中文

GraphQL service for neuroscience experimental data management (Monorepo).

CogMatrix is designed to provide a unified way to organize, manage, and query neuroscience experiment datasets (subjects, tasks, sessions, recordings, annotations, and metadata).

The project is inspired by alyx, implemented with Node.js + TypeScript, and uses GraphQL (instead of a REST-only API) for flexible, researcher-oriented data access.

CogMatrix provides a unified way to organize and query neuroscience experiment data.

• Request only required fields to avoid over/under-fetching
• Model complex relationships naturally (experiment -> session -> recording -> annotation)
• Aggregate related sources in a single query
• Improve exploratory workflows with self-documented schema
• Evolve schema incrementally without breaking existing workflows

The system targets research-oriented, interactive data workflows rather than extreme QPS.

• Developer/researcher query experience over peak throughput
• Low-latency interactive querying for daily exploration
• Predictable behavior over premature micro-optimizations
• Clarity and maintainability of data modeling

• Node.js >= 20
• pnpm

pnpm install

From repository root:

pnpm --filter api start:dev

From apps/api:

pnpm run start:dev

Default server port is 3000 (or PORT if provided).

From repository root:

pnpm --filter api build
pnpm --filter api test
pnpm --filter api test:e2e
pnpm --filter api lint

The following starter notes are intentionally kept for now and can be deleted later when repository cleanup is complete:

• Template origin: Turborepo with Vite React example
• Existing utility stack: TypeScript, ESLint, Prettier