vibetree/README.md

# inventree

A small, self-contained IT asset inventory system. The heart of inventree is a
nested folder tree — like a file manager — for organizing physical and
virtual IT assets (servers, switches, laptops, printers, etc.).

- Single compiled Go binary, embedded HTML templates and static assets
- SQLite storage (WAL mode, foreign keys, migrations)
- Bootstrap 5 UI, minimal vanilla JavaScript (no Node/React/build step)
- Session-based auth with two roles: Admin and User
- Custom, admin-defined asset types with custom fields
- CSV import/export, full-text-ish search with breadcrumbs

## Quick start

```sh
go build -o inventree ./cmd/inventree
./inventree serve
```

Then open http://localhost:8080 and log in with:

```
Username: admin
Password: admin
```

(This default account is created intentionally on first run. Change or
disable it via the Users admin page once you're up and running.)

The database file is created automatically at `data/inventree.db` on first
run, with migrations applied and seed data (default statuses, item types,
and the admin account) inserted.

## CLI

```sh
inventree                    # same as `inventree serve`
inventree serve              # start the web server
inventree version            # print the version
inventree database migrate   # apply pending migrations and exit
inventree database backup [path]  # write a consistent VACUUM INTO snapshot
```

`serve` flags:

| Flag               | Default               | Description                                |
|---------------------|------------------------|--------------------------------------------|
| `-addr`             | `:8080`                | Listen address                             |
| `-db`               | `data/inventree.db`    | Path to the SQLite database file           |
| `-secure-cookies`   | `false`                 | Mark the session cookie `Secure` (enable when serving over HTTPS) |

`/healthz` is available for load balancer / container health checks.

## The inventory tree

The Inventory page is the default page after login and the center of the
application: a folder/asset tree on the left, and a detail panel on the
right that shows whatever is selected (a folder's contents, or a single
asset).

- **Folders** can nest arbitrarily deep, and contain both subfolders and
  assets.
- **Expand/collapse** state is remembered per user (persisted server-side),
  and folder contents are fetched lazily via small AJAX requests — the page
  never fully reloads when you browse the tree.
- **Drag and drop** folders and assets onto another folder (or onto
  "Inventory" itself) to move them. Admin only.
- **Right-click for everything**: it's meant to feel like a native file
  manager. Right-click a folder, an asset, a breadcrumb crumb, or empty
  tree/panel background for the actions available there (New Subfolder, Add
  Asset, Rename, Check Out, Move, Print Label, Delete, …) instead of hunting
  for a toolbar button. The main panel itself stays blank until you actually
  click into a folder or asset — browsing happens in the tree.
- **Deleting a folder** that still contains subfolders or assets requires
  confirmation, and offers a safe "move contents elsewhere" option instead
  of an accidental cascade. A "permanently delete everything inside" option
  is also available for when that's genuinely what you want.
- **Search** (top of the tree pane) looks across asset ID, name, folder
  name, item type, assigned user, notes, and custom field values, and shows
  each result's full folder breadcrumb.

## Item types and custom fields

Admins can define arbitrary asset types (Laptop, Switch, Firewall, …) under
**Item Types**, each with its own set of custom fields (text, long text,
number, date, boolean, dropdown, URL, IP address), marked required or not,
with a default value and display order. A representative starter set of
types and fields is seeded on first run — feel free to edit or delete it.

## Checkout, quantity, and contacts

- **Check out / check in**: right-click an asset → **Check Out…** to assign
  it to a user (or clear the assignment to check it back in). Works from any
  asset row, not just its detail view.
- **Quantity vs. unique**: item types have a **Unique** flag (checked by
  default). Unique types — a laptop, a server — get one asset row per
  physical unit, same as before. Uncheck it for bulk/consumable types (a USB
  cable, a box of screws) and assets of that type are tracked as a single
  row with a **Quantity** and +/- steppers in the asset panel and folder
  tables, instead of one row per unit. Flipping a type back to Unique clamps
  any existing quantities back down to 1.
- **Contacts**: under Settings → Users, a third role alongside Admin and
  User. Contacts can't sign in at all — they exist purely so you can assign
  assets to end users (e.g. "this laptop belongs to Jane in Accounting")
  without creating them a login account. A blank password is fine when
  creating one.

