Example using textTract

detect-doc-text-async.sh

#!/bin/bash
# @meta version 1.0.0
# @option --pdf   path to pdf file.
# The file is processed and saved as 'raw-text.txt'
# in the current directory.
`eval "$(argc --argc-eval "$0" "$@")"`

if [ -z "${argc_pdf}" ]; then
  echo "Try --help"
  exit 1
fi

PDFDOC=$argc_pdf

aws s3 cp $PDFDOC s3://upload-docs.limpidfox.com/upload.pdf

# Variables
BUCKET_NAME="upload-docs.limpidfox.com"
DOCUMENT_NAME="upload.pdf"
REGION="us-east-1"  # Change to your desired region

# Start the document text detection
JOB_ID=$(aws textract start-document-text-detection \
    --document-location '{"S3Object":{"Bucket":"'"$BUCKET_NAME"'","Name":"'"$DOCUMENT_NAME"'"}}' \
    --region "$REGION" \
    --query "JobId" \
    --output text)

echo "Job started with JobId: $JOB_ID"

# Polling loop to check job status
STATUS="IN_PROGRESS"
while [ "$STATUS" == "IN_PROGRESS" ]; do
    echo "Waiting for job to complete..."
    sleep 5  # Wait for 5 seconds before checking again

    STATUS=$(aws textract get-document-text-detection \
        --job-id "$JOB_ID" \
        --region "$REGION" \
        --query "JobStatus" \
        --output text)
    echo "Current status: $STATUS"
done

# Check if the job completed successfully
if [ "$STATUS" == "SUCCEEDED" ]; then
    echo "Job succeeded. Fetching results..."

    # Fetch the full results
    aws textract get-document-text-detection \
        --job-id "$JOB_ID" \
        --region "$REGION" > textract-results.json

    echo "Results saved to textract-results.json"

    # Optionally extract raw text using jq
    jq -r '.Blocks[] | select(.BlockType == "LINE") | .Text' textract-results.json > raw-text.txt
    echo "Raw text saved to raw-text.txt"
else
    echo "Job failed with status: $STATUS"
fi