# Project-specific configuration

This repo (`ia_dev`) is intended to be used as a **git submodule** inside each project. Project-specific parameters are stored in `projects/<id>/conf.json` (e.g. `projects/lecoffreio/conf.json`). The `<id>` is the project identifier and the name of the directory under `projects/`.

## Current project selection

There is no longer a "slug" in the URL or path; the project **id** (directory name under `projects/`) is resolved by:

**1. Mail ticketing**  
From the **To** address of an email: search all `projects/*/conf.json` for `tickets.authorized_emails.to` equal to that address (case-insensitive). The matching project directory name is the id.

**2. API ai_working_help**  
The token is found by scanning as described above: traverse all projects and all envs, reading each `projects/<id>/.secrets/<env>/ia_token` and comparing its content (or content + env) to the Bearer token. The first match yields the project id and the env (e.g. `test`, `pprod`, `prod`).

**3. Scripts / fallback** (when no request context)  
- **`MAIL_TO`** (env): same as (1), resolve id by email "to".  
- **`AI_AGENT_TOKEN`** (env): same as (2), resolve id (and env) by token; sets `PROJECT_SLUG` and `PROJECT_ENV`.  
- **`IA_PROJECT`** (env)  
- **`.ia_project`** at repository root (one line)  
- **`ai_project_id`** at repository root (one line). When `ia_dev` is a submodule, this file lives at the host repo root (parent of `ia_dev`).

When running from a repo that has `ia_dev` as a submodule, the root is the parent repo; the script resolves `ia_dev` either as `./ia_dev` or `./deploy` (symlink to `ia_dev/deploy`).

## Schema

One JSON file per project: `projects/<id>/conf.json` (e.g. `projects/lecoffreio/conf.json`). The `<id>` is the directory name; the config file is always named `conf.json`.

| Field | Required | Description |
|-------|----------|-------------|
| `name` | yes | Human-readable project name |
| `project_path` | no | Relative path to project from ia_dev (e.g. `../lecoffre_ng_test`); used when running from ia_dev standalone |
| `build_dirs` | no | List of directories (relative to repo root) where `npm run build` is run before push. If missing or empty, build check is skipped |
| `version` | no | Version/bump configuration |
| `version.package_json_paths` | no | List of paths (relative to repo root) to `package.json` files to update on bump |
| `version.splash_app_name` | no | App name used in splash message template |
| `mail` | no | Mail/imap bridge config |
| `git` | no | Git hosting: `wiki_url`, `token_file` (path relative to repo root for token file). Ticketing URL is under `tickets`, not `git`. |
| `tickets` | no | Ticketing: `ticketing_url` (Gitea issues URL), `authorized_emails` (`to`: alias or list of env-keyed objects, `from`: list of allowed sender addresses). **to** resolves the project id when processing mails. It may be a single string or a list of one object with env keys: `[{ "test": "AI.<project_id>.TEST@4nkweb.com", "pprod": "...", "prod": "..." }]`. Pattern: `AI.<project_id>.<env>@4nkweb.com` (`project_id` and `env` may be uppercase). See projects/ia_dev/docs/TICKETS_SPOOL_FORMAT.md. |

The API token for ai_working_help is **not** in conf.json. It is found by scanning all projects and all envs (as described above): each file `projects/<id>/.secrets/<env>/ia_token` is read and the Bearer is compared to the file content or to (file content + env). Token value is **base** + **env**; **env** is the environment name (test, pprod, prod), to be adapted per environment. The first match gives project id and env.

## Example (minimal)

```json
{
  "name": "My App",
  "build_dirs": ["backend", "frontend"]
}
```

## Example (full)

See `projects/lecoffreio/conf.json`.