added notes

This commit is contained in:
2026-07-18 08:06:43 -07:00
parent a68d1b4060
commit b318dc28f4
3 changed files with 169 additions and 7 deletions
-7
View File
@@ -4,10 +4,3 @@ These are containers that have been requested to add to the repo. This used to
## The List
* Glance - dashboard
## Apps That Will Not be in the Repo
1. [Harbor](https://goharbor.io/) — Self-hosted container image registry. Harbor's compose file is generated by its own installer from harbor.yml (~10 tightly coupled services); a static compose file in this repo would drift immediately. Use the official installer.
2. [Exportify](https://github.com/watsonbox/exportify) — Spotify playlist CSV exporter. No published container image (Dockerfile only), and the hosted version at https://exportify.app/ runs entirely in the browser with no server-side data storage.
+52
View File
@@ -0,0 +1,52 @@
# Exportify
Exportify exports your Spotify playlists (tracks, artists, albums, and audio features) to CSV files. It has no published container image — only a Dockerfile you'd have to build yourself — and the hosted version runs entirely client-side in the browser with no server-side data storage, so there's nothing to self-host that isn't better served by the hosted app. See `container-requests.md` for the "will not be in the repo" rationale.
- Project Repository: https://github.com/watsonbox/exportify
- Hosted App: https://exportify.app/
## Tutorial: Exporting your Spotify playlists
### Option A — use the hosted app (recommended)
1. Go to https://exportify.app/
2. Click **Log in with Spotify** and authorize the app (it only requests read access to your playlists — it never modifies your Spotify account)
3. Your playlists load as a list; click **Export** next to a single playlist to download its CSV, or **Export All** to download every playlist as a zip of CSVs
4. Each CSV includes: Track URI, Track Name, Artist, Album, Release Date, Duration, Popularity, and Spotify audio-feature columns (Danceability, Energy, Key, Loudness, Tempo, etc.)
Since everything runs in your browser and talks directly to Spotify's API, no data passes through or is stored on any third-party server.
### Option B — run it yourself (no published image, build from source)
Only worth doing if you don't trust the hosted instance or want to self-host on an internal network.
```bash
git clone https://github.com/watsonbox/exportify.git
cd exportify
```
1. Create a Spotify app at https://developer.spotify.com/dashboard to get a Client ID
2. Add your redirect URI (e.g. `http://localhost:5000/callback` or your own domain) in the Spotify app settings
3. Copy the example env and set your client id/redirect uri:
```bash
cp .env.example .env
nano .env
```
4. Build and run with the project's own Dockerfile:
```bash
docker build -t exportify .
docker run -d --name exportify -p 5000:5000 --env-file .env exportify
```
5. Open `http://localhost:5000` and log in with Spotify as above
### Using the exported CSVs
- Import into a spreadsheet to dedupe playlists, sort by audio features, or build custom smart playlists
- Feed into other tools (e.g. `pandas.read_csv`) for playlist analysis
- Re-import isn't supported by Exportify itself — CSV export is one-way; use a separate tool (e.g. `spotify-backup`, `soundiiz`) if you need CSV-to-Spotify import
## Gotchas
- Spotify's API rate-limits large libraries; exporting many playlists back-to-back may need retries
- Audio-feature columns depend on Spotify's Audio Features API remaining available — Spotify has deprecated/restricted parts of this API for new apps in the past, so some columns may be blank depending on your app's API access tier
- The self-hosted route requires you to register and maintain your own Spotify Developer app (Client ID + redirect URI), which the hosted version already handles for you
+117
View File
@@ -0,0 +1,117 @@
# Harbor
Harbor is a self-hosted OCI container image registry with vulnerability scanning, image signing, replication, and RBAC — basically a private Docker Hub. Its compose file is generated by Harbor's own installer from `harbor.yml`, so a static `compose.yaml` in this repo would drift immediately (~10 tightly-coupled services: core, registry, database, redis, jobservice, portal, trivy, nginx, etc). Use the official installer instead of hand-rolling compose. See `container-requests.md` for the "will not be in the repo" rationale.
