rankt is a suite of tools for tracking and testing search relevancy
rankt requires Python 3
`pyvenv venv`
`pip install -r requirements.txt`
rankt init -h for more info.
Basic example:
rankt init --search-url [the url of your solr core select handler]
rankt create -h for more info.
A case is used to organize relevancy tuning.
rankt create [name of case, e.g., top100_keywords]
You will now have a cases/top100_keywords/queries path for your query experiments.
Queries support jinja2 template syntax.
rankt snapshot -h for more info.
Basic example:
rankt snapshot [case name] --query [query template filename] --keywords [path to keywords file]
This will save the document at each position for each keyword in --keywords file.