Skip to content

Changelog

All notable changes to this project will be documented in this file.

The format is based on Keep a Changelog and this project adheres to Semantic Versioning.

0.8.3 - 2025-04-14

Compare with 0.8.2

Bug Fixes

  • adding presigns() to init (b4764e6 by yada).
  • adding more robust readme handling (61d0cd4 by yada).
  • openai generation broken (5e3cece by yada).

0.8.2 - 2025-02-22

Compare with 0.8.1

Bug Fixes

  • blocking var breaks mac compatability (cb38b48 by yada).

0.8.1 - 2025-02-22

Compare with 0.8.0

0.8.0 - 2025-02-22

Compare with 0.7.0

Features

  • adding dataset-card generation tool (a410353 by yada).

Bug Fixes

  • modifying file protect config (b26b632 by yada).
  • local or s3 backends: prevent overwriting important system files (39867dc by yada).
  • s3 backend: variable unbound when specifying path (7e97559 by yada).
  • ub.ls(): working on both huggingface model repo and datasets (70b070e by yada).
  • huggingface dataset: using proper saves() to work (and update) huggingface repo (7d382af by yada).
  • huggingface dataset: shortcircuit to download directly instead of double save (c0ef5e0 by yada).
  • broken s3 loading | broken: huggingface loading after backend downloads (36aaae2 by yada).
  • hf backends loading | broken: s3 loading (fbf3240 by yada).
  • use correct testing script (1984a08 by yada).
  • missing credentials / dependencies for test cases (672330a by yada).

Code Refactoring

  • allow loading from huggingface datasets as a file (aa3dc50 by yada).
  • let loader handle huggingface loads (400f7e1 by yada).

0.7.0 - 2025-02-20

Compare with 0.6.0

Features

  • adding dataset split control when using ub.saves("hf://org/dataset_repo") (20735f9 by yada).
  • adding toml / yaml loaders (3955e49 by yada).

Bug Fixes

  • missing type checks after update (f001f15 by yada).
  • s3 backend: missing arguments for ls (7c1d8e5 by yada).
  • incorrect huggingface data upload behavior when using datasets (bc5d773 by yada).

0.6.0 - 2025-02-13

Compare with 0.5.2

Features

  • adding improved huggingface dataset (with hfapi / hf datasets mixed backend) (96486d2 by yada).

0.5.2 - 2025-02-05

Features

  • Add LLM API utility (b0bc83d by openhands-agent)

Bug Fixes

  • HuggingFace URI loading incorrectly (ef8c9757f95 by openhands-agent)
  • gallery() to handle None values and display images properly (6f2cb39 by openhands-agent)

Compare with 0.5.1

0.5.1 - 2025-01-28

Compare with 0.5.0

0.5.0 - 2025-01-04

Compare with v0.4.13

Features

  • adding back ub.peeks() support (ac76e3a by yada).
  • adding proper colorized logger (9e0d758 by yada).
  • adding basic huggingface upload tools (f412a18 by yada).
  • adding huggingface storage backend (4c93076 by yada).
  • adding basic test suite and txt loader (bc16177 by yada).
  • adding basic working loader and tests (aa65789 by yada).
  • adding skeleton loader classes (a1e299f by yada).

Bug Fixes

  • adding colorama dep (1a9dbe4 by trojblue).
  • missing init in code (2799f19 by yada).
  • huggingface uploading an datasets object; s3 incorrect uri passed in (e5bb8d2 by yada).
  • adding colorlog dependency (a746230 by yada).
  • adding datasets dependency (071e5cc by yada).
  • color control characters getting written to logs (ecf1781 by yada).
  • image_loader: properly handling image loaders (cf5e535 by yada).
  • double write penalty at ub.saves() (ec07b9b by yada).
  • adding convert to rgb when using gallery (0e5826d by yada).

