@@ -312,7 +312,7 @@ async def stream_generator():
312312 # Non-streaming logic
313313 spinner .text = f"Waiting for { model }
314314 response : OpenAIChatCompletion = await client .chat .completions .create (** request_params )
315-
315+
316316 content = response .choices [0 ].message .content
317317 spinner .succeed ("Request completed" )
318318
@@ -385,6 +385,7 @@ async def wrapper(
385385 gpt_4o_mini = custom_model ("gpt-4o-mini" )
386386 o1_mini = custom_model ("o1-mini" )
387387 o1_preview = custom_model ("o1-preview" )
388+ gpt_4_5_preview = custom_model ("gpt-4.5-preview" )
388389
389390
390391class AnthropicModels :
@@ -687,7 +688,7 @@ async def stream_generator():
687688 return compressed_content , None
688689 except ValueError as e :
689690 return "" , e
690-
691+
691692 # For non-JSON responses, keep original formatting but make single line
692693 logger .debug (f"[LLM] API Response: { ' ' .join (content .strip ().splitlines ())} )
693694 return content .strip (), None
@@ -727,12 +728,14 @@ async def wrapper(
727728 haiku_3_5 = custom_model ("anthropic/claude-3.5-haiku" )
728729 sonnet = custom_model ("anthropic/claude-3-sonnet" )
729730 sonnet_3_5 = custom_model ("anthropic/claude-3.5-sonnet" )
731+ sonnet_3_7 = custom_model ("anthropic/claude-3.7-sonnet" )
730732 opus = custom_model ("anthropic/claude-3-opus" )
731733 gpt_3_5_turbo = custom_model ("openai/gpt-3.5-turbo" )
732734 gpt_4_turbo = custom_model ("openai/gpt-4-turbo" )
733735 gpt_4 = custom_model ("openai/gpt-4" )
734736 gpt_4o = custom_model ("openai/gpt-4o" )
735737 gpt_4o_mini = custom_model ("openai/gpt-4o-mini" )
738+ gpt_4_5_preview = custom_model ("openai/gpt-4.5-preview" )
736739 o1_preview = custom_model ("openai/o1-preview" )
737740 o1_mini = custom_model ("openai/o1-mini" )
738741 gemini_flash_1_5 = custom_model ("google/gemini-flash-1.5" )
@@ -837,7 +840,7 @@ async def stream_generator():
837840 options = {"temperature" : temperature , "num_predict" : max_tokens },
838841 stream = True ,
839842 )
840-
843+
841844 for chunk in response :
842845 if chunk and "message" in chunk and "content" in chunk ["message" ]:
843846 content = chunk ["message" ]["content" ]
@@ -1012,7 +1015,7 @@ async def stream_generator():
10121015 return compressed_content , None
10131016 except ValueError as e :
10141017 return "" , e
1015-
1018+
10161019 # For non-JSON responses, keep original formatting but make single line
10171020 logger .debug (f"[LLM] API Response: { ' ' .join (content .strip ().splitlines ())} )
10181021 return content .strip (), None
@@ -1084,7 +1087,7 @@ async def send_together_request(
10841087 content = []
10851088 if isinstance (image_data , str ):
10861089 image_data = [image_data ]
1087-
1090+
10881091 for i , image in enumerate (image_data , start = 1 ):
10891092 content .append ({"type" : "text" , "text" : f"Image { i } })
10901093 if image .startswith (("http://" , "https://" )):
@@ -1097,7 +1100,7 @@ async def send_together_request(
10971100 "type" : "image_url" ,
10981101 "image_url" : {"url" : f"data:image/jpeg;base64,{ image } }
10991102 })
1100-
1103+
11011104 # Add original text content
11021105 content .append ({"type" : "text" , "text" : last_user_msg ["content" ]})
11031106 last_user_msg ["content" ] = content
@@ -1120,19 +1123,19 @@ async def stream_generator():
11201123 response_format = {"type" : "json_object" } if require_json_output else None ,
11211124 stream = True
11221125 )
1123-
1126+
11241127 for chunk in response :
