tesseract.js

What's Changed

• Significant improvements to recognition speed (by @fanchenkong1 in https://github.com/naptha/tesseract.js/pull/1039)
  • A new relaxedsimd build takes advantage of the latest WASM and hardware capabilities.
  • Upgrading from v6 to v7 reduces runtimes by ~15-35% (depending on device and use-case). The highest uplifts are currently seen on the latest Intel processors. *...

What's Changed

• Switched from using node-fetch to using built-in fetch implementation for Node.js v18 and newer.
  • This resolves a nuisance dependency-related warning message reported in #876
• Various minor updates

Full Changelog: https://github.com/naptha/tesseract.js/compare/v6.0.0...v6.0.1

What's Changed

• Fixed memory leaks (#977)
  • This version fixed a long-standing issue where memory would rise over time, eventually leading to a crash.
• Reduced runtime and memory usage for most users by updating default formats (#916).
• Fixed compatibility with Electron main process (#925)
• Fixed bug where user-provided parameters were overwritten by defaults (#975).

Breaking...

What's Changed

• Fixed bug where incorrect language data was loaded when oem is 2 (#938)
• Fixed bug where boolean values in blocks output format would always be false (#907)

Full Changelog: https://github.com/naptha/tesseract.js/compare/v5.1.0...v5.1.1

What's Changed

• Added line size metrics to blocks output (#906)
  • line objects now include a property named rowAttributes, which is an object containing ascenders, descenders, and row_height metrics
  • These metrics allow for manual font size calculations that are more accurate than using the font_size property.
• Updates to documentation, types, and dependencies

N...

What's Changed

• Fixed bug triggered by running worker.recognize while a previous call to worker.recognize is still running (#875)
  • Sending multiple jobs to the same worker at the same time is still not recommended.
  • Instead, schedulers should be used to coordinate running jobs in parallel (see [...

What's Changed

• Fixed support for setting "init only" parameters using config option of createWorker (#862)
  • For example, load_number_dawg is an "init only" parameter that cannot be set using either worker.setParameters or the options argument of worker.recognize.
  • However, load_number_dawg can be set by the following createWorker statement.
    • `createWorker(...

What's Changed

• Minor changes to types, documentation, and build

New Contributors

• @dora-micha made their first contribution in https://github.com/naptha/tesseract.js/pull/843

Full Changelog: https://github.com/naptha/tesseract.js/compare/v5.0.2...v5.0.3

What's Changed

• Fixed bugs with wrong lang data being loaded per #834 and #835 by @Balearica in https://github.com/naptha/tesseract.js/pull/836

Version 5.0.1 is nearly identical to 5.0.2 and was the latest version for under a day, so does not have its own release notes.

Full Changelog: https://github.com/naptha/tesseract.js/compare/v5.0.0...v5.0.2

What's Changed

Major New Features

1. Significantly smaller file sizes
   1. 54% smaller file sizes for English, 73% smaller for Chinese (see #806 for details)
   2. This results in a ~50% decrease in runtime for first-time users (who do not yet have the data downloaded/cached)
2. Significantly lower memory usage
   1. Worker memory utilization in the [web benchmark](https://g...