Analyzes commits to the OpenMRS repositories, e.g for use in an annual report
- docker
- nodejs
- npm
- Run the elasticsearch docker container:
// maybe you've done this before...
sudo docker start es-ac
// make sure you have the latest
sudo docker pull docker.elastic.co/elasticsearch/elasticsearch:6.7.0
// this will not preserve data across restarts
sudo docker run --name es-ac -p 9200:9200 -p 9300:9300 -e "discovery.type=single-node" docker.elastic.co/elasticsearch/elasticsearch:6.7.0
- Download dependencies from npm
npm install
- Fetch the data from github and load it into elasticsearch (use your GitHub username and password to avoid being rate-limited by the GitHub REST API)
node fetch-from-github.js
If you want more details about committers
node fetch-committers-from-github.js
node search-committers.js bahmni // full-text search for "bahmni" on anyone who committed to OpenMRS
- To see one committer's work over time
node analyze-user.js username
- To analyze who committed how much, in a given year
node analyze-committers-for-year.js year [repo]
- The real analysis we want (you may need to edit this file to set the correct YEAR constant)
node analyze-by-quarter.js
Kibana makes is trivial to do ad-hoc analysis on the commit data.
TODO change this to use an image from the docker.elastic.co registry
docker run --name k-ac --link es-ac:es-ac -e ELASTICSEARCH_URL=http://es-ac:9200 -p 5601:5601 -d kibana:latest
// it should be running on http://localhost:5601
// in the UI you have to do Settings, then set index name to "commits", choose the "date" field, and click Create
// then click Discover and change the time frame from the top right
There's a toy webapp for viewing this data as a dashboard. (I wanted to see what writing a node.js webapp using Express was like.)
cd webapp
// only need to do this once
npm install
// if elasticsearch is running on http://192.168.99.100:9200 for you like it is for me:
npm start
// otherwise specify the elasticsearch host, like
npm start http://localhost:9200
// the webapp will start on http://localhost:3000/