lhumina_code/hero_drive
22
0
Fork 0
Code Issues 1 Pull requests 1 Projects Releases Packages Wiki Activity Actions

Complete remaining file manager features from spec #6

New issue
Open
opened 2026-02-25 14:45:32 +00:00 by mik-tf · 0 comments
mik-tf commented 2026-02-25 14:45:32 +00:00
Owner
Copy link

Summary

Remaining features from the full file manager spec (geomind_code/home#41) that are not yet implemented in hero_drive.

The core file manager (v1) is complete — 10 RPC methods, Dioxus UI with browse/search/viewers/markdown editor, 5-crate workspace, CI green. This issue tracks the remaining items to reach full spec compliance.

Architecture: Dual-Mode Operation

hero_drive supports two deployment modes:

Standalone Mode (default)

  • Embeds hero_fossil as a library (features = ["server"])
  • Spawns its own WebDAV server on port 3387
  • Self-contained: cargo run just works, no external dependencies
  • Use case: development, testing, single-machine deployments

Ecosystem Mode (via hero_zero)

  • Connects to an external hero_fossil service as a WebDAV client
  • WEBDAV_URL env var points to hero_fossil (e.g., via hero_proxy at http://localhost:8080/hero_fossil/webdav/default)
  • Listens on Unix socket at ~/hero/var/sockets/hero_drive.sock
  • hero_proxy routes /hero_drive/... to the socket
  • hero_auth provides JWT-based auth via shared HERO_SECRET
  • heroindex provides full-text search + audit log
Standalone:                          Ecosystem (hero_zero):

hero_drive                           hero_fossil (separate service)
├── embedded hero_fossil (:3387)       └── WebDAV on hero_fossil.sock
├── OServer RPC (:3394)                        ↑
└── Dioxus UI                        hero_drive (WebDAV client)
                                     ├── OServer RPC
                                     ├── Dioxus UI
                                     └── hero_drive.sock
                                               ↑
                                     hero_proxy (:8080) → browser

The WEBDAV_URL env var controls which mode: if set, hero_drive uses the external URL and skips spawning the embedded WebDAV server. If unset, it starts the embedded server on the default port.

Remaining Features

1. Documentation (README.md)

  • Architecture overview (dual-mode: standalone vs ecosystem)
  • Quick start for standalone mode (cargo run)
  • Ecosystem integration guide (hero_zero, hero_proxy, WEBDAV_URL)
  • Environment variables reference (WEBDAV_URL, HERO_SECRET, RUST_LOG)
  • API reference (10 RPC methods, WebDAV endpoints)

2. Ecosystem Integration

  • CLI flag or auto-detect to skip embedded WebDAV when WEBDAV_URL is set
  • --bind unix:<path> argument for Unix socket binding (alongside TCP fallback)
  • GET /health endpoint for hero_proxy health checks
  • hero_zero service template (services/hero_drive.toml)
  • hero_proxy routing entry (hero_drive:http)

3. Auth & Permissions (integrate with hero_auth)

  • Axum middleware: extract Authorization: Bearer <JWT> header
  • Local JWT validation using shared HERO_SECRET env var (HS256, zero network calls)
  • Scope-based access control: read (browse/download), write (upload/edit/move/copy), admin (delete permanently, manage perms)
  • Login redirect flow via hero_auth (/login?redirect_uri=...)
  • UI: login page, token storage in state, inject Bearer header in API calls
  • Hide UI actions the user's scope doesn't permit
  • Standalone mode: auth optional (disabled by default, enabled with --require-auth)

4. Advanced Search (via heroindex)

  • Add heroindex_client dependency
  • Index file metadata into heroindex (name, path, mime_type, size, modified_at)
  • Search by date range (modified_at gte/lte) — currently missing from UI
  • Full-text content search for text/markdown files (spec: index-based, via heroindex)
  • Replace recursive WebDAV walk with heroindex queries for better performance
  • Standalone mode: fall back to current recursive WebDAV search when heroindex unavailable

5. Server-side Audit Log (via heroindex)

  • Index file operation events (create, delete, rename, move, copy, upload, download) into heroindex
  • Each event: timestamp, user (from JWT sub), action, path, detail
  • RPC method to query audit log with date range and filters
  • UI: server-persisted activity panel (replace current client-side-only in-memory log)
  • Standalone mode: fall back to in-memory ring buffer (current behavior)