Code Refactoring

  • remoivng old files (5e59240 by yada).
  • adding project template (fd95d45 by yada).

v0.4.13 - 2024-11-17

Compare with v0.4.12

Features

  • adding ub.label_gallery() tool for data labelling (0fdac23 by yada).

v0.4.12 - 2024-09-30

Compare with v0.4.11

Features

  • allowing human-readable date in presigns() expiration (1381b9a by yada).

v0.4.11 - 2024-09-30

Compare with v0.4.10

Features

  • s3_client: adding generate_presigned_uri function; removing unused code (ccfcbc8 by yada).

v0.4.10 - 2024-07-22

Compare with v0.4.9

Bug Fixes

  • further ipython import fix (faac26f by yada).

v0.4.9 - 2024-07-22

Compare with v0.4.8

Bug Fixes

  • missing ipython dependency when calling ub.peeks() (989b471 by yada).

v0.4.8 - 2024-07-18

Compare with v0.4.7

Bug Fixes

  • concurrent_loads: fixing load order (74ac31f by yada).

v0.4.7 - 2024-07-18

Compare with v0.4.6

Features

  • ub.gallery(): adding notebook gallery (4f577a3 by yada).
  • adding ub.ls() wrapper for shorter ub.traverses() (e991c05 by yada).
  • uni_peeker: adding peek_df functionality (ed78393 by yada).
  • adding concurrent_loads() function (c3201d3 by yada).

Bug Fixes

  • ub.loads: tempfile naming error on windows (1d454d5 by yada).

v0.4.6 - 2024-06-28

Compare with v0.4.5

Bug Fixes

  • UniSaver: replace NaN with null if saving to dict or jsonl (5babfb8 by yada).

v0.4.5 - 2024-06-28

Compare with v0.4.4

Bug Fixes

  • adding graceful handling for errors when a line is unbale to be read (0965249 by yada).

v0.4.4 - 2024-06-14

Compare with v0.4.3

Bug Fixes

  • ub.saves(); a bug where ub.saves(list[str]) won't correctly save (3849e63 by yada).

v0.4.3 - 2024-06-13

Compare with v0.4.2

Features

  • add ability to save various formatted image files as png file (8142695 by yada).

Bug Fixes

  • resolved bug that prevents loading files from url (c9cb709 by yada).

v0.4.2 - 2024-05-30

Compare with v0.4.1

Features

  • extend include_extensions at ub.traverses() to take more than extensions (593b69c by yada).

v0.4.1 - 2024-05-18

Compare with v0.4.0

Bug Fixes

  • ub.traverses(): traversing a s3 directory will not return the dir itself (f0d85db by yada).

v0.4.0 - 2024-05-08

Compare with v0.3.20

Bug Fixes

  • resizer hangs when handling large number of images to be resized (9627894 by yada).

v0.3.20 - 2024-03-13

Compare with v0.3.19

Features

  • adding load feather support on uniloader (fee98ff by yada).

v0.3.19 - 2024-03-09

Compare with v0.3.18

Bug Fixes

  • incorrect behavior signature on ub.loads() on s3 uri (ca00b40 by yada).

v0.3.18 - 2024-03-06

Compare with v0.3.17

v0.3.17 - 2024-03-06

Compare with v0.3.16

Bug Fixes

  • unclosed image at uniresizer (b554630 by yada).

v0.3.16 - 2024-03-05

Compare with v0.3.15

Bug Fixes

  • adding back max_workers in uni_resizer (7c533f6 by yada).

v0.3.15 - 2024-02-12

Compare with v0.3.14

Features

  • adding debug_print argument to unibox.traverses() (66d1c02 by yada).

v0.3.14 - 2023-12-18

Compare with v0.3.13

Bug Fixes

  • using unibox.loads() with s3 (2ba1a57 by yada).

v0.3.13 - 2023-12-14

Compare with v0.3.12

Bug Fixes

  • include ipykernel version to avoid tqdm issues (a081b70 by yada).
  • traverses() with folder: incomplete s3 uri (3b4d25b by yada).
  • adding unit for traverse s3 (c7d00d7 by yada).

v0.3.12 - 2023-12-12

Compare with v0.3.11

Bug Fixes

  • traverses(s3): allowing traverses() to return dir info (0af5fdb by yada).

v0.3.11 - 2023-12-11

Compare with v0.3.10

Features

  • unibox.peeks(): adding list peek support & proper command use (c346f99 by yada).

