File size: 13,264 Bytes
60058de |
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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 |
import json
from typing import Dict, Iterator, List, Literal, Optional, Tuple, Union
from qwen_agent.agents.fncall_agent import FnCallAgent
from qwen_agent.llm import BaseChatModel
from qwen_agent.llm.schema import ASSISTANT, DEFAULT_SYSTEM_MESSAGE, Message
from qwen_agent.settings import MAX_LLM_CALL_PER_RUN
from qwen_agent.tools import BaseTool
from qwen_agent.utils.utils import format_as_text_message, merge_generate_cfgs
from openai import OpenAI
import time
TOOL_DESC = (
'{name_for_model}: Call this tool to interact with the {name_for_human} API. '
'What is the {name_for_human} API useful for? {description_for_model} Parameters: {parameters} {args_format}')
PROMPT_REACT = """Digging through the buttons to find quailty sources and the right information. You have access to the following tools:
{tool_descs}
Use the following format:
Question: the input question you must answer
Thought: you should always think about what to do
Action: the action to take, should be one of [{tool_names}]
Action Input: the input to the action
Observation: the result of the action
Action: the action to take, should be one of [{tool_names}]
Action Input: the input to the action
Observation: the result of the action
... (this Thought/Action/Action Input/Observation can be repeated zero or more 20 times)
Notice:
- You must take action at every step. When you take action, you must use the tool with the correct format and output the action input.
- When you can not find the information you need, you should visit page of previous page recursively until you find the information you need.
- You can not say "I'm sorry, but I cannot assist with this request."!!! You must explore.
- You do not need to provide the final answer, but you must explore.
- Action Input should be valid JSON.
Begin!
{query}
"""
# def call_(self, **kwargs):
# user_prompt = self.template.format(
# query = kwargs["query"],
# document = kwargs["document"]
# )
# messages = []
# messages.append({"role": "system", "content": self.system})
# messages.append({"role": "user", "content": user_prompt})
# response = litellm.completion(
# model = "gpt-4o",
# response_format={"type": "json_object"},
# messages=messages
# )
# try:
# return eval(response.choices[0].message.content)
class Seeker(FnCallAgent):
"""This agent use ReAct format to call tools"""
def __init__(self,
function_list: Optional[List[Union[str, Dict, BaseTool]]] = None,
llm: Optional[Union[Dict, BaseChatModel]] = None,
system_message: Optional[str] = DEFAULT_SYSTEM_MESSAGE,
name: Optional[str] = None,
description: Optional[str] = None,
files: Optional[List[str]] = None,
**kwargs):
super().__init__(function_list=function_list,
llm=llm,
system_message=system_message,
name=name,
description=description,
files=files,
**kwargs)
self.extra_generate_cfg = merge_generate_cfgs(
base_generate_cfg=self.extra_generate_cfg,
new_generate_cfg={'stop': ['Observation:', 'Observation:\n']},
)
self.client = OpenAI(
api_key=llm['api_key'],
base_url=llm['model_server'],
)
self.llm_cfg = llm
self.momery = []
def observation_information_extraction(self, query, observation):
SYSTEM_PROMPT = """You are an information extraction agent. Your task is to analyze the given observation and extract information relevant to the current query. You need to decide if the observation contains useful information for the query. If it does, return a JSON object with a "usefulness" value of true and an "information" field with the relevant details. If not, return a JSON object with a "usefulness" value of false.
**Input:**
- Query: "<Query>"
- Observation: "<Current Observation>"
**Output (JSON):**
{
"usefulness": true,
"information": "<Extracted Useful Information> using string format"
}
Or, if the observation does not contain useful information:
{
"usefulness": false
}
Only respond with valid JSON.
"""
user_prompt = "- Query: {query}\n- Observation: {observation}".format(query=query, observation=observation)
# print(user_prompt)
messages = [
{'role': 'system', 'content': SYSTEM_PROMPT},
{'role': 'user', 'content': user_prompt}]
max_retries = 10
for attempt in range(max_retries):
try:
response = self.client.chat.completions.create(
model=self.llm_cfg['model'],
response_format={"type": "json_object"},
messages=messages
)
print(response.choices[0].message.content)
# response_content = json.loads(response.choices[0].message.content)
if "true" in response.choices[0].message.content:
try:
return json.loads(response.choices[0].message.content)["information"]
except:
return response.choices[0].message.content
else:
return None
except Exception as e:
print(e)
if attempt < max_retries - 1:
time.sleep(1 * (2 ** attempt)) # Exponential backoff
else:
raise e # Raise the exception if the last retry fails
def critic_information(self, query, memory):
SYSTEM_PROMPT = """You are a query answering agent. Your task is to evaluate whether the accumulated useful information is sufficient to answer the current query. If it is sufficient, return a JSON object with a "judge" value of true and an "answer" field with the answer. If the information is insufficient, return a JSON object with a "judge" value of false.
**Input:**
- Query: "<Query>"
- Accumulated Information: "<Accumulated Useful Information>"
**Output (JSON):**
{
"judge": true,
"answer": "<Generated Answer> using string format"
}
Or, if the information is insufficient to answer the query:
{
"judge": false
}
Only respond with valid JSON.
