistral-7b-instruct"
         system_prompt     = "act like a calculator, just reply with the answer"
-        user_prompt       = "40+2"
-        expected_response = '42'
+        user_prompt       = "1+2"
         async_mode        = False
         messages = [ { "role"    : "system"      ,
                        "content" : system_prompt },
                      { "role"    : "user",
                        "content" : user_prompt}]
         response  = self.api_perplexity.ask_using_messages(messages, model=model, async_mode=async_mode)
-        assert '42' in response
+        assert '3' in response
 
 
     def test_ask_using_system_prompts(self):

Original file line number	Diff line number	Diff line change
`@@ -0,0 +1,3 @@`
	`1`	`+[submodule "modules/OSBot-Fast-API"]`
	`2`	`+ path = modules/OSBot-Fast-API`
	`3`	`+ url = [email protected]:owasp-sbot/OSBot-Fast-API.git`
Original file line number	Diff line number	Diff line change
`@@ -1 +1 @@`
`1`		`-uvicorn server:app --reload`
	`1`	`+uvicorn server__fastapi__llms:app --reload --host 0.0.0.0 --port 5010`