Author: Eugene Yang
This is a course project of Information Retrieval in Georgetown University.
- This project is run under nodeJS v7.0.0.
Fast and clean way to install nodeJS on Linux:
apt-get install nvm
nvm install node
- Please install dependencies by npm using the following command under the project directory
npm install
- In order to run container in a safe and clean way, please expose garbage collection function by adding argument
--expose-gc
when executenode
.
-
Configurations are listed in
config.json
, including memory constraint and flush bunch settings. -
Please assign -1 to
memory_limit
if do not want any memory constraint. -
Default similarity parameter settings are listed in
config.json
.
For building index:
node build [trec-files-directory-path] [index-type] [output-dir]
Or for using default settings, you can execute build.bat
on Windows or build.sh
on Unix-like systems.
For executing query file:
node query_search [index-directory-path] [query-file-path] [retrieval-model] [index-type] [search-type] [optional: results-file]
Or use default directories:
node query_search -a [retrieval-model] [index-type] [search-type] [optional: results-file]
Apache License 2.0