This Python script is designed to read a text file and a prompt file, process the text file by splitting it by page numbers, and then send each page along with the prompt to a ChatGPT API. The script handles these operations in parallel, using up to 8 workers for efficiency.
- File Reading: Reads a specified text file and splits it by page numbers.
- Prompt Processing: Reads a prompt file and replaces a placeholder with the actual page text.
- ChatGPT Integration: Sends processed data to a ChatGPT API and retrieves responses.
- Parallel Processing: Handles multiple pages concurrently, improving processing speed.
- JSON Validation: Validates the JSON response from the API.
- Response Handling: Prints the JSON response (can be modified to store the response).
python3 process.py --txtfile DAVEBUSTER\'SENTERTAINMENTINC_20220629_8-K_EX-101_CreditLoanAgreement.PDF.txt --prompt page_data.prompt.txt
The output of the ChatGPT API should be something like:
{
"page_number": 1,
"tranche": null,
"quantum": null,
"financial_maintenance_covenant": null,
"addbacks_cap": null,
"MFN_threshold": null,
"MFN_exceptions": null,
"portability": null,
"lender_counsel": null,
"borrower_counsel": null,
"borrower": "DAVE & BUSTER’S, INC., as the Borrower",
"guarantor": "THE OTHER GUARANTORS PARTY HERETO FROM TIME TO TIME",
"admin_agent": "DEUTSCHE BANK AG NEW YORK BRANCH, as Administrative Agent and Collateral Agent",
"collat_agent": "DEUTSCHE BANK AG NEW YORK BRANCH, as Collateral Agent",
"effective_date": "June 29, 2022"
}