Spaces:
Sleeping
Sleeping
File size: 5,902 Bytes
e646764 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 |
import json
from typing import List, Dict, Any, Union
from gradio.components.chatbot import ChatMessage
from llm import LLM_Client
import os
from config import session_keys
import logging
logging.basicConfig(level=logging.INFO)
logger = logging.getLogger(f"🤖 {__name__}")
class LLMNode:
def __init__(self, session_id):
self.session_id = session_id
settings = session_keys.get(session_id, {})
self.provider = settings.get("provider", "OpenAI")
self.tool_model = settings.get("tool_call_model", "gpt-4o-mini")
self.response_model = settings.get("response_model", "gpt-4o-mini")
logger.info(f"Provider: {self.provider}")
logger.info(f"Tool_model: {self.tool_model}")
logger.info(f"Response_model: {self.response_model}")
self.tool_selector = LLM_Client(session_id, sourceAI="openai")
self.generator = None
if self.provider.lower() == "openai":
if settings.get("OPENAI_API_KEY"):
self.generator = LLM_Client(session_id, sourceAI="openai")
elif self.provider.lower() == "nebius":
if settings.get("NEBIUS_API_KEY"):
self.generator = LLM_Client(session_id, sourceAI="nebius")
def build_prompt(
self,
history: List[Union[Dict[str, Any], ChatMessage]],
message: str,
image_base64: str = None,
vision_enabled: bool = False,
type: str = "generate",
encryptId: str = None,
history_len = 20,
face_data: dict = None,
color_season: str = None
) -> List[Dict[str, Any]]:
prompts = []
chat_prompt = []
for msg in history:
if isinstance(msg, ChatMessage):
role, content = msg.role, msg.content
else:
role, content = msg.get("role"), msg.get("content")
if role in ["user", "assistant", "system"]:
chat_prompt.append({"role": role, "content": content})
user_data_str = ""
if face_data:
user_data_str += f"User facial parts' color: {face_data}.\n"
if color_season:
user_data_str += f"User color season: {color_season}\n"
if encryptId:
user_data_str += f"tmp_id: {encryptId}"
user_message = ""
if user_data_str != "":
user_message += f"UserData info: {user_data_str}\n\n"
user_message += message
if vision_enabled:
if image_base64:
chat_prompt.append({
"role": "user",
"content": [
{"type": "text", "text": user_message},
{"type": "image_url", "image_url": {"url": f"data:image/png;base64,{image_base64}"}}
]
})
else:
chat_prompt.append({
"role": "user",
"content": [
{"type": "text", "text": user_message}
]
})
else:
chat_prompt.append({"role": "user", "content": message})
# Construct Main System Prompt
text_prompt_path = "promptsDB"
with open(os.path.join(text_prompt_path, "system_prompt.txt"), "r", encoding="utf-8") as f:
system_prompt_str = f.read().strip()
system_prompt = {
"role": "system",
"content": system_prompt_str
}
prompts.append(system_prompt)
# Construct User Data
logger.info(f"User Message: \n{user_message}")
logger.info(f"Image Exist: {True if image_base64 else False}")
chat_prompt
prompts = prompts + chat_prompt[-history_len:]
return prompts
def call_tool_step(
self,
messages: List[Union[Dict[str, str], Dict[str, Any]]],
tools: List[Dict[str, Any]]
) -> Dict[str, Any]:
return self.tool_selector.get_completion(
model=self.tool_model,
max_tokens=300,
messages=messages,
tools=tools,
tool_choice="auto"
)
def call_final_step(self, messages: List[Dict[str, str]]) -> Dict[str, Any]:
return self.generator.get_completion(
model=self.response_model,
max_tokens=300,
messages=messages,
tool_choice="none"
)
def sanitize_messages_for_nebius(self, messages: List[Dict[str, Any]]) -> List[Dict[str, str]]:
sanitized = []
for msg in messages:
if msg["role"] in ["user", "assistant", "system"]:
sanitized.append({"role": msg["role"], "content": msg.get("content", "")})
elif msg["role"] == "tool":
content = msg.get("content", "")
sanitized.append({"role": "assistant", "content": f"Tool '{msg['name']}' returned:\n{content}"})
return sanitized
def call_generation_step(
self,
message: str,
history: List[Union[Dict[str, Any], ChatMessage]],
tool_result: str = None,
face_data: dict = None,
color_season: str = None
) -> List[Dict[str, Any]]:
messages = self.build_prompt(
history=history,
message=message,
image_base64=None,
vision_enabled=False,
face_data=face_data,
color_season=color_season
)
if tool_result:
messages.append({
"role": "assistant",
"content": f"Information (in your system) - use this if you required to answer that: \n\n{tool_result}"
})
step2 = self.call_final_step(messages)
return [{"role": "assistant", "content": step2["choices"][0]["message"]["content"]}]
|