Skip to content
Service and mobile app for handling corporate reimbursements
Branch: master
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Type Name Latest commit message Commit time
Failed to load latest commit information.


The main backend service for the Kulu app.

Set up

  • Clone the app

  • Create the secret config from the sample:

cp config/kulu-backend-secrets-sample.edn config/kulu-backend-secrets.edn
  • Fill in the placeholder vals in the above file. Read on to learn more about our configuration management.

  • Ensure that you're running postgres 9.5 or 9.6 (postgres >=10 is not supported)

  • Set up dev and test dbs:

$ psql
create database kulu_backend_dev;
\c kulu_backend_dev
create extension "uuid-ossp";

create database kulu_backend_test;
\c kulu_backend_test
create extension "uuid-ossp";
  • lein deps to get all the dependencies
  • Migration script expects DATABASE_URL in the enviroment
DATABASE_URL=postgres://$USER\:@localhost:5432/kulu_backend_dev lein clj-sql-up migrate


In dev, run bin/server-start. It sets up some needed env vars and starts the server with lein. Or you can directly run lein run -m kulu-backend.web 3001 (takes defaults for the env vars).

Once started, visit http://localhost:3001/

In heroku

  • Push to heroku with git push heroku master
  • Interact on
  • To run migrations, heroku run lein clj-sql-up migrate

Developer info

Peculiarities in requests and responses

You may have noticed that our HTTP requests come in with under_scored params but everywhere inside the app we use the kabab-case. This is done with kulu-backend.utils.api/idiomatize-keys and it's friends:

  • On incoming HTTP requests (at the handler level) we convert params to the kabab-case (dasherize them).
  • Once the response is ready, we convert the response body back to the under_score flavor.

See example here and read the code documentation here for more details.

Similarly, we apply the inverse transform (under_scorize on the way in, dasherize on the way out) while sending data to SQL.


We use nomad to read our app's configuration from edn files. The config related files are:


In prod, we read the DATABASE_URL env var. The dev and test databases are specified in the config. In test, every DB command is run in a transaction and rolled back after the test, so we always a clean slate.

The connection string in test comes from Nomad configs

(expand further)


We use seperate IAM accounts for dev and prod (we stub out AWS usage in test).

See the AWS Console's IAM tab (for for the various IAM accounts we use. Similarly we have separate S3 buckets and SQS queues for dev/prod and IAM account's permissions make sure that the dev AWS IAM account can't read/write from the prod resources and vice-versa.


  • brew install elasticsearch

  • Install the KOPF UI plugin to handle all ES administration: in /usr/local/var/lib/elasticsearch/plugins

  • $ lein run -m kulu-backend.tasks setup dev to migrate indices.

  • Run elasticsearch

  • Go to and download the latest version of elasticsearch (assumingly as zip).

  • unzip it.

  • chmod +x bin/elasticsearch

  • $ lein run -m kulu-backend.tasks setup dev to migrate indices.

  • bin/elasticsearch

  • this starts a process that binds to port 9200.

  • curl -X GET http://localhost:9200/

to get something like:

  "status" : 200,
  "name" : "Mop Man",
  "version" : {
    "number" : "1.3.4",
    "build_hash" : "a70f3ccb52200f8f2c87e9c370c6597448eb3e45",
    "build_timestamp" : "2014-09-30T09:07:17Z",
    "build_snapshot" : false,
    "lucene_version" : "4.9"
  "tagline" : "You Know, for Search"


Run lein test to run tests. It will automatically set the NOMAD_ENV=test and pick up the right configuration.

Make sure you run elasticsearch before running tests, as some of the tests in the suite test end-to-end making real calls to ES on a test index.

Companion Services

You can’t perform that action at this time.