A Python wrapper and JSON-RPC server for Stanford CoreNLP, providing NLP tools like parsing, tagging, and coreference resolution.
Stanford CoreNLP Python is a Python wrapper for Stanford University's CoreNLP natural language processing tools. It provides programmatic access to advanced NLP capabilities including part-of-speech tagging, parsing, named entity recognition, and coreference resolution through either direct module import or a JSON-RPC server interface. The wrapper handles communication with the Java-based CoreNLP system and converts outputs to Python-friendly formats.
Python developers and researchers who need access to Stanford's comprehensive NLP tools without working directly in Java. It's particularly useful for applications requiring production-ready linguistic analysis with minimal setup overhead.
Developers choose this wrapper because it provides reliable access to Stanford's well-established NLP models through a clean Python interface, eliminating the need for complex Java-Python integration. The JSON-RPC server option makes it suitable for scalable applications where NLP processing needs to be decoupled from main application logic.
Python wrapper for Stanford CoreNLP tools v3.4.1
Open-Awesome is built by the community, for the community. Submit a project, suggest an awesome list, or help improve the catalog on GitHub.
Provides Python interface to Stanford CoreNLP's full suite, including part-of-speech tagging, parsing, named-entity recognition, and coreference resolution, as detailed in the README's feature list.
Can run as a JSON-RPC server outputting structured JSON data, enabling easy integration into web applications or distributed systems, with examples in client.py.
Outputs parse trees compatible with NLTK for further natural language processing, facilitating seamless use within Python's NLP ecosystem.
Supports pronoun dereferencing with detailed reference tracking across sentences, a powerful feature for advanced text analysis, as explained in the coref section.
Tied to Stanford CoreNLP version 3.4.1 from 2014, which lacks newer models and improvements, and the wrapper may break with updates, as noted in the README.
Requires 3GB of RAM and takes minutes to load models, making it unsuitable for quick startups or low-memory environments, as warned in the Questions section.
Depends on a separate Java process and specific jar files, adding setup complexity and potential compatibility issues, with manual download and configuration steps.