[Feature] Advanced Highlighting

[valh1996]

[Feature] Advanced Highlighting

[valh1996]

Maybe we can take inspiration from this:

• https://github.com/AaronMorais/pdf-highlighter/blob/master/src/pages/index.js#L184-L211
• https://github.com/EvolutionJobs/pdf-viewer/blob/master/pdf-viewer-page.ts

[hrynko]

Hi @valh1996,

Do you think updating PDFJS to 2.13.216 would cover the diacritics insensitive search problem (check this PR)? As for "several words to search", I'm not sure how this is supposed to work.

[valh1996]

Hi @valh1996,

Do you think updating PDFJS to 2.13.216 would cover the diacritics insensitive search problem (check this PR)? As for "several words to search", I'm not sure how this is supposed to work.

Hi @hrynko,

Yes thanks, I think it would be perfect for the diacritics. However, how to access the pdfFindController to execute a find with your package?

I think it should be possible to do it via the ref making pdfFindController public:

pdfEmbedRef.pdfFindController.executeCommand('find', {
  caseSensitive: false,
  findPrevious: undefined,
  highlightAll: true,
  phraseSearch: false,
  query: query
});

As for the "several words to search", the problem is that if I run the above find with the word "Alice", and then re-run it with the word "Alex", then it will overwrite all previous matches with the word "Alice".

But for that, I'm not sure if you can do something at your level. I could for example simply modify this part of the PDF-JS lib with patch-package for exemple.

Therefore, could you please do an update to make access to the find command? And when a new version with the changes on master will be available (for PDFJS > 2.13.216) ?

EDIT : It looks like we have to go through the eventBus now to highlight the text instead of executeCommand. I don't know if you have an example to highlight in this case?

[hrynko]

I've tried exposing something like the following, with no success so far:

import { EventBus, PDFFindController } from 'pdfjs-dist/legacy/web/pdf_viewer.js'
...
const findController = new PDFFindController({
  eventBus: new EventBus(),
  linkService: this.linkService,
})
findController.setDocument(this.document)

I'm not sure if this will work outside of PDFViewer yet, but if you could continue this experiment, I would appreciate a PR.

[valh1996]

I've tried exposing something like the following, with no success so far:

import { EventBus, PDFFindController } from 'pdfjs-dist/legacy/web/pdf_viewer.js'
...
const findController = new PDFFindController({
  eventBus: new EventBus(),
  linkService: this.linkService,
})
findController.setDocument(this.document)

I'm not sure if this will work outside of PDFViewer yet, but if you could continue this experiment, I would appreciate a PR.

Yes that's what I tried too, but as you say it can't work without a viewer I think. I tried to ask the question, but it seems that in this case we have to initialize everything manually...

Wouldn't it be easier to refactor using the simple viewer?

What are the advantages of having created the component outside the viewer?

[hrynko]

What are the advantages of having created the component outside the viewer?

I expected that not using the viewer component could be more flexible and predictable, although it would have some limitations.

Wouldn't it be easier to refactor using the simple viewer?

It might be, but it would require additional refactoring of the component and could lead to unexpected side effects. So if it can be done without using a viewer, I would do it like this. Otherwise, I would postpone it until the next minor release.

[valh1996]

I tried an alternative solution to make the highlight when rendering the textlayer since we get the text with the exact position. But we lose all the advantages of the PDFFindController.

So, after trying several things, I can't get a conclusive result if you can help me on this please?

[hrynko]

I'm having issues updating PDFJS, so I'd like to resolve them first. Will have another look at the highlighting issue afterwards.