starcoach-hackathon/zarinpal-analytics-deployment
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

feat: start scheduler in local development

Browse Source
This commit is contained in:
Amirhossein Khalili
2026-08-22 17:25:00 +08:00
parent 35a1101c8b
commit e2decb9e01
3 changed files with 8 additions and 4 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
docs/LOCAL-DEVELOPMENT.md
View File

@@ -1,13 +1,13 @@
# Local development without Docker Compose
This Windows-only helper runs the API, analytics worker, and frontend directly from the three sibling repositories. PostgreSQL and Redis must already be reachable locally. It does not create databases, run migrations/imports, generate credentials, edit service-owned environment files, or start Docker.
This Windows-only helper runs the API, analytics worker, growth-coach scheduler, and frontend directly from the three sibling repositories. PostgreSQL and Redis must already be reachable locally. It does not create databases, run migrations/imports, generate credentials, edit service-owned environment files, or start Docker.
## Prerequisites
- Go and npm are on `PATH`.
- PostgreSQL is reachable at `127.0.0.1:5432`.
- Redis is reachable at `127.0.0.1:6379`.
- The backend has `cmd/api` and `cmd/worker`, plus a protected backend-owned `.env`.
- The backend has `cmd/api`, `cmd/worker`, and `cmd/scheduler`, plus a protected backend-owned `.env`.
- The frontend has a `package.json` `dev` script and its frontend-owned local environment file.
- Dependencies, migrations, the one-time importer, and demo merchant creation have already completed through their owning repositories.
@@ -41,6 +41,6 @@ After `start`, open the frontend URL declared by the frontend development server
The AI growth analyst is controlled by service-owned environment files; no AvalAI credential belongs in this deployment repository or in the frontend.
For local deterministic verification, set `AI_GROWTH_ENABLED=true` in the ignored backend `.env` and `PUBLIC_AI_GROWTH_ENABLED=true` in the ignored frontend `.env`. With `AVALAI_API_KEY` empty, the worker must not make a provider request: it persists and serves a deterministic `PARTIAL` report instead.
For local verification, set `AI_COACH_ENABLED=true`, `AI_PAGE_INSIGHTS_ENABLED=true`, and `AI_GROWTH_ENABLED=true` in the ignored backend `.env`, and `PUBLIC_AI_GROWTH_ENABLED=true` in the ignored frontend `.env`. With `AVALAI_API_KEY` empty, no generation is misrepresented as a fresh AI response.
The backend `.env.example` is authoritative for the complete contract. The production operator must explicitly provide `AVALAI_API_KEY`, `AVALAI_MODEL`, `AI_DAILY_REQUEST_BUDGET`, and `AI_DAILY_TOKEN_BUDGET` before enabling the feature. `AVALAI_BASE_URL`, `AI_MAX_CONCURRENT`, `AI_CREATE_PER_MINUTE`, and `AI_UNCACHED_SCOPES_PER_DAY` remain configurable safety limits. Never place `AVALAI_API_KEY` in a public frontend variable.