## QR asset tags and printable labels

Every asset can have a QR-coded label printed for it: right-click an asset
for **Print Label**, or right-click a folder for **Print Labels…** (every
asset in that folder, recursively). The QR code encodes an absolute
`/scan/<Asset ID>` URL, so scanning it with a phone camera — or a USB
barcode/QR scanner focused on the browser's address bar — opens straight to
that asset.

The print page supports two label sheet layouts, switchable from a dropdown
on the page itself:

- **Avery 5160** (1" x 2-5/8", 30/sheet) — standard address labels, exact
  published dimensions.
- **61526** asset tag labels (3/4" x 2", ~48/sheet) — the grid/margins here
  are a best-effort layout, not a verified spec sheet, so print one test
  page on plain paper and check it against a real label before committing a
  full (often non-refundable) sheet.

## CSV import / export

- **Export**: from the tree's right-click menu, or `/export/assets.csv`
  (optionally scoped to a folder, with `?recursive=1` to include
  subfolders). Includes a `Quantity` column and one `cf_<field_key>` column
  per distinct custom field in use.
- **Folder-scoped import**: right-click a folder → **Import CSV…** imports
  every row directly into that folder. No `Folder` column needed — that's
  the whole point of doing it from a specific folder. Never available at the
  top level.
- **Global import**: Settings → Import / Export has the full-featured
  version — a CSV with a `Folder` column (a `/`-separated path, e.g.
  `Main Office/Network`) that can optionally auto-create missing folders —
  plus a separate CSV import for bulk-adding/updating Statuses (`Name`,
  `Color`). Both matching-by-name/Asset-ID semantics and full column
  references with examples are documented right on that page.

A full backup/restore of everything (folders, assets, types, custom fields,
statuses, users) is just `inventree database backup` — it's a consistent
snapshot of the whole SQLite database.

## Project layout

```
cmd/inventree/        CLI entrypoint (serve, version, database migrate/backup)
internal/
  auth/                session management, secure cookies
  database/            SQLite connection, pragmas, migrations
  migrations/           embedded .sql migration files
  models/               shared domain types
  folders/               nested folder tree
  assets/                 asset CRUD, search, CSV import/export
  itemtypes/              admin-defined types + custom field definitions
  users/                  accounts
  statuses/               configurable asset statuses
  settings/               key/value app settings
  web/                    HTTP router, handlers, template rendering
web/
  templates/              Go html/template pages and partials (embedded)
  static/                 Bootstrap 5, Bootstrap Icons, app.css, app.js (embedded)
deploy/                  systemd unit example
```

## Development

```sh
go build ./...
go vet ./...
go test ./...
```

The SQLite driver is [modernc.org/sqlite](https://pkg.go.dev/modernc.org/sqlite),
a pure-Go implementation, so `CGO_ENABLED=0` builds work out of the box and
the compiled binary has no C library dependencies.

## Docker

```sh
docker compose up --build
```

This builds the image, starts inventree on port 8080, and persists the
database in a named volume (`inventree-data`, mounted at `/app/data`).

Or directly:

```sh
docker build -t inventree .
docker run -p 8080:8080 -v inventree-data:/app/data inventree
```

## systemd

See `deploy/inventree.service` for an example unit. Adjust the paths, then:

```sh
sudo useradd --system --home /opt/inventree --shell /usr/sbin/nologin inventree
sudo mkdir -p /opt/inventree/data
sudo cp inventree /opt/inventree/inventree
sudo chown -R inventree:inventree /opt/inventree
sudo cp deploy/inventree.service /etc/systemd/system/
sudo systemctl daemon-reload
sudo systemctl enable --now inventree
```

## Scope

inventree is intentionally a focused MVP: folders, assets, custom types,
statuses, users, search, and CSV import/export. It does not include
ticketing, purchasing, monitoring, RMM, MDM, or accounting features — if you
need those, look elsewhere and integrate via CSV.