# remote-data-ssh-sync (`scripts/remote-data-ssh-sync.sh`)

Pulls **deployed environment data** over SSH into a **local mirror** (not versioned in Git), then optionally ingests that mirror into **AnythingLLM**.

## Configuration source (per project)

`projects/<id>/conf.json`:

- `smart_ide.remote_data_access.environments.<env>.ssh_host_alias`
- `smart_ide.remote_data_access.environments.<env>.remote_data_directories[]`
- `smart_ide.anythingllm_workspace_slug[env]` (optional; required for ingestion)

## Mirror location

Default:

- `<smart_ide_root>/.data/remote-data/<projectId>/<env>/<role>/`

This directory is ignored by Git (see `.gitignore`).

Override:

- `SMART_IDE_REMOTE_DATA_MIRROR_ROOT=/abs/path`

## AnythingLLM ingestion

By default, the script attempts ingestion and skips explicitly if config is missing.

Inputs:

- `~/.config/4nk/anythingllm-sync.env` (optional): provides `ANYTHINGLLM_BASE_URL` + `ANYTHINGLLM_API_KEY`
- `projects/<id>/conf.json`: provides the workspace slug for the selected env

Implementation:

- calls `scripts/anythingllm-pull-sync/sync.mjs` with `--upload-all` on each mirrored role directory

## Usage

```bash
./scripts/remote-data-ssh-sync.sh --project enso --env test
```

Fetch only (no ingestion):

```bash
./scripts/remote-data-ssh-sync.sh --project enso --env test --no-anythingllm
```

Ingest only specific roles:

```bash
./scripts/remote-data-ssh-sync.sh --project enso --env test --roles docv_dp_git_data
```

Dry-run (prints rsync command lines):

```bash
./scripts/remote-data-ssh-sync.sh --project enso --env test --dry-run
```