CAVaT toolkit: Corpus Analysis and Validation for TimeML
LICENSE:
Copyright 2009-2012 Leon Derczynski
Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with the License. You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the specific language governing permissions and limitations under the License.
To cite or attribute CAVaT in your work, please include the following information:
@inproceedings{derczynski2010cavat, title = {{Analysing Temporally Annotated Corpora with CAVaT}}, booktitle = {Proceedings of the 7th International Conference on Language Resources and Evaluation}, year = {2010}, author = {Derczynski, L. and Gaizauskas, R.}, pages = {398--404} }
REQUIREMENTS:
Global: Python 2.6.4 - 2.x
Python add-ons: NLTK 2.0b2+ PyParsing 1.5.2+ SQLite 3 (or MySQL)
INSTALLATION: Unpack cavat into its own directory. Databases will be stored in ~/.cavat/ .
RUNNING CAVAT: Change into CAVaT's main directory, and run "python cavat.py". When the system has loaded, you should see a "cavat>" prompt. You can quit with Ctrl-C, "quit" or "exit".
From here, if you have no corpora loaded, you may want to import one. Enter the following to load a TimeML corpus into the "test" database - it's important to include the trailing / in the path:
cavat> corpus import /home/user/corpus/data/ to test
Depending on your disk and CPU speeds, this might take about 10-20 seconds per megabyte of TimeML. If it seems to take longer, you can get more information about what CAVaT is doing during import by enabling debug mode before import:
cavat> debug on
Leave debug mode with a simple:
cavat> debug off
Once the corpus has loaded, you can use "corpus info" to see metadata about the
import, or "corpus list" to see an available list of corpora. To switch between
corpora, and to select a newly loaded one, enter "corpus use
QUERIES:
The show command is used for generating reports on the currently loaded corpus. Reports focus on one tag type, and give information about their attributes. One can view all values for a tag with "list" reports, or the distribution of values with "distribution" reports, or simply see how many instances of that tag list a value for a field with "state" reports.
Reports can be provided in multiple formats; there is: screen - for screen or fixed-width font output csv - comma separated values tex - LaTeX table format
The general format for report generation is:
show
To try a simple query, enter:
cavat> show distribution of tlink reltype
You should see a table listing the values listed for relType in the current corpus' TLINK tags, as well as their frequencies. To see how many TLINKs use a signal, and use the result in a LaTeX document, you can try:
cavat> show state of tlink signalid as tex
FURTHER HELP:
Try the Google Code pages for CAVaT: http://code.google.com/p/cavat/
Or email the author: leon@dcs.shef.ac.uk