scribelabsai / amazon-trp-node

Amazon Textract Response Parser library for Node.

amazon-trp-node

Amazon Textract Response Parser library for TS.

---

This is a TS port of the Python TRP script from AWS.

Installation

npm install @scribelabsai/amazon-trp

Usage

1. Load the blocks into a document (e.g. using GetDocumentAnalysisCommand from @aws-sdk/client-textract)

import { TextractClient, GetDocumentAnalysisCommand } from '@aws-sdk/client-textract';
import { Document } from '@scribelabsai/amazon-trp';

import type { BlockStruct } from '@scribelabsai/amazon-trp';

const client = new TextractClient();
const resp = await client.send(new GetDocumentAnalysisCommand({ JobId: 'MY_JOBID }));
const doc = new Document(resp.blocks as BlockStruct[]);

2. Do something with the document (e.g. getting tables)

doc.pages.forEach((p) => {
  p.tables.forEach((t) => {
    // Do something with the table
  });
});

License

MIT, see LICENSE file.

amazon-trp-node's Issues

Difficulty Accessing Table Content in Textract Output

I'm aware that this question may seem rather basic, but I'm facing an issue with accessing the complete content of tables in my code. In the example below, the console.log("p", p.text) successfully displays all the content of the image. However, console.log("c.text", c.text) appears to access only some elements. Could someone provide an example of how to iterate over the table properly?

const params = {
  Document: {
    Bytes: blob,
  },
  FeatureTypes: ["TABLES"],
};

const command = new AnalyzeDocumentCommand(params);
try {
  const data = await textractClient.send(command);

  const doc = new Document(data.Blocks as BlockStruct[]);
  doc.pages.forEach((p) => {
    console.log("p", p.text);
    p.tables.forEach((t) => {
      t.rows.forEach((l) => {
        console.log("l", l.cells);
        l.cells.forEach((c) => {
          console.log("c.text", c.text);
        });
      });
    });
  });
}
I'm looking for assistance in understanding why c.text is not accessing the entire content and how to properly iterate over the table to access all its elements. Any insights or examples would be greatly appreciated.

Tahnks its not easy to be a noob

Publish to NPM

It looks like the last version that was published to NPM was 2 years ago v1.0.0.