jamie-lemon
commented
2 months ago Agree there is no case sensitivity for the search at present. However I do believe MuPDF.js detects line-wrapped words, If I use your file with:
let results = page.search("Hello world")
It delivers 3 results as follows:
[
[
[
72,
75.22499084472656,
149.67677307128906,
75.22499084472656,
72,
91.24498748779297,
149.67677307128906,
91.24498748779297
]
],
[
[
287.1199951171875,
75.22499084472656,
360.5367431640625,
75.22499084472656,
287.1199951171875,
91.24498748779297,
360.5367431640625,
91.24498748779297
]
],
[
[
505.17999267578125,
75.22499084472656,
540.9913940429688,
75.22499084472656,
505.17999267578125,
91.24498748779297,
540.9913940429688,
91.24498748779297,
72,
96.28498840332031,
109.5767593383789,
96.28498840332031,
72,
112.30498504638672,
109.5767593383789,
112.30498504638672
]
]
]These are QuadPoints which represent the areas with the text found ( see: https://mupdfjs.readthedocs.io/en/latest/how-to-guide/node/document/index.html#searching-a-document ).
I also tested here: https://casper.mupdf.com/wasm/demo/ and uploaded your test.pdf file and performed a search for "hello world" the UI then highlighted these areas on the document:
So I think the search method does find of the three instances that you expect with the correct bounding box data with the QuadPoints!
xybei
The search function can not match line-wrapped words (as pdf.js does in the figure below). It also does not support "Match Case" or "Whole Words". Hope it can be improved, thanks!
test.pdf