Update utils/llms.py

utils/llms.py

@@ -1,18 +1,27 @@
 import os
 import helpers.helper as helper
 
-import google.generativeai as genai
 from g4f.client import Client
 from litellm import completion
 import random
 import json
+import os
+# from dotenv import load_dotenv
 
-from g4f.Provider import DeepInfraChat,Glider,Liaobots,Blackbox,ChatGptEs,LambdaChat,TypeGPT
+# load_dotenv()
+from g4f.Provider import DeepInfraChat,Glider,LambdaChat,TypeGPT
 
 gemini_api_keys=json.loads(os.environ.get("GEMINI_KEY_LIST"))
-groq_api_keys=["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","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","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"]
+groq_api_keys=json.loads(os.environ.get("GROQ_API_KEYS"))
+chutes_key=os.environ.get("CHUTES_API_KEY")
+github_key=os.environ.get("GITHUB_API_KEY")
+
 DeepInfraChat.models = ["google/gemma-3-27b-it","deepseek-ai/DeepSeek-R1-Turbo","Qwen/QwQ-32B","deepseek-ai/DeepSeek-R1","deepseek-ai/DeepSeek-V3-0324","meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8","meta-llama/Llama-4-Scout-17B-16E-Instruct","microsoft/Phi-4-multimodal-instruct"]
+
 deepinframodels=["meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8","microsoft/Phi-4-multimodal-instruct","google/gemma-3-27b-it","meta-llama/Llama-4-Scout-17B-16E-Instruct"]
+chutes_models={"MAI-DS-R1-FP8":"microsoft/MAI-DS-R1-FP8","DeepSeek-V3-0324":"deepseek-ai/DeepSeek-V3-0324","deepseek-reasoner":"deepseek-ai/DeepSeek-R1","GLM-4-32B-0414":"THUDM/GLM-4-32B-0414","GLM-Z1-32B-0414":"THUDM/GLM-Z1-32B-0414"}
+github_models={"gpt4.1":"gpt-4.1","gpt-4o":"gpt-4o","o4-mini":"o4-mini"}
+
 REASONING_CORRESPONDANCE = {"DeepSeek-R1-Glider":Glider, "DeepSeekR1-LAMBDA":LambdaChat,"DeepSeekR1":DeepInfraChat,"deepseek-slow":TypeGPT}
 os.environ["GEMINI_API_KEY"] =random.choice(gemini_api_keys)
 
@@ -26,20 +35,6 @@ clienty = OpenAI(
 CHAT_CORRESPONDANCE = {"DeepSeek-V3":DeepInfraChat}
 
 client = Client()
-genai.configure(api_key="AIzaSyAQgAtQPpY0bQaCqCISGxeyF6tpDePx-Jg")
-modell = genai.GenerativeModel('gemini-1.5-pro')
-generation_config = {
-  "temperature": 1,
-  "top_p": 0.95,
-  "top_k": 40,
-  "max_output_tokens": 8192,
-  "response_mime_type": "text/plain",
-}
-
-model2flash = genai.GenerativeModel(
-  model_name="gemini-2.0-flash-thinking-exp",
-  generation_config=generation_config,
-)
 
         
 
@@ -54,20 +49,12 @@ def clear():
         helper.q.task_done()
 
 def gpt4(messages,response_format,model="gpt-4"):
-    if response_format!=None:
-        output = clienty.chat.completions.create(
-        model="google/gemini-2.0-flash-001",
-        messages= messages,
-        response_format=response_format,  
-        )
-        return str(output.choices[0].message.content)
-    else:
-        print(messages)
         if len(messages) ==1:
             messages[0]["role"]="user"
         response = completion(
             model="gemini/gemini-2.0-flash", 
-            messages=messages
+            messages=messages,
+            response_format=response_format
         )
         return str(response.choices[0].message.content)
 
@@ -87,7 +74,6 @@ def gpt4stream(messages,model,api_keys):
                 messages=messages,
                 stream=True
 
