Azure – How do I read the original pdf file in set indexer datasource in a custom WebApiSkill after enabling "Allow Skillset to read file data"

I see the following under my indexer settings:

When hovering over it I read the following:

True means the original file data obtained from your blob data source
is preserved. This allows passing the original file to a custom skill,
or to the Document Extraction skill.

How do I read the original pdf file in the associated blob data source in a custom WebApiSkill?

file_data_base64 = value.get('data', {}).get('file_data', '')
...

EDIT

I enabled Allow Skillset to read file data in the indexer. My full setup:

• WebApiSkill inputs

inputs=[
    InputFieldMappingEntry(name="file_data", source="/document/file_data")
],

• WebApiSkill input reading

import azure.functions as func
import datetime
import json
import logging
import base64
import fitz
from io import BytesIO

app = func.FunctionApp()
logging.basicConfig(level=logging.INFO)


@app.route(route="CustomSplitSkill", auth_level=func.AuthLevel.FUNCTION)
def CustomSplitSkill(req: func.HttpRequest) -> func.HttpResponse:
    logging.info('Python HTTP trigger function processed a request.')

    try:
        req_body = req.get_json()
        logging.info('Request body parsed successfully.')
    except ValueError:
        logging.error(f"Invalid input: {e}")
        return func.HttpResponse("Invalid input", status_code=400)

    # 'values' expected top-level key in the request body
    response_body = {"values": []}
    for value in req_body.get('values', []):
        recordId = value.get('recordId')
        file_data_base64 = value.get('data', {}).get('file_data', '').get('data', '')
        if not file_data_base64:
            logging.error("No file_data found in the request.")
            return func.HttpResponse("Invalid input: No file_data found", status_code=400)

        try:
            file_data = base64.b64decode(file_data_base64)

            try:
                pdf_document = fitz.open(stream=BytesIO(file_data), filetype='pdf')
            except fitz.FileDataError as e:
                logging.error(f"Failed to open PDF document: {e}")
                return func.HttpResponse("Failed to open PDF document", status_code=400)
            except Exception as e:
                logging.error(f"An unexpected error occurred while opening the PDF document: {e}")
                return func.HttpResponse("An unexpected error occurred", status_code=500)
            
            if pdf_document.page_count == 0:
                logging.error("No pages found in the PDF document.")
                return func.HttpResponse("Invalid PDF: No pages found", status_code=400)

            extracted_text = ""
            for page_num in range(pdf_document.page_count):
                page = pdf_document.load_page(page_num)
                extracted_text += page.get_text()

            combined_list = [{'textItems': ['text1', 'text2'], 'numberItems': [0, 1]}]  # i deleted the chunking and associated page extraction for simplicity

            response_record = {
                "recordId": recordId,
                "data": {
                    "subdata": combined_list
                }
            }
            response_body['values'].append(response_record)
        except Exception as e:
            logging.error(f"Error processing file_data: {e}")
            return func.HttpResponse("Error processing file_data", status_code=500)

    logging.info('Function executed successfully.')
    return func.HttpResponse(json.dumps(response_body), mimetype="application/json")

The error:

Message:
Could not execute skill because the Web Api request failed.

Details:
Web Api response status: 'NotFound', Web Api response details: ''

Given that I have projections I cannot debug this properly as debugging is not supported with projections. The logging does not seem to log the specific error either despite the error handling and checks.