"""
memory = "-".join(memory)
user_prompt = "- Query: {query}\n- Accumulated Information: {memory}".format(query = query, memory=memory)
messages = [
{'role': 'system', 'content': SYSTEM_PROMPT},
{'role': 'user', 'content': user_prompt}]
response = self.client.chat.completions.create(
model=self.llm_cfg['model'],
response_format={"type": "json_object"},
messages=messages
)
max_retries = 10
for attempt in range(max_retries):
try:
response = self.client.chat.completions.create(
model=self.llm_cfg['model'],
response_format={"type": "json_object"},
messages=messages
)
print(response.choices[0].message.content)
if "true" in response.choices[0].message.content:
try:
return json.loads(response.choices[0].message.content)["answer"]
except:
return response.choices[0].message.content
else:
return None
except Exception as e:
print(e)
if attempt < max_retries - 1:
time.sleep(1 * (2 ** attempt)) # Exponential backoff
else:
raise e # Raise the exception if the last retry fails
def _run(self, messages: List[Message], lang: Literal['en', 'zh'] = 'en', **kwargs) -> Iterator[List[Message]]:
text_messages = self._prepend_react_prompt(messages, lang=lang)
# print("==========================")
# print(text_messages)
# print("==========================")
num_llm_calls_available = MAX_LLM_CALL_PER_RUN
response: str = 'Thought: '
query = self.llm_cfg["query"]
action_count = self.llm_cfg.get("action_count", MAX_LLM_CALL_PER_RUN)
num_llm_calls_available = action_count
while num_llm_calls_available > 0:
num_llm_calls_available -= 1
# print(num_llm_calls_available)
# Display the streaming response
output = []
for output in self._call_llm(messages=text_messages):
if output:
yield [Message(role=ASSISTANT, content=output[-1].content)]
# print(output)
# Accumulate the current response
if output:
response += output[-1].content
has_action, action, action_input, thought = self._detect_tool("\n"+output[-1].content)
if not has_action:
if "Final Answer: " in output[-1].content:
break
else:
continue
# Add the tool result
query = self.llm_cfg["query"]
observation = self._call_tool(action, action_input, messages=messages, **kwargs)
stage1 = self.observation_information_extraction(query, observation)
if stage1:
self.momery.append(stage1+"\n")
if len(self.momery) > 1:
yield [Message(role=ASSISTANT, content= "Memory:\n" + "-".join(self.momery)+"\"}")]
else:
yield [Message(role=ASSISTANT, content= "Memory:\n" + "-" + self.momery[0]+"\"}")]
stage2 = self.critic_information(query, self.momery)
if stage2:
response = f'Final Answer: {stage2}'
yield [Message(role=ASSISTANT, content=response)]
break
observation = f'\nObservation: {observation}\nThought: '
response += observation
# yield [Message(role=ASSISTANT, content=response)]
if (not text_messages[-1].content.endswith('\nThought: ')) and (not thought.startswith('\n')):
# Add the '\n' between '\nQuestion:' and the first 'Thought:'
text_messages[-1].content += '\n'
if action_input.startswith('```'):
# Add a newline for proper markdown rendering of code
action_input = '\n' + action_input
text_messages[-1].content += thought + f'\nAction: {action}\nAction Input: {action_input}' + observation
# print(text_messages[-1].content)
def _prepend_react_prompt(self, messages: List[Message], lang: Literal['en', 'zh']) -> List[Message]:
tool_descs = []
for f in self.function_map.values():
function = f.function
name = function.get('name', None)
name_for_human = function.get('name_for_human', name)
name_for_model = function.get('name_for_model', name)
assert name_for_human and name_for_model
args_format = function.get('args_format', '')
tool_descs.append(
TOOL_DESC.format(name_for_human=name_for_human,
name_for_model=name_for_model,
description_for_model=function['description'],
parameters=json.dumps(function['parameters'], ensure_ascii=False),
args_format=args_format).rstrip())
tool_descs = '\n\n'.join(tool_descs)
tool_names = ','.join(tool.name for tool in self.function_map.values())
text_messages = [format_as_text_message(m, add_upload_info=True, lang=lang) for m in messages]
text_messages[-1].content = PROMPT_REACT.format(
tool_descs=tool_descs,
tool_names=tool_names,
query=text_messages[-1].content,
)
return text_messages
def _detect_tool(self, text: str) -> Tuple[bool, str, str, str]:
special_func_token = '\nAction:'
special_args_token = '\nAction Input:'
special_obs_token = '\nObservation:'
func_name, func_args = None, None
i = text.rfind(special_func_token)
j = text.rfind(special_args_token)
k = text.rfind(special_obs_token)
if 0 <= i < j: # If the text has `Action` and `Action input`,
if k < j: # but does not contain `Observation`,
# then it is likely that `Observation` is ommited by the LLM,
# because the output text may have discarded the stop word.
text = text.rstrip() + special_obs_token # Add it back.
k = text.rfind(special_obs_token)
func_name = text[i + len(special_func_token):j].strip()
func_args = text[j + len(special_args_token):k].strip()
text = text[:i] # Return the response before tool call, i.e., `Thought`
return (func_name is not None), func_name, func_args, text
|