v0.1.2-rc1

[0.1.2-rc1]

Added

• Expose merge threshold in python bindings. This makes it possible to merge
  communities when they share many edges together after Leiden.

v0.1.1

[0.1.1] - 2026-04-23

Fixed

• Fix output for Leiden's python binding

v0.1.0

[0.1.0] - 2026-04-16

Initial release: Rust implementation of performance-critical community-detection and graph routines from pixelator, with Python bindings, a CLI, and published wheels.

Added

• Rust library (pixelator_core)
  • Fast label propagation with pluggable assignment strategies and Parquet-based graph I/O.
  • Leiden community detection with a modularity quality function on weighted partitioned graphs.
  • Hybrid pipeline combining fast label propagation, graph aggregation, optional Leiden refinement, and multiplet recovery (new_graph).
  • Shared utilities: graph types and statistics, node partitioning (fast and Leiden-specific), linear algebra helpers, Parquet read/write for edge lists and node partitions.
• CLI (community-detection)
  • flp — run fast label propagation on a Parquet edge list and write node partitions.
  • leiden — run Leiden with configurable resolution and write partitions.
  • stats — print basic graph statistics (nodes, edge weight sum, components, largest-component fraction).
• Python bindings (pixelator_core / package name pixelgen-pixelator-core)
  • run_label_propagation, run_leiden, run_hybrid_community_detection, find_graph_statistics, and GraphProperties-style results for hybrid workflows.
  • Module __version__ aligned with the Rust crate version.

v0.1.0-rc3

[0.1.0] - TBD

Initial release: Rust implementation of performance-critical community-detection and graph routines from pixelator, with Python bindings, a CLI, and published wheels.

Added

• Rust library (pixelator_core)
  • Fast label propagation with pluggable assignment strategies and Parquet-based graph I/O.
  • Leiden community detection with a modularity quality function on weighted partitioned graphs.
  • Hybrid pipeline combining fast label propagation, graph aggregation, optional Leiden refinement, and multiplet recovery (new_graph).
  • Shared utilities: graph types and statistics, node partitioning (fast and Leiden-specific), linear algebra helpers, Parquet read/write for edge lists and node partitions.
• CLI (community-detection)
  • flp — run fast label propagation on a Parquet edge list and write node partitions.
  • leiden — run Leiden with configurable resolution and write partitions.
  • stats — print basic graph statistics (nodes, edge weight sum, components, largest-component fraction).
• Python bindings (pixelator_core / package name pixelgen-pixelator-core)
  • run_label_propagation, run_leiden, run_hybrid_community_detection, find_graph_statistics, and GraphProperties-style results for hybrid workflows.
  • Module __version__ aligned with the Rust crate version.

v0.1.0-rc2

[0.1.0] - TBD

Initial release: Rust implementation of performance-critical community-detection and graph routines from pixelator, with Python bindings, a CLI, and published wheels.

Added

• Rust library (pixelator_core)
  • Fast label propagation with pluggable assignment strategies and Parquet-based graph I/O.
  • Leiden community detection with a modularity quality function on weighted partitioned graphs.
  • Hybrid pipeline combining fast label propagation, graph aggregation, optional Leiden refinement, and multiplet recovery (new_graph).
  • Shared utilities: graph types and statistics, node partitioning (fast and Leiden-specific), linear algebra helpers, Parquet read/write for edge lists and node partitions.
• CLI (community-detection)
  • flp — run fast label propagation on a Parquet edge list and write node partitions.
  • leiden — run Leiden with configurable resolution and write partitions.
  • stats — print basic graph statistics (nodes, edge weight sum, components, largest-component fraction).
• Python bindings (pixelator_core / package name pixelgen-pixelator-core)
  • run_label_propagation, run_leiden, run_hybrid_community_detection, find_graph_statistics, and GraphProperties-style results for hybrid workflows.
  • Module __version__ aligned with the Rust crate version.

v0.1.0-rc1

[0.1.0] - TBD

Initial release: Rust implementation of performance-critical community-detection and graph routines from pixelator, with Python bindings, a CLI, and published wheels.

Added

• Rust library (pixelator_core)
  • Fast label propagation with pluggable assignment strategies and Parquet-based graph I/O.
  • Leiden community detection with a modularity quality function on weighted partitioned graphs.
  • Hybrid pipeline combining fast label propagation, graph aggregation, optional Leiden refinement, and multiplet recovery (new_graph).
  • Shared utilities: graph types and statistics, node partitioning (fast and Leiden-specific), linear algebra helpers, Parquet read/write for edge lists and node partitions.
• CLI (community-detection)
  • flp — run fast label propagation on a Parquet edge list and write node partitions.
  • leiden — run Leiden with configurable resolution and write partitions.
  • stats — print basic graph statistics (nodes, edge weight sum, components, largest-component fraction).
• Python bindings (pixelator_core / package name pixelgen-pixelator-core)
  • run_label_propagation, run_leiden, run_hybrid_community_detection, find_graph_statistics, and GraphProperties-style results for hybrid workflows.
  • Module __version__ aligned with the Rust crate version.