qrate v0.1.0-alpha.1

Overview

qrate reaches its first alpha release with a stable core data container, full project-creation workflow, and essential table operations. This release is ready for initial archivist testing and feedback before moving toward feature completeness.

What's New in v0.1.0-alpha.1

File Persistence & Editing (PR #47)

• Autosave with configurable modes — choose between timed/buffered (default), immediate, or off.
• Ctrl+S explicit save — save on demand, with an automatic flush at quit time.
• Unsaved indicator — the project title in the window shows a dot when unsaved changes are pending.
• Floating cell editor — replaces inline editing with a floating popover that overflows the cell and stays on-screen, with sticky highlighted headers (row number and column header highlight while editing).
• Data integrity — all edits are persisted atomically to the .qrate file in the original column order, preserving layout.

Details Panel Enhancements (PR #48)

• Tabular field display — replaced text lists with a bordered 2-column grid (label | value) for cleaner metadata review.
• Sticky image + scrolling fields — the image stays fixed while you scroll through field metadata, with resizable height (80–600px, 180px default).
• Type-aware file icons — resolved files now show appropriate icons (BookOpen for PDFs/docs, Play for audio, Frame for video, File for others).
• Open/reveal buttons — every resolved file type now has affordances to open or reveal in folder.
• Fixed layout bugs — addressed fieldlist clipping, rounded-border image rendering, and the bottom-dock crop affecting Details visibility.

Table Search & Filtering (PR #46)

• Global find bar — Ctrl+F search with case-insensitive, whole-word, and regex toggles (matching a Zed-style workflow).
• Per-column filtering — Google Sheets–style column header dropdowns with value checklist and search-to-narrow for high-cardinality columns.
• View/source index split — filters correctly narrow visible rows without reordering the underlying data, ensuring edits write back to the correct source row.
• Composable with virtualization — filtering and search work seamlessly with the virtualized grid on large datasets.

Project File Stability (Preceding work, ASNT-16)

• WAL mode replaced with DELETE journal — eliminated the .qrate-wal and .qrate-shm leftover-file problem by switching to DELETE mode with synchronous=FULL and busy_timeout=5000.
• Single-connection architecture — projects use a single owned connection per process (no pooling), simplifying concurrency and crash recovery.
• Layout persistence — dock position, size, and layout state are persisted with the project.

Downloads

• Windows: *-setup.exe (installer) or *-x86_64.zip (portable)
• macOS: *-universal.dmg (Intel + Apple Silicon)

Verify your download against SHA256SUMS.txt.

[!NOTE] These builds are unsigned.

• Windows: SmartScreen may warn "unknown publisher" — click More info → Run anyway.
• macOS: right-click the app → Open the first time, or run xattr -dr com.apple.quarantine "/Applications/qrate.app".

COMP 371 Final project submission (dev version 1)

This is the initial developer release (v0.0.1-dev) of qrate, a local-first desktop workspace designed to help archivists catalog cultural heritage materials without the performance bottlenecks of traditional tools.

The goal was to replace the fragmented workflow of switching between Excel and file explorers with a single, high-performance environment that doesn't crash on large datasets.

What’s working in this build:

• Performance: The "Headless Document Model" (SQLite + Rust) is fully functional. You can import 1GB+ CSV files, load them instantly, and scroll through millions of rows with minimal RAM usage.
• Unified Interface: The split-pane view is active, letting you review media files and edit metadata side-by-side without context switching.
• Data Integrity: Auto-save is implemented with ACID transactions, meaning edits are persisted immediately to disk to prevent data loss.
• Archival Tools: The background thumbnail pipeline, custom spellchecker, and cell annotation system are all hooked up and running.

Foundational work (Scaffolding):

• AI Infrastructure: The Rust backend traits and provider modules (src-tauri/src/ai/) are in place, currently set up with scaffolding for Cohere and Mock providers.
• Chat Interface: The frontend routes and UI components for the AI assistant (src/routes/chat/) have been created, laying the groundwork for the future "metadata suggestion" feature.

Tech Stack: Built using Tauri v2, Svelte 5, and Rust.

See the README.md for the full architecture breakdown and setup instructions. You can test the app using sample data attached with this release.