- Project Repository: https://github.com/goharbor/harbor
- Documentation: https://goharbor.io/docs/
- Releases (installer download): https://github.com/goharbor/harbor/releases
## Tutorial: Installing Harbor
### 1. Prerequisites
- Docker Engine 20.10.10+ and Docker Compose 2.0.0+ on the host
- A domain name (or IP) that will be Harbor's registry hostname
- A TLS certificate, or be prepared to run without HTTPS on a trusted internal network only (Docker refuses to talk to an insecure registry over plain HTTP unless explicitly configured)
### 2. Download the offline installer
```bash
cd /opt
wget https://github.com/goharbor/harbor/releases/download/v2.11.1/harbor-offline-installer-v2.11.1.tgz
tar xzvf harbor-offline-installer-v2.11.1.tgz
cd harbor
```
Check the [releases page](https://github.com/goharbor/harbor/releases) for the current version.
### 3. Configure `harbor.yml`
```bash
cp harbor.yml.tmpl harbor.yml
nano harbor.yml
```
Key fields to set:
```yaml
hostname: registry.example.com # the domain or IP clients will use
http:
port: 80
https:
port: 443
certificate: /your/certificate/path
private_key: /your/private/key/path
harbor_admin_password: ChangeThisPassword
data_volume: /data/harbor
database:
password: ChangeThisPassword
```
If you don't have a certificate yet, comment out the entire `https:` block and drop `https` from downstream instructions — but note Docker clients will then need `insecure-registries` configured to push/pull (see step 6).
### 4. Run the installer
```bash
sudo ./install.sh
```
This generates `docker-compose.yml` from `harbor.yml` and brings the stack up. Optional flags:
```bash
sudo ./install.sh --with-trivy # image vulnerability scanning
sudo ./install.sh --with-chartmuseum # Helm chart repository (deprecated upstream, skip for new installs)
```
### 5. Log in to the UI
Open `https://registry.example.com` and log in as `admin` with the password set in `harbor_admin_password`. Immediately:
1. Create a new project (e.g. `homelab`) — set it Private or Public
2. Create additional users under Administration → Users, or configure LDAP/OIDC
### 6. Push and pull images
```bash
docker login registry.example.com
docker tag myapp:latest registry.example.com/homelab/myapp:latest
docker push registry.example.com/homelab/myapp:latest
# on another machine:
docker pull registry.example.com/homelab/myapp:latest
```
If running without TLS, add the registry to `/etc/docker/daemon.json` on every client first:
```json
{
"insecure-registries": ["registry.example.com:80"]
}
```
then `sudo systemctl restart docker`.
### 7. Use it as a pull-through cache / replication target (optional)
Administration → Registries lets you add Docker Hub (or another registry) as an endpoint, then Administration → Replications to mirror images in on a schedule — useful for caching frequently-pulled base images locally.
### 8. Day-2 operations
```bash
cd /opt/harbor
sudo docker compose down
sudo ./install.sh # re-run after editing harbor.yml to regenerate and restart
```
Back up `data_volume` (default `/data/harbor`) and the Postgres data inside it before any upgrade. Upgrades are major-version-specific — always read the [upgrade guide](https://goharbor.io/docs/edge/administration/upgrade/) for your version jump, don't just re-run the new installer over old data blindly.
## Gotchas
- Trivy's vulnerability database downloads on first scan and needs outbound internet access; air-gapped installs need `--with-trivy` plus manually seeding the DB.
- The installer's generated `docker-compose.yml` is meant to be managed by `install.sh`, not edited by hand — re-running the installer after config changes is the supported workflow.
- Harbor's own `database` and `redis` containers store state locally by default; point `harbor.yml` at external Postgres/Redis instead if you already run those services elsewhere.