subs2strudel

Creating semantic concept-feature norms using STRUDEL

Collecting the Data

Below is the process you need to run in order to contribute a language to the collection.

1. Open up process_strudels.Rmd in the root (~/) folder. It is the same folder that contains this README.md.
2. Select a language not currently completed. See ~/data/udpipe_languages.csv's Completed column.
3. Select a number of sub processes. Each sub process takes one core and ~2GB of RAM.
4. Run, NOT KNIT, everything. This will:
   • Download a new language file into the ~/data folder.
   • Download a new language udpipe control file into the ~/ folder.
   • Splits the langage file into smaller files for parallel processing. This makes more files in the ~/data folder.
   • Runs each smaller file in its own process. This generates files in the ~/concept-feature folder.
   • Combines the files into a single file.
5. Upload the combined file to the releases in GitHub.
6. Update the releases and ~/data/udpipe_languages.csv noting the progress.