-                # Add any other necessary parameters
             )
             for part in response:
                 cunk=cunk+(part.choices[0].delta.content or "")
@@ -97,7 +83,6 @@ def gpt4stream(messages,model,api_keys):
                     clear()
                     break 
         except Exception as e:
-            
             pass
         helper.q.put_nowait("RESULT: "+cunk)
 
@@ -128,7 +113,7 @@ def gpt4stream(messages,model,api_keys):
                 break 
         helper.q.put_nowait("RESULT: "+cunk)
 
-    elif model=="deepseek-r1-distill-llama-70b":
+    elif model=="groq/deepseek-r1-distill-llama-70b":
         os.environ["GROQ_API_KEY"] =random.choice(groq_api_keys)
 
         response = completion(model="groq/deepseek-r1-distill-llama-70b", messages=messages, stream=True)
@@ -142,7 +127,7 @@ def gpt4stream(messages,model,api_keys):
                 clear()
                 break 
         helper.q.put_nowait("RESULT: "+cunk)
-    elif model=="qwq-32b":
+    elif model=="groq/qwq-32b":
         os.environ["GROQ_API_KEY"] =random.choice(groq_api_keys)
         response = completion(model="groq/qwen-qwq-32b", messages=messages, stream=True)
 
@@ -169,34 +154,40 @@ def gpt4stream(messages,model,api_keys):
                 clear()
                 break 
         helper.q.put_nowait("RESULT: "+cunk)
+    elif model in chutes_models:
+        response = completion(model=f"openai/{chutes_models[model]}",api_key=chutes_key,base_url="https://llm.chutes.ai/v1", messages=messages, stream=True)
 
-    elif model=="gemini-2.0-flash-thinking-exp-01-21":
-        for key in gemini_api_keys:
-            try:
-                os.environ["GEMINI_API_KEY"] =key
+        cunk=""
+        for part in response:
+            cunk=cunk+(part.choices[0].delta.content or "")
+            
+            if "```json" not in cunk:
+                helper.q.put_nowait(part.choices[0].delta.content or "")
+            if helper.stopped: 
+                clear()
+                break 
+        helper.q.put_nowait("RESULT: "+cunk)
 
-                response = completion(model="gemini/gemini-2.0-flash-thinking-exp-01-21", messages=messages, stream=True)
+    elif model in github_models:
+        response = completion(model=f"github/{github_models[model]}",api_key=github_key, messages=messages, stream=True)
 
-                cunk=""
-                for part in response:
-                    cunk=cunk+(part.choices[0].delta.content or "")
-                    if "```json" not in cunk:
-                        helper.q.put_nowait(part.choices[0].delta.content or "")
-                    if helper.stopped: 
-                        clear()
-                        break 
-                break         
-            except Exception as e:
-                helper.q.put_nowait(str(e))
-                pass            
+        cunk=""
+        for part in response:
+            cunk=cunk+(part.choices[0].delta.content or "")
+            
+            if "```json" not in cunk:
+                helper.q.put_nowait(part.choices[0].delta.content or "")
+            if helper.stopped: 
+                clear()
+                break 
         helper.q.put_nowait("RESULT: "+cunk)
 
-    elif model=="gemini-2.5-pro-exp-03-25" :
+    elif "gemini" in model: 
         for key in gemini_api_keys:
             try:
                 os.environ["GEMINI_API_KEY"] =key
 
-                response = completion(model="gemini/gemini-2.5-pro-exp-03-25", messages=messages, stream=True)
+                response = completion(model=f"gemini/{model}", messages=messages, stream=True)
 
                 cunk=""
                 for part in response:
@@ -251,7 +242,7 @@ def gpt4stream(messages,model,api_keys):
         helper.q.put_nowait("RESULT: "+cunk)
 
 
-    elif model=="deepseek-reasoner" :
+    elif model=="qwq-32b" :
         helper.q.put_nowait("<think>")
         cunk=""
         providers=REASONING_QWQ