language_study
Learning languages the wrong way
Currently this project is designed to build a simple probability model from sample inputs. The model can then be queried to provide options for the continuation of a sentence.
Goals:
- Provide a language adaptable interface to build models for multiple languages.
- Visualize the model to some degree.
- Provide some methods for integraing external API tools.
- Expose the model to users such that they can learn correlations in unfamiliar languages.
Notes:
- The current probability model is built on the unproven assumption that sentences are defined linearly through progression. This may be more suited to a Markov network that functions off of thought expression instead of representation.
Quickstart
- Clone the repository
- Install python dependacies
- Set environment variable(s)
DB_URI=sqlite:///localpath_to_your_db(export/set depending on your os) - Instantiate the database
$ ./dbsuite.py erase_db - Find a sample book to use and clean it.
$ ./db_suite.py clean_book "title of novel in your books directory" - Start loading in some sample data
$ ./db_suite.py push_book "title of novel in your books directory" - Follow directions to load information in and mess around with queries.