Skip to content

Roadmap

Below is the development roadmap for MicroJSON, showing completed phases and current/future work. The project has evolved from a lightweight annotation format into a high-performance tiling platform with Rust acceleration.

Phase 1: Consolidation and Documentation (Complete)

  1. Refinement of Core Model:

    • Finalized and stabilized the MicroJSON core model with Pydantic v2.
    • Published updated documentation and specifications.
    • Implemented hierarchical references (parentId, ref).

  2. Community Engagement:

    • Established communication channels via GitHub issues.
    • Gathered feedback from stakeholders and early adopters.

Phase 2: Expanded Features and Extensions (Complete)

  1. Harmonization with GeoJSON Pydantic:

    • Full compatibility with geojson-pydantic — MicroJSON models extend GeoJSON Pydantic types.

  2. Harmonization with OME Model:

    • Integrated OME-compatible coordinate systems and multiscale metadata.
    • Provenance tracking module for data lineage.

  3. Tiling with TileJSON and binary formats:

    • TileJSON 3.0.0 metadata specification for MicroJSON tilesets.
    • 2D vector tile pipeline: GeoJSON to PBF (MVT) and tiled Parquet.
    • Python TileWriter/TileReader for legacy PBF workflows.

Phase 3: 3D Data and Rust Acceleration (Complete)

  1. Rust-Accelerated Tiling Engine:

    • Full rewrite of hot-path tiling code in Rust (PyO3 + maturin).
    • 2D pipeline: StreamingTileGenerator2D with quadtree-based clipping, Douglas-Peucker simplification, and parallel tile encoding via rayon.
    • 3D pipeline: StreamingTileGenerator with octree indexing, QEM mesh simplification, and multi-format output.

  2. 3D Geometry and Mesh Support:

    • Added PolyhedralSurface and TIN geometry types.
    • OBJ mesh ingestion with parallel parsing.
    • Fragment file format (MJF2) for sharded intermediate storage.

  3. Output Formats:

    • 3D Tiles (GLB) with meshopt or Draco compression.
    • PBF3 — custom protobuf format for 3D tile data.
    • Tiled Parquet (ZSTD) for ML training pipelines.
    • Neuroglancer precomputed format for web-based 3D visualization.
    • 2D PBF (MVT) — pure-Rust encoder/decoder, no Python dependencies.

  4. Compression:

    • Meshopt (lossless, fast decode, Brotli-friendly) — default for viewer output.
    • Draco (lossy quantization, smallest on disk) — optional.
    • ZSTD for Parquet columns.
    • Brotli HTTP transport compression for GLB serving.

  5. Dataset Pipelines:

    • MouseLight: 38 brains, 876K rows, meshopt 3D Tiles (84 min total).
    • Hemibrain: 5,000 neurons (95 cell types), Parquet tiling complete.

Phase 4: Adoption and Long-Term Sustainability (In Progress)

  1. Case Studies and Benchmarks:

    • Published benchmark tables: synthetic, MouseLight, retrieval, multi-pyramid.
    • Reproducing paper results with current pipeline (Tables 2-8).

  2. Documentation and Guides:

    • Updated specification for 3D geometry types.
    • API documentation for Rust-accelerated modules.
    • Example scripts for 2D and 3D pipelines.

  3. Remaining Work:

    • TypeScript and Java reference implementations.
    • Governance model and standards process.
    • Hemibrain full 3D Tiles with QEM + meshopt.
    • Community engagement: user meetings and feedback sessions.