Smpl prjkt for chtmll 🥴

Usage

git clone https://github.com/a0s/chtmll.git
cd chtmll

bundle install
bundle exec rake db:create
bundle exec rake db:migrate

# Then extract your dataset into ./dataset like
# ./dataset/categories.json
# ./dataset/reviews.json
# ./dataset/themes.json
# and import it with:
bundle exec rake import_dataset
# or 
bundle exec rake import_dataset path=PATH_TO_DATASET

bundle exec rails s # start server

API Endpoints

GET /reviews - filter reviews by theme_ids/category_ids/comments. Example:

curl -s "localhost:3000/reviews?comments[]=find&theme_ids[]=6345&theme_ids[]=6374&category_ids[]=1223&limit=5" | jq
[
  {
    "id": 58925314,
    "comment": "... SKIPPED ...",
    "themes": [
      {
        "sentiment": 1,
        "theme_id": 6345
      },
      {
        "sentiment": 1,
        "theme_id": 6344
      },
      {
        "sentiment": 1,
        "theme_id": 6374
      }
    ]
  },

... SKIPPED ...

]

GET /reviews/avg_by_theme - average sentiment for reviews filtered by theme_ids/category_ids/comments and splitted by theme_id. Example:

curl -s "localhost:3000/reviews/avg_by_theme?comments[]=find&category_ids[]=1223" | jq
[
  {
    "theme_id": 6344,
    "avg_sentiment": 1
  },
  {
    "theme_id": 6345,
    "avg_sentiment": 1
  },
  {
    "theme_id": 6349,
    "avg_sentiment": -1
  }
]

GET /reviews/avg_by_category - average sentiment for reviews filtered by theme_ids/category_ids/comments and splitted by theme id. Example:

curl -s "localhost:3000/reviews/avg_by_category?comments[]=find&category_ids[]=1223" | jq
[
  {
    "category_id": 1223,
    "avg_sentiment": 0.8666666666666667
  }
]

POST /review - create new review for existing theme_id

curl -s -X POST --data '{"comment":"ololo", "themes":[{"theme_id":6345,"sentiment":1}]}' -H 'Content-Type: application/json' localhost:3000/review | jq
{
  "id": 59460021,
  "comment": "ololo",
  "themes": [
    {
      "sentiment": 1,
      "theme_id": 6345
    }
  ]
}

Run tests

docker-compose --file docker/compose-test.yml up --build --abort-on-container-exit --exit-code-from chtmll --renew-anon-volumes --remove-orphans 
...
chtmll_1    | ................................................
chtmll_1    |
chtmll_1    | Finished in 2.18 seconds (files took 1.49 seconds to load)
chtmll_1    | 48 examples, 0 failures
chtmll_1    |
docker_chtmll_1 exited with code 0

TODO

mass async insert over rabbitmq
better specs
endpoints for creation themes and categories

Known issues

Local development and pg gem on MacOS

brew install postgresql@9.6
gem install pg -v '1.1.4' -- --with-pg-config=/usr/local/Cellar/postgresql@9.6/9.6.16/bin/pg_config

Source dataset included data duplication, lets skip it during import

{
    "comment": "... SKIPPED ...", 
    "themes": [
        {
            "theme_id": 6374,
            "sentiment": -1
        }, 
        {
            "theme_id": 6374,
            "sentiment": 1
        }, 
        {
            "theme_id": 6350,
            "sentiment": -1
        }, 
        {
            "theme_id": 6363,
            "sentiment": -1
        }
    ], 
    "created_at": "2019-06-18T12:22:40.000Z", 
    "id": 59421588
}

chtmll

Commits

Merge pull request #8 from a0s/dependabot/bundler/nokogiri-1.11.5

Bump nokogiri from 1.10.8 to 1.11.5

Merge pull request #7 from a0s/dependabot/bundler/puma-4.3.8

Bump puma from 4.3.5 to 4.3.8

Merge pull request #5 from a0s/dependabot/bundler/websocket-extensions-0.1.5

Merge pull request #6 from a0s/dependabot/bundler/rack-2.2.3

README