Why Hanami?
- Speed (response time is important) — Hanami has the smaller call stack than Rails
- Safety (can't be tampered) — Hanami has very small response cycle in case of invalid requests and nice validation tools.
- Data integrity — Hanami uses dry-validation.
Data Integrity
- Validation of input parameters in
apps/api/controllers/ratings/create.rb - Validation of data in tables via checks made in migration
db/migrations/20220409074044_create_ratings.rb - Protection against tampering could be done with unique
respondent_id, for example:abcd-1234
- Create
.env.developmentwith following content
# Define ENV variables for development environment
DATABASE_URL=postgresql://postgres:example@db/nps_development
SERVE_STATIC_ASSETS="true"
WEB_SESSIONS_SECRET="abc"
- Run
make dockerize. Successful output looks like:
docker-compose up --build
Building web
[+] Building 26.1s (10/10) FINISHED
=> [internal] load build definition from Dockerfile 0.0s
=> => transferring dockerfile: 37B 0.0s
=> [internal] load .dockerignore 0.0s
=> => transferring context: 2B 0.0s
=> [internal] load metadata for docker.io/library/ruby:2.6.9-alpine 1.5s
=> [1/5] FROM docker.io/library/ruby:2.6.9-alpine@sha256:14c280549fd2bd7355c8c8d76b357760f90396c993ee1fa5d0e4fd9d0fa74e18 0.0s
=> [internal] load build context 0.0s
=> => transferring context: 39.88kB 0.0s
=> CACHED [2/5] RUN apk add --no-cache build-base postgresql postgresql-dev libpq 0.0s
=> CACHED [3/5] WORKDIR /app 0.0s
=> [4/5] COPY . /app 0.0s
=> [5/5] RUN bundle install -j $(nproc) --quiet 23.3s
=> exporting to image 1.2s
=> => exporting layers 1.2s
=> => writing image sha256:584f72b9a367db7febef14f15e64fab9ba8909db4c396b30573d85ff40ed8487 0.0s
=> => naming to docker.io/library/nps_web 0.0s
Use 'docker scan' to run Snyk tests against images to find vulnerabilities and learn how to fix them
Starting nps_db_1 ... done
Recreating nps_web_1 ... done
Attaching to nps_db_1, nps_web_1
db_1 |
db_1 | PostgreSQL Database directory appears to contain a database; Skipping initialization
db_1 |
db_1 | 2022-04-09 06:34:46.321 UTC [1] LOG: starting PostgreSQL 12.3 on x86_64-pc-linux-musl, compiled by gcc (Alpine 9.3.0) 9.3.0, 64-bit
db_1 | 2022-04-09 06:34:46.321 UTC [1] LOG: listening on IPv4 address "0.0.0.0", port 5432
db_1 | 2022-04-09 06:34:46.321 UTC [1] LOG: listening on IPv6 address "::", port 5432
db_1 | 2022-04-09 06:34:46.326 UTC [1] LOG: listening on Unix socket "/var/run/postgresql/.s.PGSQL.5432"
db_1 | 2022-04-09 06:34:46.345 UTC [23] LOG: database system was interrupted; last known up at 2022-04-09 06:31:17 UTC
db_1 | 2022-04-09 06:34:46.702 UTC [23] LOG: database system was not properly shut down; automatic recovery in progress
db_1 | 2022-04-09 06:34:46.704 UTC [23] LOG: redo starts at 0/1650868
db_1 | 2022-04-09 06:34:46.704 UTC [23] LOG: invalid record length at 0/1650998: wanted 24, got 0
db_1 | 2022-04-09 06:34:46.704 UTC [23] LOG: redo done at 0/1650960
db_1 | 2022-04-09 06:34:46.719 UTC [1] LOG: database system is ready to accept connections
web_1 | [2022-04-09 06:34:48] INFO WEBrick 1.4.4
web_1 | [2022-04-09 06:34:48] INFO ruby 2.6.9 (2021-11-24) [x86_64-linux-musl]
web_1 | [2022-04-09 06:34:48] INFO WEBrick::HTTPServer#start: pid=1 port=2300
- While your docker container is running, connect to it from another terminal with command
make shell. Successful output:
docker-compose exec web sh
/app #
- Run
hanami db createandhanami db prepare
Name Method Path Action
POST /api/ratings Api::Controllers::Ratings::Create
GET, HEAD /api/ratings Api::Controllers::Ratings::Index
After successful make dockerize you should be able to send requests to http://localhost:2300
Use some kind of REST client for your browser. For example RESTED for Google Chrome.
Use POST /api/ratings for creating rating and updating (whether it is a duplicate submission): use cases A+B.
Example JSON:
{"rating": {"score": 1, "touchpoint": "realtor_feedback", "respondent_class": "seller", "object_class": "realtor", "respondent_id": 123324, "object_id": 321141}}
Use GET /api/ratings for fetching list of ratings by touchpoint, respondent_class, object_class: use case C.
Example request string:
http://localhost:2300/api/ratings?touchpoint=realtor_feedback&respondent_class=seller&object_class=realtor
- Dockerfile and Makefile https://github.com/gruz0/hanami-docker