UpToDateAI is a Python package designed to fetch and provide the latest documentation about recently released programming frameworks to AI models. This package helps bridge the gap between AI model training cut-off dates and the latest developments in the programming world.
You can install UpToDateAI using pip:
pip install uptodateaiURL of the website you want to crawl:
from uptodateai import process_docs
process_docs("https://docs.fastht.ml/")This will crawl the specified website and save the content as Markdown files in a docs directory.
- Web crawling using Scrapy
- Content extraction using newspaper3k
- HTML to Markdown conversion
- Automatic directory structure creation based on URL paths
- Customizable crawling settings
Inside your project, pass the URL of the documentation you want to crawl:
from uptodateai import process_docsprocess_docs("https://microsoft.github.io/graphrag/")-
Run the code. A new folder called
/docs(very well organized in Markdown format)will be generated within your project. This directory we will then use it as our knowledge base for AI models to enhance their understanding of the latest frameworks. -
We pass the knowledge base to AI using Cursor Composer:
- Enable it in Cursor Settings under the "Beta" section
- Use the shortcut Cmd+I (MacOS) or Ctrl+I (Windows) to open the Composer interface
- Provide instructions (eg build a web app using Fasthtml and mention relevant files for our case we will mention
@/docs)
-
The cursor composer will generate new files and code in accordance to the library documentation
-
Review and refine the generated code as needed
-
Please give it a star on GitHub if you find it helpful:
This project is licensed under the MIT License. See the LICENSE file for details.
Contributions are welcome! Please open an issue or submit a pull request.