Wiki
Clone wikiparsimonious-vole / Home
Introduction
Parsimonious Vole is a simplified systemic functional parser.
Dependencies
- Python 2.7 or latter. Python 3.x not tested.
-
GraphViz: sudo apt-get install graphviz libgraphviz-dev pkg-config
- Stanfor NER
- pyner - Python wrapper for Stanford NER annotator
- python-matplotlib
Pthon libraries:
Usage
The input to PV are the dependency parses produced by Stanford Parser. Currently 2 formats are supported: the XML or STP. The XML file can be produced either by running Stanford parser locally or from online demo version here. The STP file can be generated with magic.parse.stanford_parser.py, see example below.
Specify path to Stanford Parser in the environment variable "STANFORD_PARSER_HOME". Alternatively you can edit configuration.py and specify the path to it.
- Generating the input file:
#!python from magic.parse.stanford_parser import call_stanford_parser_file("/path/to/file/input.txt") # generates the STP file next to the TXT file
- Generating a nice HTML with Mood Constituency Graphs:
#!python from usage import output_analisys_of_parse_bundle_to_file output_file = "path/to/output.html" input_file = "path/to/input.stp" # generates the HTML file with analysis result output_analisys_of_parse_bundle_to_file(output_file, input_file)
This is a command line tool which generates a CSV file with OCD related analysis for all STP/XML files located in a given directory.
The script file(mood_graph_to_csv.py) is located in usage package and it takes exactly one argument - the path to the folder containing the STP/XML parses (the Stanford Parser output).
#!bash
python mood_graph_to_csv.py path/to/input.stp
- Mood Parsing:
#!python from magic.guess.mood import test_mood_graph sentence_id = 1 path_to_stp_input_file = "/path/to/input.stp" correct_the_input_file = True mcg = test_mood_graph(sentence_id, path_to_stp_input_file, correct_the_input_file )
- Transitivity Parsing:
#!python from magic.guess.transitivity import parse_transitivity_from_mood_graph # enriching the MCG with transitivity features res = parse_transitivity_from_mood_graph(mcg)
Recommendations
If you are in development mode I recommend using Eclipse IDE together with PyDev plugin a plugin for Python.
Future Developments
- add WSD by PyWSD
- add TempEx alignment
Notes:
Updated