Pulse Parse

Pulse Parse is a SQLite-backed Apple Health export parsing and workout analysis toolkit with a Tauri desktop app and shared Rust crates.

What it does

• preprocesses Apple Health export.xml files so malformed DTD content and invalid control characters do not break parsing
• ingests workouts and records into a local SQLite database
• links workouts to overlapping Apple Health records so exports include corresponding workout data
• inspects workout coverage, summarizes activity over a time window, and exports workout bundles as JSON or CSV

Repository layout

This is a Rust workspace with a Tauri desktop app nested under desktop/, not a single-package Tauri repo:

• desktop/ contains the Vite/React frontend and the Tauri app package
• desktop/src-tauri/ contains the Tauri host, commands, and bundling config
• crates/health-core contains shared request/event contracts and dataset metadata types
• crates/health-store contains the SQLite ingest, query, and export layer
• crates/health-service contains the desktop-facing service layer used by Tauri

That means the desktop app itself follows the usual Tauri split (src/ plus src-tauri/), while the backend logic is factored into shared Rust crates.

Desktop app

The repository includes a Tauri v2 desktop app in desktop/ with an in-process Rust backend built from the shared workspace crates.

The Rust desktop backend is organized in crates/:

• health-core owns the desktop request/event contracts and shared dataset metadata types
• health-store owns app-managed dataset path conventions and dataset state persistence
• health-service owns the desktop-facing ingest, dashboard, workout detail, and export use cases used by Tauri

Prerequisites for local desktop development and packaging:

cd desktop
pnpm install

If you use just, the repository root includes shortcuts for the common flows:

just install
just check
just test
just lint
just fmt
just dev
just bundle

Run just --list to see the full set of available recipes.

Run the desktop app in development mode:

pnpm run tauri:dev

Build the desktop bundle for your platform:

pnpm run tauri:build

If you build on Linux, install the system packages required by Tauri/WebKitGTK first. If you build on Windows, make sure WebView2 is available on the target machine.

The desktop app supports:

• loading an Apple Health export.xml directly into one app-managed SQLite dataset
• rebuilding the desktop dataset with inline progress reporting
• reopening the desktop dataset from local app state without reselecting a DB path
• filtering by a single day or date range plus workout type, including broader suggested types and custom activity-type entries
• viewing overview metrics, activity breakdowns, trends, a searchable workout overview list, selected workout detail, and workout metric drilldown modals
• exporting full JSON, condensed summary JSON, or CSV directories

The desktop app is self-contained on the backend side: it uses the bundled Rust service layer directly.

Platform support

The desktop app is structured to be cross-platform:

• app-managed datasets and config use Tauri's platform-specific app data and config directories
• SQLite is bundled through rusqlite, so the app does not depend on a system SQLite install
• the bundle config includes Windows icons (.ico) plus PNG assets used by Linux and macOS

The source layout targets macOS, Linux, and Windows, but the repository does not include a CI matrix that exercises those platforms automatically.