- Resolve state inconsistency issue. ✅
- We have made some minor improvements in the model. 😎
- We've enhanced the model to process em/en, third/quarter, thin/hair, medium math space characters & regular/narrow non breaking space characters. 🚀 🛰️
- We have made some behind the scene updates in type definitions. 😎
- Accented characters now also include diacritical marks. 🙌
- Token containing a word joiner is treated as a single token. ✅
- Shape of token containing accented chars, diacritical marks & word joiner char(s) is now determined after removing them. 👏
- We've upgraded the model to process non-breaking spaces in the same way as standard spaces. 🙌
- Updated the missed package.json version. ❤️🩹
- We have added link to the github repo in the package.json file. 🙌
- We have enhanced the model a little bit — POS tagging accuracy is now touching 95%. 🙌 🎉
- We have made some behind the scene updates. 😎
- Added missing pre-processing of fractions for NER. ✅
- We have added engines in package.json for Node.js >= 16 and link to compatible browser version in README. 😎 📔
- While loading, it automatically detects unsupported Node.js & Browsers and their versions. 😎 🙌 😇
- We have removed Node.js
Buffer.from()API calls completely. ✅ 👏 🎉
- We have made some behind the scene updates. 😎
- Squashed a typo bug in package.json in
typesfield.
- We have added support for Typescript. 🙌🎉
- Now create multiple instances of winkNLP using the model. 🔢
- Rectified a buggy regex, which failed to parse relatively long numerals. 👏✅
- Determine the document's Flesch Reading Ease Score (FRES), Reading Time, Complex words, Sentiment Score and more with
its.readabilityStatshelper. 📚📊👏
- Now use
its.lemmahelper to obtain lemma of words. 👏 🎉
- We have made some behind the scene updates to lexicon & pos model. 📔
- Custom entities meta model loading was failing. It has been fixed now. ✅
- Happy to release version 1.0.0 for you! 💫👏
- You can now build pure web browser or mobile NLP apps without requiring any server side deployment. 😇