search

naptha / tesseract.js

Pure Javascript OCR for more than 100 Languages 📖🎉🖥
http://tesseract.projectnaptha.com/
Apache License 2.0
34.09k stars 2.15k forks source link

Add new example for custom granular output [v5] #896

Closed Kishlay-notabot closed 3 months ago

Kishlay-notabot commented 4 months ago

Add new example for custom granular output, Users can generate json files which contain bbox data of words/symbols detected in the input files. These examples add the ability for users to generate bulk datasets for handwritten characters/words. See #877

Balearica commented 4 months ago

Thanks for putting this together. I reviewed, and have a couple notes aimed at making sure other users can understand and run this example.

  1. As Tesseract.js supports both browser and Node.js, the README should specify in the first few sentences that this is an example specifically for Node.js.
    1. This may seem obvious, however users trying to run Node.js code in the browser, or browser code in Node.js, is actually one of the most common sources of confusion.
  2. The example repo, and open source Node.js projects in general, should include a package.json file that installs the correct version of the relevant dependencies
    1. Running npm i should install all the correct dependencies
    2. In addition to saving other users time, specifying the expected version of packages ensures that the example still runs even when dependencies implement breaking changes in later versions.
Kishlay-notabot commented 4 months ago

@Balearica my apologies for not including the package-json file. I accidentally deleted it in the latest commit of the repo, I'll re-add it once I get back to my laptop.

Kishlay-notabot commented 4 months ago

@Balearica updated with all the necessary changes, clarified the usage method and added package-json files.

Kishlay-notabot commented 4 months ago

The contributing link was broken, fixed it with the latest commit.

Kishlay-notabot commented 3 months ago

@Balearica please review this

Balearica commented 3 months ago

I reviewed this today, and have a few more comments. As before, the overall goal is making sure new users can easily run this code.

  1. Running npm i currently does not install all the dependencies required, because package.json does not include the dependency canvas
    1. It's possible that this happened because you have canvas installed globally on your system
  2. Please add the exact commands users need to run as code blocks in the README
    1. This appears to be:
      1. node OCR-and-bbox-export.js
      2. node Crop-from-exported-json.js
  3. Please add a sample directory/image such that the commands run without the user needing to create a new folder and copy in data
    1. At present downloading and running immediately results in an error, as it appears to be looking for images in input_images despite this path not existing in the repo

If you make these adjustments such that new users can simply (1) clone, (2) run npm i, and (3) run node OCR-and-bbox-export.js/node Crop-from-exported-json.js, and everything runs, I will merge in.

Kishlay-notabot commented 3 months ago

@Balearica Added all the specific details, Please check it out!

Balearica commented 3 months ago

Great, I will merge. Thank you for this contribution to the community.