kelleyl / econ_form_corrector Goto Github PK

need to be able to render form with list of suggestions (top 5) for selected documents

Add a readme and screenshots describing how to use the platform

display a status on the data display page showing the status of each document

allow multiple files to be uploaded at once, display all uploaded files on a checklist page with a button to begin the cleaning

Table Extraction
Input -> PDF
From the original pdf, extract the page with table (should be able to filter out form pages)
OCR (tabula) form page

Table Normalization
Input -> Dirty tables in CSV, Optional List of canonical forms, Optional list of previous corrections
This stage could be simplified by doing multiple tables at once
Show some correction candidates and autofill previous corrections
Save corrections

Next Step File Generation

allow upload of pdf to identify and save pages containing tables of Schedule H