AsyncFactorial

Scala service for factorial processing with three modes:

• Kafka consumer/producer: consumes factorial.tasks, produces to factorial.results.
• HTTP server: consumes JSON and responds with factorial results.
• Console mode: reads numbers from stdin, writes results to out.txt asynchronously.

Quick Start

1. Clone repo into a sibling folder of PRAssign (for Go/Kafka stack):

   git clone https://github.com/UsatovPavel/AsyncFactorial.git
   cd AsyncFactorial

2. Create .env:

   # HTTP сервис Scala 
   SERVER_HOST=0.0.0.0
   SERVER_PORT=8080
   SERVER_PARALLELISM=4   # empty => use availableProcessors
   
   # Kafka service
   KAFKA_BOOTSTRAP_SERVERS=kafka1:9092,kafka2:9092,kafka3:9092
   KAFKA_GROUP_ID=asyncfactorial-consumer
   KAFKA_INPUT_TOPIC=factorial.tasks
   KAFKA_OUTPUT_TOPIC=factorial.results
   CONSUMER_REPLICAS=4
   
   KAFKA_EXPOSE_PORT=9092  # for local-kafka profile only

3. Run Kafka consumer (needs Go/Kafka network prassign_default):

   make service

   (reads .env, scales consumer to CONSUMER_REPLICAS).

Modes

• Kafka service: IOApp KafkaConsumerTask, consumes factorial.tasks, publishes factorial.results. Config via env/application.conf.
• Console:
  • IOApp Task parses integers from stdin until exit, runs factorials concurrently (non-blocking input), appends results to out.txt (order not guaranteed), logs parse errors.
  • Two modes: waitAll=true waits for all tasks on exit; waitAll=false cancels unfinished work and stops cleanly.
  • Built on cats-effect fibers with Supervisor (lightweight, easy to spawn).
• HTTP: POST /factorial (Tapir + Vert.x), body = JSON array of Int, optional header X-Job-Id (if missing, generated). Swagger: /docs

Tech Stack

• Language: Scala 2.13
• Concurrency/Streams: cats-effect 3, fs2
• Kafka: fs2-kafka
• HTTP: tapir
• JSON: circe
• Test: scalatest (with cats-effect)
• Build tools: Docker Compose (external prassign_default), Makefile

Env / Config

• Defaults in application.conf (server.*, kafka.*).
• Runtime overrides via env (SERVER_*, KAFKA_*, CONSUMER_REPLICAS, KAFKA_EXPOSE_PORT).

Run

• Kafka consumer: make service (requires .env, uses KAFKA_BOOTSTRAP_SERVERS, CONSUMER_REPLICAS).
• Local single-broker profile (optional): docker compose --profile local-kafka up.
• HTTP mode (if needed): docker compose --profile http up app (expects env SERVER_*).

Testing

make test

• Console pipeline: TaskSpec checks input/parse/write, waitAll true/false, performance
• Writer: NumberWriterSpec — verifies how the writer logs parse errors and normal results, and that it stops cleanly on Shutdown
• Core: FactorialAccumulatorSpec covers factorial on valid/invalid inputs and parse errors.
• HTTP: smoke/bulk/invalid-json/parallel/big-job-id specs for POST /factorial

Notes

• Compose networks.default is external prassign_default to talk to Go/Kafka stack.