k-means++ #71

jakeISAC · 2026-05-17T21:12:19+02:00

jakeISAC commented

2026-05-17 21:12:19 +02:00

No description provided.

jakeISAC added 7 commits

2026-05-17 21:12:19 +02:00

Refactor k-means clustering and optimize scheduling logic: 825c442c3e

- Implement k-means++ initialization for improved centroid selection.
- Replace natural logarithm with base-2 logarithm in `create_log_space`.
- Refine distance calculations for better precision using `OrderedFloat`.
- Remove `NOISE_RANGE` constant and redundant noise injection mechanism.
- Fix minor sorting logic in `dir_preprocessor.rs`.
- Adjust default schedule parameter in test to reflect updated logic.
- Bump version to 1.2.11 in Cargo.toml.

Remove unnecessary error propagation for empty directories in dir_structure_analysis.rs 0536bf944a

Remove unnecessary error propagation for empty directories in dir_structure_analysis.rs 947b24f1c0

Add configurable thread limit for multithreaded copies: 19c73d0dfa

- Introduce `--threads` argument to control maximum concurrent tasks.
- Limit task spawning in `copy_multithreaded` to reduce memory usage.
- Add semaphore for concurrent group control to enhance stability.

Bump version to 1.2.12 in Cargo.toml. b520f5a33a

Fix group_permit handling in copy_multithreaded to ensure proper task spawning control. decd2592f3

Rename task_count and threads to cfs_chunk_count and concurrent_limit for improved clarity: 87a108e1a2

- Updated argument names to better reflect their functionality.
- Adjusted related scheduler and copying logic accordingly.
- Improved comments and logs for enhanced code readability and maintainability.

jakeISAC merged commit 9861efff1b into main

2026-05-17 21:12:24 +02:00

jakeISAC referenced this pull request from a commit

2026-05-17 21:12:26 +02:00

Merge pull request 'k-means++' (#71) from k-means++ into main

jakeISAC deleted branch k-means++

2026-05-17 21:12:51 +02:00

Sign in to join this conversation.

No reviewers

No labels

No milestone

No project