- Example fund information in the urls.csv file. Note - first entry is the link to factsheet definition file.
- Use 'download_pdfs.py' and then 'convert_pdfs.py' to get the fund data
- Use 'mooc-analyzer.ipynb' to then step through the main application.
- If files in right directory it should generate questions and create a csv file with the extracted info that you can query.
- If you want to just see the output of processing over 100 fund pdf files using gpt-4o-mini using this method then look at the extracted_fund_data.csv file.
e.g.
ask_advisor("What is the most common holding across all of the funds? Just return the name of the one holding.")
The most common holding across all these funds is Microsoft Corporation, appearing in multiple funds with varying levels of allocation.
The application uses OpenAI gpt-4o-mini and OpenAI Swarm agent workflow
Prepared for:
