README.md

metadata

title: Extract PDF content tool from file upload
emoji: 🗂️
colorFrom: green
colorTo: pink
sdk: gradio
sdk_version: 5.11.0
app_file: app.py
pinned: true
tags:
  - tool
  - pdf

Use this tool for extracting data from pdf.

Small example:

pdf_extraction_tool = Tool.from_space(
    "matterattetatte/pdf-extractor-tool",
    name="pdf-extractor",
    description="Extract data"
)

pdf_extraction_tool("Extract all headlines from all pdfs in folder pdfs")

Full-fledged example (managed and managing agents):

from smolagents import CodeAgent, HfApiModel, ManagedAgent, ToolCallingAgent, GradioUI, Tool
from huggingface_hub import login
import os

login('hf_*******')

pdf_agent = ToolCallingAgent(
    tools=[Tool.from_space("matterattetatte/pdf-extractor-tool", name="pdf-extractor", description="Extract data")],
    model=HfApiModel(),
    max_steps=4,
)

managed_pdf_agent = ManagedAgent(
    agent=pdf_agent,
    name="extraction",
    description="Returns the content of pdf files in a string. Give it your path as an argument. Also, this agent should link to the files it is are using.",
)

manager_agent = CodeAgent(
    tools=[],
    model=HfApiModel(model_id = "Qwen/Qwen2.5-Coder-32B-Instruct"),
    managed_agents=[managed_pdf_agent],
    additional_authorized_imports=['os', 're'],
)

manager_agent.run("Read file pdfs/my_file.pdf and summarize its content for me. I want to understand how to do things")

GradioUI(manager_agent).launch()