political-speech: A Kotlin repository from artiomi

Political speech evaluation application

Implementation details

The following application provide functionality to evaluate political candidates speeches and give response to below questions:

Which politician gave the most speeches in 2013?
Which politician gave the most speeches on "homeland security"?
Which politician spoke the fewest words overall?

In order to provide above information, was developed a REST API endpoint which receives a list of URLs, which points to CSV files with political speeches details.

  curl -X 'GET' \
  'http://localhost:8080/evaluation?url1=file:/D:/data/test-input1.csv&url2=file:/D:/data/test-input2.csv' \
  -H 'accept: application/json'

Response format:

{
 "mostSpeeches": string|null,
 "mostSecurity": string|null,
 "leastWordy": string|null
}

Input CSV file format

The CSV file should contain these headers: Speaker ; Topic ; Date ; Words which are separated by ;.
Example:

Speaker;Topic;Date;Words
Alexander Abel; education policty; 2012-10-30; 5310
Bernhard Belling; coal subsidies; 2012-11-05; 1210
Caesare Collins; coal subsidies; 2012-11-06; 1119
Alexander Abel; homeland security; 2012-12-11; 911

Execution steps

Endpoint accepts query parameters in following form: param name should have format urlN, where N is an Int > 0, param value should be a valid URL to a CSV file, currently are supported onlyfile, http, https schemas. By default, the endpoint accepts up to 5 query params, this value can be updated via app.speech.validator.max-allowed-urls application property.
Content of the CSV file to which URL points is read and mapped to a POJO.
CSV records are mapped with a batchId(random value generated for each evaluation request) and stored in database for evaluation calculations from next step. This solution was chosen in order to use SQL aggregation functionality for further calculations.
Each evaluation index calculation is done by a dedicated component. Currently, all components, run a custom SQL query for specific batchId, with additional filters (if needed) and return aggregated results per candidate. Some components accept customization via properties:
- For most speeches per year, via property app.speech.aggregation.speech-year, can be updated the year for which aggregation is made. Default:2013.
- For topic related speech, via property app.speech.aggregation.speech-topic , can be updated the topic for which aggregation is made. Default: homeland security.

Build & Run

Java - 21
Kotlin - 1.9.20
Spring Boot - 3.2.0

Build

mvn clean compile

Package

mvn clean package

Note: In order to package and run all test, it is required to have docker process running, because there are integration tests which requires database instance.
It is possible to skip tests by running:

mvn clean package -DskipTests

Run

Note: Before running the app on local, it is required to start docker instance, by running below command in project's root directory: docker-compose up -d

mvn spring-boot:run

Swagger

After app startup, Swagger documentation will be available here.

artiomi/political-speech