6. Zip Download

  • Server-side zip assembly endpoint for folder download
  • Multi-file zip download (selected files)
  • UI: download button for folders and multi-select

7. Folder Upload

  • Add webkitdirectory attribute to file input for folder upload
  • Reconstruct folder paths from dropped directory entries
  • Create subdirectories on WebDAV before uploading files

8. Viewer & Editor Enhancements

  • Image viewer: fullscreen/lightbox mode with requestFullscreen API
  • PDF viewer: custom page navigation controls (page N of M), zoom (replace bare iframe)
  • Markdown editor: drag image into textarea to upload and insert link
  • Markdown editor: version conflict detection (ETag/If-Match before save)

9. UI Polish

  • Loading skeleton placeholders (animated cards instead of text spinner)
  • New file from template (pre-populated content per file type)

Ecosystem Dependencies

Dependency Purpose Standalone Ecosystem
hero_fossil WebDAV file storage Embedded (library) External service via hero_proxy
hero_auth OAuth2 + JWT auth Optional (--require-auth) Required via HERO_SECRET
heroindex Full-text search + audit log Fallback to WebDAV walk + memory log Full integration
hero_proxy Unix socket routing Not used (TCP direct) Routes /hero_drive/...
hero_zero Service orchestration Not used Manages lifecycle

