Suppressing "Corrupt JPEG data: 1 extraneous bytes before marker 0xd9" output about tesseract.js

[ImSpyke]

I am working on this project with matthieu, dont hesitate to ping me too if you have any solutions ^^

[matthieuEv]

Any news? :)

[Balearica]

Please provide an example image that can be used to replicate this issue.

[Balearica]

I would like to know if there is a way to suppress this specific output from being logged to the console. I have attempted to use custom logger functions, but the message still appears. Temporarily overriding console.log, console.warn, and console.error works, but it feels like a hack rather than a solution.

Since these warning messages are created by Tesseract rather than being something added by Tesseract.js, disabling specific warning messages at the source is outside of the scope of this repo. Indeed, the only way filter messages Tesseract prints to stderr without modifying the Tesseract codebase would be to implement the hacky solutions you describe (e.g. overriding console.log).

However, Tesseract does support saving warning messages as an output, rather than printing them to stderr (which maps to console.warn in JavaScript). Therefore, by redirecting the messages to the output, you should be able to implement something similar to what you are describing. To capture debugging messages in the output object, rather than printing them to console, set the debug to true within the output options.

const worker = await await Tesseract.createWorker("eng");
const ret = await worker.recognize(files[i], undefined, {debug: true});
console.log(ret.data.text);
console.log(`Captured message: ${ret.data.debug}`);

Captured message: Estimating resolution as 390

Once you have the debug output string, you can either ignore it or save it to a string--essentially suppressing all warning messages from the console--or write a wrapper function that prints warnings only when they meet certain criteria.

[matthieuEv]

Hello, sorry for my inactivity here😅, this method seems to work for me, thanks again 😁