image/svg+xml
CLAM
Computational Linguistics Application Mediator
Maarten van Gompel, Martin Reynaert & Antal van den Bosch
ILK Research GroupTilburg Center for Cognition and CommunicationTilburg University
CLAM development is being funded by CLARIN-NL (www.clarin.nl)
NLP Application(s)( black box )
CLAM Webservice
System wrapper script
Optional Workflow Interface(not provided by CLAM!)
Service Configuration
End-Userin webbrowser
Automated Client
CLAM Client API
CLAM Client API
HTTPCLAM-XML
HTTP
HTTPCLAM-XML + XSL= xHTML
HTTP
HTTP
HTTPxHTML
CLAM Data API
- objective -
CLAM wraps around your command-line NLP tool and turns it into a fully fledged RESTful webservice with minimal effort. CLAM also acts as a modern web application, with which end-users can interact directly from their browser.
- architecture -
- service provider perspective -
What you need to write:
Service configuration file - A configurationfile for CLAM that describes the behaviourof your application in terms of profiles(input and output definitions) and parameters.
System wrapper script - A script that CLAMinvokes and which should in turn invoke youractual NLP application.
- features -
- No need to modify actual NLP application !- RESTful webservice & modern webapplication in one- Strong metadata and provenance data support- Viewers for output data- Supports long batch processes on large corpora.- Python Data API for writing system wrapper scripts- Python Client API for writing automated clients- Runs with Apache2 (WSGI)- Input/Output Conversion (input from for example PDF or MS Word)
- user perspective -
1) create a new project 2) upload input files with metadata parameters3) set global parameters 4) start the project, wait for completion 5) view/download output files
- web application -
2011-02