A Corpus of Grand National Assembly of Turkish Parliament's Transcripts

To reference this work, use:

@InProceedings{GUNGOR18.19,
author = {Onur Gungor ,Mert Tiftikci and Çağıl Sönmez},
title = {A Corpus of Grand National Assembly of Turkish Parliament's Transcripts},
booktitle = {Proceedings of the Eleventh International Conference on Language Resources and Evaluation (LREC 2018)},
year = {2018},
month = {may},
date = {7-12},
location = {Miyazaki, Japan},
editor = {Darja Fišer and Maria Eskevich and Franciska de Jong},
publisher = {European Language Resources Association (ELRA)},
address = {Paris, France},
isbn = {979-10-95546-02-3},
language = {english}
}

==== Scroll down for English.

Bu repoda TBMM'nin 1920 ve 2015 arasındaki genel kurul tutanaklarının dijitalleştirilmiş hallerinin bir derlemini bulabilirsiniz.

Sunduğumuz kodlar kullanılarak bazı sorgular yapıp derlemi analiz edebilirsiniz. Bunun örneklerine LREC 2018 kapsamında gerçekleştirilen ParlaCLARIN çalıştayında sunduğumuz makaleden ulaşılabilir.

Eğer katkıda bulunmak isterseniz, Onur Güngör ile iletişime geçin.

==== English

This is a compilation of the transcripts of Grand National Assembly of Turkish Parliament (TBMM) meetings which span nearly a century between 1920 and 2015.

One can use supplied software to query and analyze the content. See this paper from ParlaCLARIN Workshop at LREC 2018 for sample analyses on the corpus.

If you want to contribute, please contact Onur Güngör.

We also provide the most recent version of the corpus at Releases section.

Contributing

If you want to contribute, you can follow these simple steps to create a development environment:

git clone https://github.com/onurgu/turkish-parliament-texts/releases

wget http://voltran.cmpe.boun.edu.tr/temporary_download/datasets/tbmm/corpus-dev.tar.gz
tar -zxvf corpus-dev.tar.gz

pip3 install pipenv
pipenv install --python 3

Building the corpus

nohup python -m corpus_compiler.builder --command construct_corpus --corpus_filename corpus-v0.4b/tbmm_corpus.mm --train_lda --vocabulary_filename corpus-v0.4b/vocabulary >> construct_vocab.nohup &

Using the Jupyter notebook

You can use the notebook.ipynb file to load and query the corpus.

Example code to save a figure (using the small development corpus):

pipenv shell
ipython
import corpus_loader

corpus = corpus_loader.load(corpus_filepath="./corpus-dev/tbmm_corpus.mm")

lda, topic_dist_matrix, label_vector = corpus_loader.load_lda_model(corpus, "./corpus-dev/tbmm_corpus.mm.tbmm_lda.model")

# plot distribution of "mebus" and "milletvekil" keywords
corpus_loader.corpus.plot_word_freqs_given_a_regexp_for_each_year([r"^mebus",r"^milletvekil"], ["mebus", "milletvekili"], keyword="milletvekil_and_mebus",)

# to see plotted values 
plot_values, counts, total_count, all_keywords = corpus_loader.corpus._word_freqs_given_a_regexp_for_each_year(r"^(milletvekil|vekil)", keyword="milletvekil",)