Priority Order

  1. README — document what exists and the dual-mode architecture
  2. Ecosystem integration — socket binding, health endpoint, hero_zero template
  3. Auth (hero_auth) — security foundation
  4. Advanced search (heroindex) — replaces fragile recursive walk
  5. Audit log (heroindex) — reuses same heroindex integration
  6. Zip download — commonly requested
  7. Folder upload — browser support varies
  8. Viewer/editor enhancements — polish
  9. UI polish — cosmetic
## Summary Remaining features from the full file manager spec ([geomind_code/home#41](https://forge.ourworld.tf/geomind_code/home/issues/41)) that are not yet implemented in hero_drive. The core file manager (v1) is complete — 10 RPC methods, Dioxus UI with browse/search/viewers/markdown editor, 5-crate workspace, CI green. This issue tracks the remaining items to reach full spec compliance. ## Architecture: Dual-Mode Operation hero_drive supports two deployment modes: ### Standalone Mode (default) - Embeds hero_fossil as a library (`features = ["server"]`) - Spawns its own WebDAV server on port 3387 - Self-contained: `cargo run` just works, no external dependencies - Use case: development, testing, single-machine deployments ### Ecosystem Mode (via hero_zero) - Connects to an **external** hero_fossil service as a WebDAV client - `WEBDAV_URL` env var points to hero_fossil (e.g., via hero_proxy at `http://localhost:8080/hero_fossil/webdav/default`) - Listens on Unix socket at `~/hero/var/sockets/hero_drive.sock` - hero_proxy routes `/hero_drive/...` to the socket - hero_auth provides JWT-based auth via shared `HERO_SECRET` - heroindex provides full-text search + audit log ``` Standalone: Ecosystem (hero_zero): hero_drive hero_fossil (separate service) ├── embedded hero_fossil (:3387) └── WebDAV on hero_fossil.sock ├── OServer RPC (:3394) ↑ └── Dioxus UI hero_drive (WebDAV client) ├── OServer RPC ├── Dioxus UI └── hero_drive.sock ↑ hero_proxy (:8080) → browser ``` The `WEBDAV_URL` env var controls which mode: if set, hero_drive uses the external URL and skips spawning the embedded WebDAV server. If unset, it starts the embedded server on the default port. ## Remaining Features ### 1. Documentation (README.md) - [ ] Architecture overview (dual-mode: standalone vs ecosystem) - [ ] Quick start for standalone mode (`cargo run`) - [ ] Ecosystem integration guide (hero_zero, hero_proxy, WEBDAV_URL) - [ ] Environment variables reference (WEBDAV_URL, HERO_SECRET, RUST_LOG) - [ ] API reference (10 RPC methods, WebDAV endpoints) ### 2. Ecosystem Integration - [ ] CLI flag or auto-detect to skip embedded WebDAV when `WEBDAV_URL` is set - [ ] `--bind unix:<path>` argument for Unix socket binding (alongside TCP fallback) - [ ] `GET /health` endpoint for hero_proxy health checks - [ ] hero_zero service template (`services/hero_drive.toml`) - [ ] hero_proxy routing entry (`hero_drive:http`) ### 3. Auth & Permissions (integrate with hero_auth) - [ ] Axum middleware: extract `Authorization: Bearer <JWT>` header - [ ] Local JWT validation using shared `HERO_SECRET` env var (HS256, zero network calls) - [ ] Scope-based access control: `read` (browse/download), `write` (upload/edit/move/copy), `admin` (delete permanently, manage perms) - [ ] Login redirect flow via hero_auth (`/login?redirect_uri=...`) - [ ] UI: login page, token storage in state, inject Bearer header in API calls - [ ] Hide UI actions the user's scope doesn't permit - [ ] Standalone mode: auth optional (disabled by default, enabled with `--require-auth`) ### 4. Advanced Search (via heroindex) - [ ] Add `heroindex_client` dependency - [ ] Index file metadata into heroindex (name, path, mime_type, size, modified_at) - [ ] Search by date range (modified_at gte/lte) — currently missing from UI - [ ] Full-text content search for text/markdown files (spec: index-based, via heroindex) - [ ] Replace recursive WebDAV walk with heroindex queries for better performance - [ ] Standalone mode: fall back to current recursive WebDAV search when heroindex unavailable ### 5. Server-side Audit Log (via heroindex) - [ ] Index file operation events (create, delete, rename, move, copy, upload, download) into heroindex - [ ] Each event: timestamp, user (from JWT sub), action, path, detail - [ ] RPC method to query audit log with date range and filters - [ ] UI: server-persisted activity panel (replace current client-side-only in-memory log) - [ ] Standalone mode: fall back to in-memory ring buffer (current behavior) ### 6. Zip Download - [ ] Server-side zip assembly endpoint for folder download - [ ] Multi-file zip download (selected files) - [ ] UI: download button for folders and multi-select ### 7. Folder Upload - [ ] Add `webkitdirectory` attribute to file input for folder upload - [ ] Reconstruct folder paths from dropped directory entries - [ ] Create subdirectories on WebDAV before uploading files ### 8. Viewer & Editor Enhancements - [ ] Image viewer: fullscreen/lightbox mode with `requestFullscreen` API - [ ] PDF viewer: custom page navigation controls (page N of M), zoom (replace bare iframe) - [ ] Markdown editor: drag image into textarea to upload and insert link - [ ] Markdown editor: version conflict detection (ETag/If-Match before save) ### 9. UI Polish - [ ] Loading skeleton placeholders (animated cards instead of text spinner) - [ ] New file from template (pre-populated content per file type) ## Ecosystem Dependencies | Dependency | Purpose | Standalone | Ecosystem | |---|---|---|---| | hero_fossil | WebDAV file storage | Embedded (library) | External service via hero_proxy | | hero_auth | OAuth2 + JWT auth | Optional (`--require-auth`) | Required via `HERO_SECRET` | | heroindex | Full-text search + audit log | Fallback to WebDAV walk + memory log | Full integration | | hero_proxy | Unix socket routing | Not used (TCP direct) | Routes `/hero_drive/...` | | hero_zero | Service orchestration | Not used | Manages lifecycle | ## Priority Order 1. **README** — document what exists and the dual-mode architecture 2. **Ecosystem integration** — socket binding, health endpoint, hero_zero template 3. **Auth** (hero_auth) — security foundation 4. **Advanced search** (heroindex) — replaces fragile recursive walk 5. **Audit log** (heroindex) — reuses same heroindex integration 6. **Zip download** — commonly requested 7. **Folder upload** — browser support varies 8. **Viewer/editor enhancements** — polish 9. **UI polish** — cosmetic
mik-tf changed title from WIP: Complete remaining file manager features from spec to Complete remaining file manager features from spec 2026-02-25 14:46:40 +00:00
Sign in to join this conversation.
No Branch/Tag specified
Branches Tags
development
main
No results found.
Labels
Clear labels
No items
No labels
Milestone
Clear milestone
No items
No milestone
Projects
Clear projects
No items
No project
Assignees
Clear assignees
No assignees
1 participant
Notifications
Due date
The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set.

Reference
lhumina_code/hero_drive#6
No description provided.