TextServer: Cloud-Based Multilingual Natural Language Processing

Lluís Padró, Jordi Turmo · 2015

TextServer is an efficient language analysis platform which offers a variety of robust NLP services for a wide range of languages. Services can be easily accessed via a web interface, where document collections can either be uploaded and sent to batch processing, or processed in real-time via interactive calls to a web-service from any application running on a remote computer or mobile device. The services run on a high-performance-computing cluster (HPCC), and a scheduler takes care of having enough instances loaded to satisfy incoming client requests with no delay caused by initializations. TextServer offers a variety of analysis levels, ranging from tokenization or lemmatization to coreference resolution and semantic graph extraction, and aims to become a useful resource for data mining applications working on monolingual or multilingual textual data which require to obtain structured data from unrestricted text.

Read the paper · More papers on PaperTik