What new

From Anote2Wiki
Jump to: navigation, search

Short list of what's new in @ Note2

Summary of features - Anote2 vs Anote

General:

  • Eliminates dependency on Perl; fully implemented in Java; easier to install and run in different platforms;
  • MySql database used to store all relevant information (including annotations); data model extended and re-organized to provide the core of the application;
  • Organization in a set of plug-ins taking advantage of AI Bench plug-in system; improves modularity; makes easier adding new functionality and alternatives for different processes;
  • Code was re-organized using Java interfaces for all relevant concepts and processes, making easier the integration of third party applications; MVC notions from AI Bench are better used to provide organized GUIs; core plug-ins define basic data structures and the interface of basic processes;
  • Indication of progress in time consuming processes.

PubMedSearch

  • Now a plug-in; new implementation more efficient based on eUtils.
  • Query management system; larger number of options available
  • Possibility of including external publications
  • Possibility of updating the results of queries
  • System to avoid publication crawling

Resources

  • Better management of lexical resources (dictionaries, lookup tables, etc)
  • Introduction of rules, ontologies and general lexical words (Stop Words)
  • More loaders and an interface allowing the easy implementation of new loaders

Corpus / Corpora

  • Notion of corpus/ corpora added to @Note in a new plug-in
  • Allows creating corpus from the publication manager
  • Information extraction processes are-organized over corpora
  • Corpus loader: Allows creating corpus from other available corpora (Genia Events, AIMED Protein and @NoteV1).

NER Resources

  • Developed fully in Java and adapted to the new resources
  • External NERs can be added with an easier interface

Curator

  • Organized in a new plug-in
  • Uses only HTML

Relation

  • New Plug-in for Relation Extraction based in Natural Language processing
  • Supports RE over corpora annotated by any NER Process