mathtrail-task

Task factory — turns Mentor's strategic parameters into concrete math problems by orchestrating LLM generation and template selection.

Mission & Responsibilities

• Receive generation parameters from Mentor (type, difficulty, topic)
• Select task generation method (LLM via taskgen, or template DB)
• Assemble final task with metadata (hints, solution, difficulty)
• Serve tasks to UI via REST API
• Track task lifecycle (created → assigned → attempted → completed)

Tech Stack

• Language: Go
• Framework: Gin (HTTP), GORM (ORM)
• Database: PostgreSQL (task templates, task instances)
• Events: Kafka pub/sub

API & Communication

• Inbound: REST API (GET /tasks/next, GET /tasks/{id}), HTTP invocation from UI
• Outbound: HTTP invoke → mathtrail-llm-taskgen (generate problem)
• Subscribes: mentor.strategy.updated (receive generation parameters)
• Publishes: task.attempt.completed (notify Profile + Mentor)

Data Persistence

• PostgreSQL: task_templates, task_instances, task_attempts

Secrets

• DB_PASSWORD — PostgreSQL password
• Vault path: secret/data/{env}/mathtrail-task/

Infrastructure

Standard infra/ layout:

• infra/helm/ — Helm chart + environment overlays (dev, on-prem, cloud)
• infra/terraform/ — Database module (K8s pod vs managed RDS)
• infra/ansible/ — On-prem node preparation

Development

• just dev — Skaffold dev loop
• just test — Go unit tests

Roadmap

1. Define Task data model (templates, instances, attempts)
2. Implement Kafka subscriber for mentor.strategy.updated
3. Build REST API for task delivery (GET /tasks/next with difficulty filter)