57. Local development¶

This page describes the shape of a local GAP Agentic development environment at a conceptual level. Exact commands, versions, and service endpoints belong with the code and can change, so for the authoritative setup steps follow the project's own installation documentation — see quick installation — rather than any fixed command list here.

57.1 Target of a local environment¶

A local environment exists to reproduce the same contracts the platform uses in production, connected only to non-production services, so that an engineer can produce a synthetic advisory without touching real farmer data. It is not an alternative deployment process; it mirrors the production contracts locally.

A working local setup generally provides:

• the application runtime and its pinned dependencies;
• a relational database, a cache, and an object-storage substitute;
• applied migrations and synthetic seed data;
• configuration supplied through a local, ignored environment file (never committed);
• health checks for the interface, orchestration, cache, and persistence.

flowchart LR
    M[Clean machine] --> DEP[Pinned dependencies]
    DEP --> SVC[Local DB · cache · object store]
    SVC --> MIG[Migrations + synthetic seed]
    MIG --> HC[Health checks pass]
    HC --> ADV[Synthetic advisory]

57.2 Configuration and secrets¶

Local configuration is supplied through an ignored environment file created from the project's example template. Secrets are never committed and never appear in documentation; obtain any development credentials through the project's approved process. See Running an advisory for what a successful local run should produce, and Troubleshooting for common issues.