What is this?
A Java suite of core NLP tools that takes raw human language text and returns linguistic analysis: tokenization, parts of speech, named entities, parsing, coreference, and sentiment.
A Java suite of core NLP tools that takes raw human language text and returns linguistic analysis: tokenization, parts of speech, named entities, parsing, coreference, and sentiment.Is it a library I call from my own Java code, or a separate service I run?A library: you add it as a dependency and call it from your code. The README calls it an integrated framework - starting from plain text, all the tools run in about two lines of code.
A library: you add it as a dependency and call it from your code. The README calls it an integrated framework - starting from plain text, all the tools run in about two lines of code.How do I add it to my project?Take a release from Maven Central as your Gradle or Maven dependency, or build the latest code yourself with 'mvn package' or Ant. The model jars are a separate download.
Take a release from Maven Central as your Gradle or Maven dependency, or build the latest code yourself with 'mvn package' or Ant. The model jars are a separate download.Do I need anything besides Java installed?No - the code is pure Java, with no Python, GPU, or other runtime required. You do need the model jars on your CLASSPATH, and Ant or Maven only if you build from source.
No - the code is pure Java, with no Python, GPU, or other runtime required. You do need the model jars on your CLASSPATH, and Ant or Maven only if you build from source.What trips people up the first time they run it?The models are not fully bundled: the default jar covers basic English, while larger models like the shift-reduce parser and WikiDict require the English extra and KBP jars, and every other language needs its own models jar.
The models are not fully bundled: the default jar covers basic English, while larger models like the shift-reduce parser and WikiDict require the English extra and KBP jars, and every other language needs its own models jar.Is this a research toy, or do production systems actually use it?Production-grade: the tools are described as stable and well-tested, and they are widely used by groups in academia, industry, and government.
Production-grade: the tools are described as stable and well-tested, and they are widely used by groups in academia, industry, and government.Can I ship it inside a proprietary product I distribute?No - it is the full GPL (v2 or later). Most free uses are fine, but you cannot distribute it as part of proprietary software.
No - it is the full GPL (v2 or later). Most free uses are fine, but you cannot distribute it as part of proprietary software.What can it actually tell me about a piece of text?Base forms of words, parts of speech, whether a token is a company or person name, normalized dates, times and quantities, sentence structure as syntactic phrases or dependencies, and which noun phrases refer to the same entity.
Base forms of words, parts of speech, whether a token is a company or person name, normalized dates, times and quantities, sentence structure as syntactic phrases or dependencies, and which noun phrases refer to the same entity.Does it work on languages other than English?Yes. Originally English-only, it now provides varying levels of support for Modern Standard Arabic, mainland Chinese, French, German, Hungarian, Italian, and Spanish, each with its own models jar hosted on the Hugging Face Hub.
Yes. Originally English-only, it now provides varying levels of support for Modern Standard Arabic, mainland Chinese, French, German, Hungarian, Italian, and Spanish, each with its own models jar hosted on the Hugging Face Hub.How often does it get updated?A new version is distributed several times a year, cut from a stable commit. Between releases you can always build the latest development code from GitHub HEAD yourself.
A new version is distributed several times a year, cut from a stable commit. Between releases you can always build the latest development code from GitHub HEAD yourself.Under the hood, is it hand-written rules or neural networks?A mix: the different tools use rule-based, probabilistic machine learning, and deep learning components.
A mix: the different tools use rule-based, probabilistic machine learning, and deep learning components.What are similar projects?HanLP, spaCy, bert, awesome-nlp, flair, NLP-progress, funNLP, allennlp
HanLP, spaCy, bert, awesome-nlp, flair, NLP-progress, funNLP, allennlpWhich topics is it in?nlp-library, named-entity-recognition, sentiment-analysis, natural-language-processing, academic-research, multilingual, nlp, artificial-intelligence, java, gplv3, deep-learning, framework
nlp-library, named-entity-recognition, sentiment-analysis, natural-language-processing, academic-research, multilingual, nlp, artificial-intelligence, java, gplv3, deep-learning, framework