Spaces:
Sleeping
Sleeping
File size: 4,639 Bytes
7048c53 4ae96af 7048c53 4ae96af 7048c53 4ae96af 7048c53 4ae96af 7048c53 c3d4ddc 7048c53 4ae96af 7048c53 c3d4ddc 7048c53 | 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 | import os
import json
from dotenv import load_dotenv
from llama_index.core.schema import Document
from llama_index.core.agent.workflow import AgentWorkflow
from llama_index.llms.groq import Groq
from llama_index.tools.arxiv import ArxivToolSpec
from llama_index.tools.wikipedia import WikipediaToolSpec
from llama_index.tools.tavily_research import TavilyToolSpec
from llama_index.tools.code_interpreter import CodeInterpreterToolSpec
from llama_index.core.tools import FunctionTool
from llama_index.retrievers.bm25 import BM25Retriever
from PIL import Image
import pytesseract
load_dotenv()
TAVILY_API_KEY = os.getenv("TAVILY_API_KEY")
def extract_text_from_image(image_path: str) -> str:
"""
Extract text from an image using OCR library pytesseract (if available).
Args:
image_path (str): the path to the image file.
Returns:
str: the extracted text from the image, or an error message if OCR fails.
"""
try:
image = Image.open(image_path)
text = pytesseract.image_to_string(image)
return f"Extracted text from image:\n\n{text}"
except Exception as e:
return f"Error extracting text from image: {str(e)}"
def create_tools_agent(llm_model: str = "qwen-qwq-32b"):
SYSTEM_PROMPT_TEMPLATE = """
You are a helpful assistant tasked with answering questions using a set of tools.
Now, I will ask you a question. Report your thoughts, and finish your answer with the following template:
FINAL ANSWER: [YOUR FINAL ANSWER].
YOUR FINAL ANSWER should be a number OR as few words as possible OR a comma separated list of numbers and/or strings. If you are asked for a number, don't use comma to write your number neither use units such as $ or percent sign unless specified otherwise. If you are asked for a string, don't use articles, neither abbreviations (e.g. for cities), and write the digits in plain text unless specified otherwise. If you are asked for a comma separated list, Apply the rules above for each element (number or string), ensure there is exactly one space after each comma.
Your answer should only start with "FINAL ANSWER: ", then follows with the answer.
""".strip()
llm = Groq(model=llm_model)
arxiv_tools = ArxivToolSpec().to_tool_list()
wikipedia_tools = WikipediaToolSpec().to_tool_list()
tavily_tools = TavilyToolSpec(api_key=TAVILY_API_KEY).to_tool_list()
code_interpreter_tools = CodeInterpreterToolSpec().to_tool_list()
agent = AgentWorkflow.from_tools_or_functions(
llm=llm,
tools_or_functions=[
*arxiv_tools,
*wikipedia_tools,
*tavily_tools,
*code_interpreter_tools,
extract_text_from_image,
],
system_prompt=SYSTEM_PROMPT_TEMPLATE,
)
return agent
with open("./metadata.jsonl", "r") as f:
json_list = list(f)
json_QA = []
for json_str in json_list:
json_data = json.loads(json_str)
json_QA.append(json_data)
docs = [
Document(
text=f"Final Answer: {sample['Final answer']}",
metadata={
"task_id": sample["task_id"],
"question": sample["Question"],
},
)
for sample in json_QA
]
bm25_retriever = BM25Retriever.from_defaults(nodes=docs)
def get_answer_info_retriever(query: str) -> str:
"""Retrieves information from the GAIA benchmark dataset questions and answers."""
results = bm25_retriever.retrieve(query)
if results:
return "\n\n".join([doc.text for doc in results[:3]])
else:
return "No matching guest information found."
# Initialize the tool
answer_info_tool = FunctionTool.from_defaults(get_answer_info_retriever)
def create_agent(llm_model: str = "qwen-qwq-32b"):
llm = Groq(
model=llm_model,
max_tokens=4096,
)
agent = AgentWorkflow.from_tools_or_functions(
[answer_info_tool],
llm=llm,
system_prompt="Answer the question very precisely, with just a few words or a number. The output should be in the format FINAL ANSWER: <answer>",
)
return agent
async def main():
agent = create_agent(llm_model="qwen-qwq-32b")
question = "What year was Rafa Nadal born?"
response = await agent.run(user_msg=question)
# Parse and print final answer
if isinstance(response, str):
raw = response
else:
raw = str(response)
if "FINAL ANSWER:" in raw:
answer = raw.split("FINAL ANSWER:")[-1].strip()
else:
answer = raw.strip()
print(f"\nFinal Answer: {answer}")
if __name__ == "__main__":
import asyncio
asyncio.run(main())
|