Summary

This PR deletes the extracted_invoice_json directory and removes the trailing space from the 'extracted_invoice_json ' directory. This fixes an invalid path error when git cloning on Windows machines:
invalid_path_git_clone

Motivation

These changes are necessary to allow Window users to successfully clone the repo on their machine. This fixes 2 open issues (#1934, #1837), improving the quality of the repository's file structure and reducing confusion. Both directories contain the same filenames but the files themselves differ slightly due to the non-deterministic nature of LLMs. A typo must've been made when creating the cookbook and then re-run without the space, creating the two directories. Due to the differing files, we do an analysis to find the correct directory.

Correctness

The choice of which directory to delete is based on it's associated cookbook and notebook. Using the json showed at the end of part 1, we can determine the referenced filename using a simple search (leading us to premierinn_GABCI19014325_extracted.json). From there, we can determine which of the 2 directories is the one used for the cookbook. We check the filename in both directories and which ever one matches is the correct one, leading to the deletion of the non-matching directory.

We also see that the correct directory name is specified in the notebook with extracted_invoice_json_path = "./data/hotel_invoices/extracted_invoice_json". So we remove the trailing space from the remaining directory. This analysis can be found in this notebook.