haptify 0.4.0+1

Generate haptic feedback files from audio: .ahap for iOS Core Haptics plus Android waveforms, via a CLI or a composable Dart pattern API.

Changelog

0.4.0+1

• Add a project banner to the README (renders on GitHub and pub.dev). Docs-only rebuild of 0.4.0.

0.4.0

First stable release. From here, stable versions are released from main and -dev.N prereleases track the dev branch on pub.dev.

New capabilities:

• Android primitive compositions: pattern.toPrimitives() and the --formats primitives CLI output map patterns onto VibrationEffect.Composition primitives (click/tick/thud/rises/falls/ spin) with scales, delays, and a minApiLevel.
• AHAP parsing: HapticPattern.fromAhap tolerantly parses .ahap files, and the CLI accepts them as inputs — convert existing iOS haptic libraries to Android formats without re-analyzing audio.
• The analyzer emits time-varying sharpness curves (iOS): each continuous segment's brightness (zero-crossing rate) is tracked over time and encoded as additive HapticSharpnessControl deviations around the event's sharpness. Curves are only emitted when the brightness actually moves; disable with --no-sharpness-curves or AnalysisOptions(sharpnessCurves: false).

Accuracy fixes for long and complex sounds:

• Fixed AHAP curve timing on iOS: parameter-curve control point times are now encoded relative to the curve's start, per the Core Haptics spec. Curves that started after t=0 — every segment after the first in a long sound — previously played with doubled time offsets.
• Intensity-curve point budgets now scale with segment length (--curve-rate, default 16 points/second), so long uploads keep their envelope detail instead of being crushed into 32 points; the RDP simplifier binary-searches its tolerance to use the full budget.
• Android waveform amplitudes are sampled at step midpoints, removing the half-step lag that dulled ramps.

Polish:

• Android conversions report one curveParameterUnsupported warning per parameter type (with a count) instead of one per curve.
• Docs: a "Tuning the output" guide mapping symptoms to CLI flags and AnalysisOptions, a haptics/sound learning-resources section, and a rewritten example walking through audio conversion, loading .ahap files, DSL authoring, and analysis tuning.
• Roadmap: playback and the Flutter companion package were removed; the demo app is the reference integration.

0.2.0-dev.4

• CLI: inputs may now be folders, and running haptify convert with no input scans the current directory for audio files.
• CLI: new default output layout — outputs group by type in a haptify-output/ folder placed next to the source folder (ahap/, waveform/), and Dart sources go to lib/generated/ so they are importable straight away. An explicit --out directory still receives everything flat.
• CLI: generated Dart file names are sanitized to lower_case_with_underscores (piano-loop.wav → piano_loop_haptic.dart), keeping the file_names lint happy.

0.2.0-dev.3

• Convert audio loaded at runtime: decodeAudioBytes(bytes) and AudioAnalyzer.analyzeBytes(bytes) decode WAV/MP3 straight from a Uint8List — no file path, filesystem, or ffmpeg — for user uploads, recordings, and downloads. Format is auto-detected from the bytes.
• Internal: the file decoder now reuses the shared byte decoder.

0.2.0-dev.2

• Restore the roadmap section in the README (primitive compositions, AHAP parsing, presets, Flutter companion).
• Add a tag-triggered GitHub Actions workflow that publishes to pub.dev.

0.2.0-dev.1

Dev preview — APIs and output details may still change before a stable release. Feedback welcome on the issue tracker.

• Built-in MP3 decoding in pure Dart (vendored minimp3 port) — ffmpeg is no longer required for MP3 input. Handles ID3v2 tags and trims LAME encoder delay/padding so haptic timing matches the original audio.
• Expanded installation instructions (global activation, dev dependency, per-platform ffmpeg setup for other formats).

0.1.0

• Initial release.
• haptify convert: generate haptic files from WAV/MP3 audio — .ahap for iOS Core Haptics, waveform JSON for Android VibrationEffect, and Dart constant sources that compile patterns into the app.
• Audio analysis: RMS loudness envelope, energy-flux onset detection, zero-crossing-rate sharpness estimation, silence gating, and Ramer-Douglas-Peucker curve simplification, all tunable via CLI flags.
• Library API: immutable haptic pattern model, composable DSL (then, overlay, repeat, scaleIntensity, ... plus 250.ms duration sugar), lossless AHAP encoder, and Android waveform encoder with conversion warnings.