DEV Community

Yan Wang profile picture

Yan Wang

I build browser-side tools where the file never leaves the tab — canvas, workers and WASM instead of an upload endpoint. Currently lensup.ai, a document scanner running in 20 languages.

Joined Joined on  Personal website https://lensup.ai/

Work

Independent — building browser-side web tools

When a PDF Export Turns Every Page Into a Picture, Define the Text Contract First

When a PDF Export Turns Every Page Into a Picture, Define the Text Contract First

Comments
3 min read

Want to connect with Yan Wang?

Create an account to connect with Yan Wang. You can also sign in below to proceed if you already have an account.

Already have an account? Sign in
A PDF Cover Needs More Than a Screenshot as Its Test Oracle

A PDF Cover Needs More Than a Screenshot as Its Test Oracle

Comments
3 min read
Report missing and repeated pages in a PDF split plan

Report missing and repeated pages in a PDF split plan

Comments
3 min read
Parse a page selection completely or not at all

Parse a page selection completely or not at all

Comments
3 min read
Test a JPG Name Collision Policy With Executable Examples

Test a JPG Name Collision Policy With Executable Examples

Comments
3 min read
Count OOXML Payload Bytes Without Pretending to Predict Compression

Count OOXML Payload Bytes Without Pretending to Predict Compression

Comments
3 min read
Plan EPUB Reading Order From the Spine, Not ZIP Filenames

Plan EPUB Reading Order From the Spine, Not ZIP Filenames

Comments
4 min read
A contain rectangle is a contract, not a PDF fidelity test

A contain rectangle is a contract, not a PDF fidelity test

Comments
3 min read
A Centimeter-to-Pixel Contract With Explicit Limits

A Centimeter-to-Pixel Contract With Explicit Limits

Comments
3 min read
Validate a PDF reorder as a permutation before you test the export

Validate a PDF reorder as a permutation before you test the export

Comments 1
3 min read
Test PDF merge navigation separately from page order

Test PDF merge navigation separately from page order

Comments
6 min read
Test OCR text and Markdown structure separately with a small token check

Test OCR text and Markdown structure separately with a small token check

Comments
4 min read
Page positions, page labels, and printed numbers are different PDF data

Page positions, page labels, and printed numbers are different PDF data

Comments
4 min read
A PDF merge can preserve mixed page sizes: verify the download, not the thumbnails

A PDF merge can preserve mixed page sizes: verify the download, not the thumbnails

Comments 1
4 min read
300 DPI means four times the pixels: budget your canvas before rendering a PDF

300 DPI means four times the pixels: budget your canvas before rendering a PDF

Comments
4 min read
60 % of our image decode failures were HEIC files named .jpg

60 % of our image decode failures were HEIC files named .jpg

Comments 3
6 min read
A scan is never perfectly square

A scan is never perfectly square

1
Comments 1
6 min read
The adaptive threshold that deleted the thing we were looking for

The adaptive threshold that deleted the thing we were looking for

Comments 1
5 min read
Three things that bit us moving a document scanner fully into the browser

Three things that bit us moving a document scanner fully into the browser

Comments 3
6 min read
loading...