v0.3.10 - 2023-12-11

Compare with v0.3.9

Features

  • adding support for s3 dir in unibox.traverses() (b90a97a by yada).
  • adding unipeeker and unibox.peeks() method for previewing data (7b7a7cd by yada).
  • adding traverse() in s3 client (3d786b3 by yada).

v0.3.9 - 2023-12-10

Compare with v0.3.8

Bug Fixes

  • unibox.loads(): add ability to properly load files from url (32c5593 by yada).

v0.3.8 - 2023-11-21

Compare with v0.3.6

Features

  • support saving string as a txt file (5748e96 by yada).
  • adding url support for unibox.loads() (9cc351b by yada).

v0.3.6 - 2023-11-12

Compare with v0.3.5

Features

  • support s3 uri in unibox.saves() (7e0db63 by yada).
  • support s3 uri in unibox.loads() (7850ed4 by yada).

Bug Fixes

  • incorrect filename when using unibox.saves() on s3 (27a4121 by yada).

v0.3.5 - 2023-11-02

Compare with v0.3.4

Features

  • s3client | bump version to 0.3.5 (a0db95a by yada).

v0.3.4 - 2023-11-01

Compare with v0.3.3

Bug Fixes

  • merges at init.py (ca658ed by yada).

v0.3.3 - 2023-11-01

Compare with v0.3.2

v0.3.2 - 2023-11-01

Compare with v0.3.1

Features

  • UniMerger: unibox.merges(data1, data2) (d156e29 by yada).

v0.3.1 - 2023-09-17

Compare with v0.3.0

v0.3.0 - 2023-09-03

Compare with v0.2.14

Features

  • image resizer: adding debug prints (bcb55df by yada).
  • image resizer: adding ability to skip existing images (c745646 by yada).
  • adding resizer-next (2ee732c by yada).

Bug Fixes

  • find existing images and remove them from jobs list (4ba16d9 by yada).
  • adding lower() to suffix before doing checks (cc17d5c by yada).

v0.2.14 - 2023-09-02

Compare with v0.2.13

Features

  • adding traverses() method for unibox (fe997a2 by yada).

v0.2.13 - 2023-08-26

Compare with v0.2.12

v0.2.12 - 2023-08-23

Compare with v0.2.11

Features

  • adding wip file renameer (148280c by yada).

v0.2.11 - 2023-08-21

Compare with v0.2.10

Bug Fixes

  • typing alias issue in python38 (6179f3e by yada).

v0.2.10 - 2023-08-20

Compare with v0.2.9

Features

  • reducing minimum python dependency from 3.10 to 3.8 (bcf96e6 by yada).

v0.2.9 - 2023-08-19

Compare with v0.2.8

v0.2.8 - 2023-08-18

Compare with v0.2.7

v0.2.7 - 2023-08-17

Compare with v0.2.6

v0.2.6 - 2023-08-15

Compare with v0.2.5

v0.2.5 - 2023-08-15

Compare with v0.2.3

Features

  • updating UniTraverser for stateful calls and filepath store (0e385e9 by yada).
  • adding UniTraverser class: code that traverses trhough directory (2e6ebff by yada).

v0.2.3 - 2023-08-14

Compare with v0.2.2

Features

  • remove pandas / pyarrow dep version (fe83df4 by yada).

v0.2.2 - 2023-08-13

Compare with v0.1.4.3

Features

  • adding unisaver (62d2eb2 by yada).
  • adding UniSaver and unibox.saves() method; bump version number to 0.2.0 (4eec7b2 by yada).

v0.1.4.3 - 2023-07-16

Compare with v0.1.4

Features

  • updating loads() for jsonl files (ccf54f3 by yada).

v0.1.4 - 2023-07-14

Compare with v0.1.3.5

Features

  • optimizing UniLoader class for csv & parquet (383215f by yada).

v0.1.3.5 - 2023-07-14

Compare with v0.1.3.4

Features

  • adding file mover; update image resizer (4d46099 by yada).

v0.1.3.4 - 2023-07-10

Compare with v0.1.3.3

Features

  • using ProcessPool instead of ThreadPool; before: 80it/s -> now: 105it/s (0180ad1 by yada).

v0.1.3.3 - 2023-07-10

Compare with v0.1.3.2

Bug Fixes

  • missing _resize (4d5fa90 by yada).

v0.1.3.2 - 2023-07-10

Compare with v0.1.3

Features

  • updating version number (3c25239 by yada).

