Riksdagens Protokoll Part-Of-Speech Tagging (Parla-Clarin Workflow)

This package implements Stanza part-of-speech annotation of Riksdagens Protokoll Parla-Clarin XML files.

Prerequisites

• A bash-enabled environment (Linux or Git Bash on windows)
• Git
• Python 3.8.5^
• GNU make (install i)

Parla-Clarin to penelope pipeline

How to install

How to configure

How to setup data

Riksdagens corpus

Create a shallow clone (no history) of repository:

make init-repository

Sync shallow clone with changes on origin (Github):

make update-repositoryupdate_repository_timestamps

Update modified date of repository file. This is necessary since the pipeline uses last commit date of each XML-files to determine which files are outdated, whilst git clone sets current time.

$ make update-repository-timestamps
or
$ scripts/git_update_mtime.sh path-to-repository

How to annotate speeches

make annotate
or
$ nohup poetry run snakemake -j4 --keep-going --keep-target-files &

Windows:

poetry shell
bash
nohup poetry run snakemake -j4 -j4 --keep-going --keep-target-files &

Run a specific year:

poetry shell
bash
nohup poetry run snakemake --config -j4 --keep-going --keep-target-files &

Install

(This workflow will be simplified)

Verify current Python version (pyenv is recommended for easy switch between versions).

Create a new Python virtual environment (sandbox):

cd /some/folder
mkdir westac_parlaclarin_pipeline
cd westac_parlaclarin_pipeline
python -m venv .venv
source .venv/bin/activate

Install the pipeline and run setup script.

pip install westac_parlaclarin_pipeline
setup-pipeline

Initialize local clone of Parla-CLARIN repository

Run PoS tagging

Move to sandbox and activate virtual environment:

cd /some/folder/westac_parlaclarin_pipeline
source .venv/bin/activate

Update repository:

make update-repository
make update-repository-timestamps

Update all (changed) annotations:

make annotate

Update a single year (and set cpu count):

make annotate YEAR=1960 CPU_COUNT=1

Configuration

work_folders: !work_folders &work_folders
  data_folder: /data/riksdagen_corpus_data

parla_clarin: !parla_clarin &parla_clarin
  repository_folder: /data/riksdagen_corpus_data/riksdagen-corpus
  repository_url: https://github.com/welfare-state-analytics/riksdagen-corpus.git
  repository_branch: dev
  folder: /data/riksdagen_corpus_data/riksdagen-corpus/corpus

extract_speeches: !extract_speeches &extract_speeches
  folder: /data/riksdagen_corpus_data/riksdagen-corpus-exports/speech_xml
  template: speeches.cdata.xml
  extension: xml

word_frequency: !word_frequency &word_frequency
  <<: *work_folders
  filename: riksdagen-corpus-term-frequencies.pkl

dehyphen: !dehyphen &dehyphen
  <<: *work_folders
  whitelist_filename: dehyphen_whitelist.txt.gz
  whitelist_log_filename: dehyphen_whitelist_log.pkl
  unresolved_filename: dehyphen_unresolved.txt.gz

config: !config
    work_folders: *work_folders
    parla_clarin: *parla_clarin
    extract_speeches: *extract_speeches
    word_frequency: *word_frequency
    dehyphen: *dehyphen
    annotated_folder: /data/riksdagen_corpus_data/annotated