lughen12/query-tool

query-tool

This CLI is made to fullfill screening accessment for 90 Seconds.

Prerequisite

1. Install ruby (tested with ruby 2.2.6)
2. Install bundle gem install bundler
3. Install dependancies in project directory bundle install

Example Usage

1. Get all results for resource

ruby main.rb -r user

2. Query a resource

ruby main.rb -r user -q name="Raylan Givens",tags=Springville

3. Query a resource with limit (default is 5)

ruby main.rb -r user -l 1

4. Query a resource with offset (default is 0)

ruby main.rb -r user -o 1

5. Change output format (default: text)

ruby main.rb -r user -f json

6. Display helps

ruby main.rb -h

7. Run tests

bundle exec rspec spec/

Evaluation Criteria

1. Extensibility (separation of concerns) The project is splitted into these following directories:

• models: Classes to represent resources. Currently only support user and project. For supporting new resource, just create another model class.
• presenter: Classes to format ruby hash. Currently only support formatted text and json. For supporting new format, just create another presenter class.
• services: Classes to do some specific logics. Currently there are 2 services. CommandService to parse CLI command to options hash, and QueryService to execute query from the option hash.
• specs: All the test in the application.

2. Simplicity

• Minimal dependancies

3. Test Coverage

• 100% coverage for models, presenters and services

4. Performance

• Support limit and offset as CLI options. When data set become larger, user can paginate their query commands.
• The solution load the whole file to memory, and loop through the whole array (O(n) in time complexity) to perform a query. This approach will not work well if data set increases.
• Building an in-memory tree before searching has been taking into consideration, so that search can be faster. However, building an in-memory tree costs time and memory (to hold that data stucture), and it is redundant in real life, because query optimization should be a job of a database engine.
• To improve search speed when data set increase, we should use a proper DB. So we can rely on it indexes to make query faster.
• In memory cache or Redis cache of a search results might be helpful too, if the same search is reapeat frequently in the application.

5. Robustness

• Report errors for invalid commands, missing resources, invalid resources, invalid format, invalid query ...

6. Extras

• -f option to return different format. It is good if user can select output format that they need
• Pretty print results, errors with color. This will made CLI output easier to read