11251128 if chunk .choices [0 ].delta .content :
11261129 content = chunk .choices [0 ].delta .content
11271130 full_message += content
11281131 yield content
11291132 logger .debug ("Stream complete" )
11301133 logger .debug (f"Full message: { full_message } )
1131- yield "\n "
1134+ yield "\n "
11321135 except Exception as e :
11331136 logger .error (f"An error occurred during streaming: { e } )
11341137 yield ""
1135- yield "\n "
1138+ yield "\n "
11361139
11371140 return stream_generator ()
11381141
@@ -1148,7 +1151,7 @@ async def stream_generator():
11481151
11491152 content = response .choices [0 ].message .content
11501153 spinner .succeed ("Request completed" )
1151-
1154+
11521155 # Compress the response to single line if it's JSON
11531156 if require_json_output :
11541157 try :
@@ -1158,7 +1161,7 @@ async def stream_generator():
11581161 return compressed_content , None
11591162 except ValueError as e :
11601163 return "" , e
1161-
1164+
11621165 # For non-JSON responses, keep original formatting but make single line
11631166 logger .debug (f"[LLM] API Response: { ' ' .join (content .strip ().splitlines ())} )
11641167 return content .strip (), None
@@ -1216,7 +1219,7 @@ async def send_gemini_request(
12161219 """
12171220 # Create spinner only once at the start
12181221 spinner = Halo (text = f"Sending request to Gemini ({ model } , spinner = "dots" )
1219-
1222+
12201223 try :
12211224 # Start spinner
12221225 spinner .start ()
@@ -1229,7 +1232,7 @@ async def send_gemini_request(
12291232 "temperature" : temperature ,
12301233 "max_output_tokens" : max_tokens ,
12311234 }
1232-
1235+
12331236 if require_json_output :
12341237 generation_config .update ({
12351238 "response_mime_type" : "application/json"
@@ -1247,15 +1250,15 @@ async def send_gemini_request(
12471250 last_user_message = next ((msg ["content" ] for msg in reversed (messages ) if msg ["role" ] == "user" ), "" )
12481251 full_message = ""
12491252 logger .debug ("Stream started" )
1250-
1253+
12511254 try :
12521255 response = model_instance .generate_content (last_user_message , stream = True )
12531256 for chunk in response :
12541257 if chunk .text :
12551258 content = chunk .text
12561259 full_message += content
12571260 yield content
1258-
1261+
12591262 logger .debug ("Stream complete" )
12601263 logger .debug (f"Full message: { full_message } )
12611264 except Exception as e :
@@ -1264,13 +1267,13 @@ async def send_gemini_request(
12641267 else :
12651268 # Non-streaming: Use chat format
12661269 chat = model_instance .start_chat (history = [])
1267-
1270+
12681271 # Process messages and images
12691272 if messages :
12701273 for msg in messages :
12711274 role = msg ["role" ]
12721275 content = msg ["content" ]
1273-
1276+
12741277 if role == "user" :
12751278 if image_data and msg == messages [- 1 ]:
12761279 parts = []
@@ -1508,7 +1511,7 @@ async def stream_generator():
15081511 # Non-streaming logic
15091512 spinner .text = f"Waiting for { model }
15101513 response = await client .chat .completions .create (** request_params )
1511-
1514+
15121515 if model == "deepseek-reasoner" :
15131516 reasoning = response .choices [0 ].message .reasoning_content
15141517 content = response .choices [0 ].message .content
@@ -1564,7 +1567,7 @@ async def wrapper(
15641567 )
15651568
15661569 return wrapper
1567-
1570+
15681571 # Model-specific methods using custom_model
15691572 chat = custom_model ("deepseek-chat" )
15701573 reasoner = custom_model ("deepseek-reasoner" )
0 commit comments