egmaziero / dizer3

Rhetorical Structure Theory Parser (for Brazilian Portuguese)

DiZer 3

dizer3 is a discoursive analyzer for Brazilian Portuguese. It is under development...

Installation

• Clone this repository git clone https://github.com/egmaziero/dizer3.git.

• Install dependencies (installing inside a virtual environment is advisable...) pip install -r requirements.txt.

• Download models using the link: https://drive.google.com/file/d/1z9WuR5USFHrangVNGpvMEA2oyjlGob8A/view?usp=sharing. Extract the directory modelsand put it in dizer3 directory.

Basic usage

dizer3 performs discourse (based on Rhetorical Structure Theory) analysis on raw text files.

To run the parser use:

python3 dizer3\run_pipeline.py --f file_path

or

python3 dizer3\run_pipeline.py --d directory_path

where

• --f indicates the a raw text file path. This argument takes precedence over -d
• --d indicates the directory with raw text files to be analyzed

Citation

To cite this work, use one of the following reference:

Maziero, E.G.; Hirst, G.; Pardo, T.A.S. (2015). Semi-Supervised Never-Ending Learning in Rhetorical Relation Identification. In the Proceedings of the Recent Advances in Natural Language Processing - RANLP, pp. 436-442. September 5-11. Hissar/Bulgaria.

Maziero, E. G. (2016). Análise retórica com base em grande quantidade de dados. Tese de Doutorado, Instituto de Ciências Matemáticas e de Computação, Universidade de São Paulo, São Carlos. doi:10.11606/T.55.2017.tde-13012017-103446. Recuperado em 2020-06-10, de www.teses.usp.br