aws-textract.js

const config = require('config');
const AWS = require('aws-sdk');
const fs = require('fs');
const path = require('path');

AWS.config.region = config.get('textract').region;
AWS.config.credentials = new AWS.Credentials(config.get('s3'));

const awsTextract = new AWS.Textract();

/**
 * Generate CSV Data from Image using AWS Textract
 * 
 * @params inputFile - Path of the input file
 *  
 */
async function generateAwsTextract({ inputFile }) {
  let promise = new Promise((resolve, reject) => {
    
    let inputBuffer = fs.readFileSync(inputFile);

    const params = {
      Document: {
        Bytes: Buffer.from(inputBuffer),
      },
      FeatureTypes: [
        'TABLES'
      ]
    };

    awsTextract.analyzeDocument(params, (err, data) => {
      if (err) reject(err); // an error occurred
      let blocks = data.Blocks;
      let blocks_map = {};
      let table_blocks = [];
      let csv = '';

      for(let block of blocks) {
        blocks_map[block['Id']] = block;
        
        if (block.BlockType == 'TABLE') {
          table_blocks.push(block);
        }
      }

      if(table_blocks.length == 0) {
        let parsedData = {};
        return resolve(parsedData);
        // return reject('Not Found');
      } else {
        table_blocks.forEach((table, index) => {
          csv += generateAwsTextractTableCsv(table, blocks_map, index);
          csv += '\n\n';
        });

        resolve(csv);

        });

      }
    });


  });
  return promise;
}


function getAwsTextractText(result, blocks_map) {
  let text = '';
  if (result.Relationships) {
    result.Relationships.forEach(relationship => {
      if (relationship.Type === 'CHILD') {
        relationship.Ids.forEach(child_id => {
          let word = blocks_map[child_id];
          if (word.BlockType === 'WORD') {
            text += word.Text + ' ';
          }

          if (word.BlockType === 'SELECTION_ELEMENT' && word.SelectionStatus == 'SELECTED') {
            text += 'X ';
          }
        });
      }
    });
  }
  return text;
}

function generateAwsTextractRowsColumnMap(table_result, blocks_map) {
  let rows = {};
  for (let relationship of table_result.Relationships) {
    if (relationship && relationship.Type === 'CHILD') {

      relationship.Ids.forEach(id => {
        let cell = blocks_map[id];
        if (cell.BlockType === 'CELL') {
          let row_index = cell.RowIndex;
          let col_index = cell.ColumnIndex;
          if (typeof rows[row_index] == 'undefined') {
            rows[row_index] = {};
          }
          rows[row_index][col_index] = getAwsTextractText(cell, blocks_map);
        }
      });
    }
    return rows;
  }
}


function generateAwsTextractTableCsv(table_result, blocks_map, table_index) {
  let rows = generateAwsTextractRowsColumnMap(table_result, blocks_map);
  let table_id = 'Table_' + table_index;
  let csv = `BOM Table: ${table_id+1}\n`;
  for (let rowKey of Object.keys(rows)) {
    for (let colKey of Object.keys(rows[rowKey])) {
      csv += rows[rowKey][colKey] + ',';
    }
    csv += '\n';
  }
  csv += '\n\n\n';
  return csv;
}

module.exports = {
  generateAwsTextract
};