View Code? Open in Web Editor NEW

Latent semantic analysis model, functionality visualized on the task of finding similar articles in a database

License: MIT License

Shell 0.32% Python 12.12% Jupyter Notebook 72.82% TeX 13.42% CSS 0.05% HTML 1.26% Procfile 0.02%

lsa-model's Introduction

LSA model

Implementation of Latent semantic analysis model. Functionality visualized on the task of finding similar articles in a database.

Authors

David Mašek and Kristýna Klesnilová

Installation

$ ./init.sh

Install packages as needed on errors and rerun init.

Getting started

Running the server

$ ./run.sh # displays address and port of server (e.g. localhost:5000)

Modifying server data

replace data inside server/data/
update source code to reflect changes
recompute LSA - $ flask update

Modifying LSA parameters

$ export FLASK_APP=server # if not already set
modify values in server/lsa_config.json
recompute LSA - $ flask update

Initializing/Recomputing LSA

$ flask update # may take some time

Project structure

├── lsa                # LSA package
│   ├── data                     # selected usable data
│   ├── raw_data                 # unprocessed/large data - not saved to git
│   ├── data.ipynb               # notebook for data exploration
│   ├── lsa.py                   # LSA related code
│   └── select_articles.py       # script to select articles
├── server             # demo server
│   ├── static
│   ├── templates
│   ├── cache                    # LSA cache files used by server
│   ├── data                     # data files used by server
│   ├── __init__.py              # server + CLI initialization
│   ├── lsa_config.py            # config file with LSA parameters
│   └── views.py                 # server request handling
├── articles_source.txt          # data sources
├── init.sh                      # install this package
├── README.md
├── run.sh                       # start server
├── setup.py                     # package setup file
└── .gitignore