fasttext_flutter 0.1.1

On-device text classification and sentence embeddings using fastText .ftz quantized models via Dart FFI. Verified on Android. Contributions welcome

0.1.1

• Fix (Android): Added missing aligned.h header from fastText's bundled intgemm library — resolves fatal error: 'aligned.h' file not found when compiling with Android NDK clang.
• Fix (Android): Added -D_POSIX_C_SOURCE=200809L build flag to guarantee posix_memalign availability on all Android API levels ≥ 24.

0.1.0

• Initial release.
• Load .ftz (quantized) and .bin (full) fastText models from the file system or Flutter assets.
• predict(text, {k, threshold}) — top-k text classification for supervised models, returns List<FastTextPrediction>.
• computeEmbedding(text) — dense sentence-vector embeddings as Float32List, works for all model types.
• FastTextModel.cosineSimilarity(a, b) — static cosine-similarity helper.
• FastTextModelType enum — detect whether the model is supervised, cbow, or skipgram at runtime.
• All inference runs in a background Isolate — the Flutter UI thread is never blocked.
• NativeFinalizer-based automatic memory management with explicit close() for deterministic cleanup.
• Platform support: Android (arm64, arm, x86_64), iOS, macOS, Linux, Windows.
• fastText C++ library compiled from source via native_toolchain_c build hooks — no pre-built binaries.