v1.0.3

v1.0.3 (2026-04-19)

Bug Fixes

• IoU-based instance matching (#93, 4da7757)

Chores

• Fix build of Docker images in CI pipeline (#91, bbba18a)

---

Detailed Changes: v1.0.2...v1.0.3

v1.0.2

v1.0.2 (2026-04-09)

Bug Fixes

• Memory efficient majority voting (#92, 39ee416)

Continuous Integration

• Move Docker build to self-hosted runner (#89, 2cf7519)

---

Detailed Changes: v1.0.1...v1.0.2

v1.0.1

v1.0.1 (2026-03-31)

Bug Fixes

• Trigger new release with updated unzip dependencies (#90, 0da044d)

Refactoring

• Replace zipfile-deflate64 by stream-unzip (#87, d286acc)

• Unzip (#88, 3f745b5)

---

Detailed Changes: v1.0.0...v1.0.1

v1.0.0

v1.0.0 (2026-03-29)

Bug Fixes

• Dockerfiles (#86, 23a9563)

• Float precision errors in unravel_flat_index (#80, 344b26d)

• Handling of empty inputs in resolve_instance_overlaps (#58, f5153ca)

Chores

• Drop support for Python 3.9 and 3.10 and upgrade to numpy >= 2.3 (#85, 9ddfee5)

• Pin numpy version to < 2.3 to avoid conflicts with cuml / numba (#79, e643e1d)

• Update docker image (#78, f4222b3)

Continuous Integration

• Add Docker image without CUDA dependencies (#82, 523f158)

• Refactor test setup (#76, ecb12c4)

• Update CI pipeline (#83, b6d0f89)

Features

• Add method to check whether flash attention is installed (#56, 697b5e3)

• Instance matching (#72, 246330e)

• Majority voting (#69, 5de343c)

• Operation to resolve instance overlaps (#57, 8b4512b)

• Pcd reader and writer (#51, 479a461)

• Ply reader and writer (#55, d9e1de1)

• Support Python 3.14 and newer PyTorch versions (#75, d5fe2a3)

Refactoring

• Adjust unzipping progress bar total value (#68, 226a35d)

• Change data type of pcd files (#53, 0cf30d6)

• CI pipeline (#81, 63ca3fe)

• CI pipeline (#60, c45db9d)

• Data PCD reader and writer (#54, 1aadb2b)

• Dependency installation in CI pipeline (#71, 8f52f7d)

• Download_file (#65, 88e5b88)

• Instance matching metrics (#74, 2fd722f)

• Introduce type aliases (#59, b0d8a71)

• Make_labels_consecutive (#61, 061d7f1)

• Migrate PCD reader and writer to pypcd4 (#63, 371d481)

• Resolve_instance_overlaps (#77, cadbfbd)

• Support uppercase file endings in file readers and writers (#73, c8d6e07)

• Total size in progress bar for file unzipping (#70, 9fbb523)

• Type hints in PCD writer (#66, 5480ac8)

• Unzip (#67, 8f5e22e)

• Update type hints (#52, 91c3a30)

---

Detailed Changes: v0.2.0...v1.0.0

v0.2.0

v0.2.0 (2025-02-26)

Bug Fixes

• Load specified number of rows from point cloud file (#29, d5c79c9)

• Return type of make_labels_consecutive (#25, bdfea64)

Continuous Integration

• Add tests for Python 3.13 and PyTorch 2.5 (#36, 165e1f4)

• Install g++ in Docker image (#46, d5b9a17)

• Install g++ in Docker image (#50, 86142a8)

• Refactor code quality pipeline (#35, 0708d27)

• Update CI pipeline (#47, 7c4e96e)

• Update docs pipeline (#49, d87cd59)

Documentation

• Fix issues in docs build (#24, b1d4e59)

• Update CI badge in readme (#48, d9325c7)

Features

• File download (#30, f492736)

• File unzipping (#32, 70380b3)

• Fitting of oriented bounding box (#41, b1e60f0)

• Load specified number of rows from point cloud file (#26, 560b594)

• Max pooling (#37, 90160c3)

• Non-maximum suppression (#40, a3386d3)

• Point cloud shuffling (#39, ae22c01)

• Random sampling (#38, 5b95d86)

• Read and write CRS information (#28, f454bf6)

Refactoring

• Add pre-commit to dev dependencies (#34, f752def)

• Extend tests for pointtorch.io.download_file (#33, 5808abd)

• Fix mypy errors for pointtorch core module (#27, 764ed00)

• Laz writer (#42, 3e7f09c)

• Remove torch dependencies from pyproject.toml (#43, 36fd411)

• Remove torch dependencies from pyproject.toml (#44, a554893)

• Update code formatting (#31, 758a361)

• Update docs (#45, 01398bc)

---

Detailed Changes: v0.1.0...v0.2.0

v0.1.0

v0.1.0 (2024-12-14)

Continuous Integration

• Fix code coverage (#11, d64da1e)

• Update Dockerfile (#12, 82a7b1c)

Documentation

• Update setup instructions (#5, d1b5b96)

Features

• Knn search (#10, a7740bb)

• Operation to make labels consecutive (#16, bfda0f7)

• Point cloud data structure (#1, dac1516)

• Point cloud I/O (#3, b8c748e)

• Radius search (#19, fec74f2)

• Read method for point clouds (#6, a793141)

• Voxel downsampling (#14, 8daa0d7)

Refactoring

• Error message in PointCloud class (#21, f8a1d39)

• Fix pytyped marker (#2, d849582)

• Fix warnings in tests (#13, 03541c0)

• Hdf reader (#18, 5174bb5)

• Make_labels_consecutive operation (#23, bde3a1a)

• Numpy type annotations (#20, 390892f)

• Package imports (#22, dc60b4d)

• Reading and writing of point cloud identifiers in HDF files (#15, 0047612)

• Reformat package imports (#4, c3b90c6)

• Voxel downsampling (#17, b430358)

Testing

• Add tests for PointCloud.to() method (#8, b5825cb)

• Adds tests for pointtorch.read (#7, 354a51e)

• Extend tests for pointtorch.io.LasWriter (#9, 7fc1429)