Extract and download key-value pairs, tables, and paragraphs from your scanned pdf, jpg, and png documents as CSV files.

1. Run npm i in frontend folder followed by npm run build
2. Run pip install -r requirements.txt in root folder
3. Create a .env file with the below content:

Create a Azure FormRecognizer service and copy the Endpoint and KEY1 from Keys and Endpoint. These will be the ENDPOINT and KEY respectively. Next create an azure storage account, and create a container in it. Go to Shared access tokens and click Generate SAS token and URL. Copy the Blod SAS URL. The part to the left of ? goes in BLOB_ENDPOINT and the part to the right goes in BLOB_QUERY

ENDPOINT = "https://xyz.cognitiveservices.azure.com"
KEY = "12345something"
BLOB_ENDPOINT = "https://xyz.blob.core.windows.net/containerName/"
BLOB_QUERY = "?xyz=xyz&xyz=xyz..."

4. Run with py main.py

Run py extract.py -i "input/file/path.pdf" -o "output/file/path.csv"