Bug Fixes

  • not resizing image when min_size > actual size (85acf5d by yada).

v0.1.3 - 2023-07-10

Compare with 0.1.21

Features

  • updating cli & click requirement version (cd324bd by yada).
  • adding image resizer; refactor dir (8586022 by yada).

0.1.21 - 2023-07-06

Compare with v0.1.2.1

v0.1.2.1 - 2023-07-06

Compare with v0.1.2

v0.1.2 - 2023-07-06

Compare with v0.1

Bug Fixes

  • update version number (d5dc53a by yada).

v0.1 - 2023-07-06

Compare with first commit

Features

  • adding pipeline logger & loader (56043e4 by yada).
  • adding basic functionality (42d83cb by yada).

v0.5.0 - 2025-01-04

Compare with v0.4.13

Features

  • adding proper colorized logger (9e0d758 by yada).
  • adding basic huggingface upload tools (f412a18 by yada).
  • adding huggingface storage backend (4c93076 by yada).
  • adding basic test suite and txt loader (bc16177 by yada).
  • adding basic working loader and tests (aa65789 by yada).
  • adding skeleton loader classes (a1e299f by yada).

Bug Fixes

  • huggingface uploading an datasets object; s3 incorrect uri passed in (e5bb8d2 by yada).
  • adding colorlog dependency (a746230 by yada).
  • adding datasets dependency (071e5cc by yada).
  • color control characters getting written to logs (ecf1781 by yada).
  • image_loader: properly handling image loaders (cf5e535 by yada).
  • double write penalty at ub.saves() (ec07b9b by yada).
  • adding convert to rgb when using gallery (0e5826d by yada).

Code Refactoring

  • remoivng old files (5e59240 by yada).
  • adding project template (fd95d45 by yada).

v0.4.13 - 2024-11-17

Compare with v0.4.12

Features

  • adding ub.label_gallery() tool for data labelling (0fdac23 by yada).

v0.4.12 - 2024-09-30

Compare with v0.4.11

Features

  • allowing human-readable date in presigns() expiration (1381b9a by yada).

v0.4.11 - 2024-09-30

Compare with v0.4.10

Features

  • s3_client: adding generate_presigned_uri function; removing unused code (ccfcbc8 by yada).

v0.4.10 - 2024-07-22

Compare with v0.4.9

Bug Fixes

  • further ipython import fix (faac26f by yada).

v0.4.9 - 2024-07-22

Compare with v0.4.8

Bug Fixes

  • missing ipython dependency when calling ub.peeks() (989b471 by yada).

v0.4.8 - 2024-07-18

Compare with v0.4.7

Bug Fixes

  • concurrent_loads: fixing load order (74ac31f by yada).

v0.4.7 - 2024-07-18

Compare with v0.4.6

Features

  • ub.gallery(): adding notebook gallery (4f577a3 by yada).
  • adding ub.ls() wrapper for shorter ub.traverses() (e991c05 by yada).
  • uni_peeker: adding peek_df functionality (ed78393 by yada).
  • adding concurrent_loads() function (c3201d3 by yada).

Bug Fixes

  • ub.loads: tempfile naming error on windows (1d454d5 by yada).

v0.4.6 - 2024-06-28

Compare with v0.4.5

Bug Fixes

  • UniSaver: replace NaN with null if saving to dict or jsonl (5babfb8 by yada).

v0.4.5 - 2024-06-28

Compare with v0.4.4

Bug Fixes

  • adding graceful handling for errors when a line is unbale to be read (0965249 by yada).

v0.4.4 - 2024-06-14

Compare with v0.4.3

Bug Fixes

  • ub.saves(); a bug where ub.saves(list[str]) won't correctly save (3849e63 by yada).

v0.4.3 - 2024-06-13

Compare with v0.4.2

Features

  • add ability to save various formatted image files as png file (8142695 by yada).

Bug Fixes

  • resolved bug that prevents loading files from url (c9cb709 by yada).

v0.4.2 - 2024-05-30

Compare with v0.4.1

Features

  • extend include_extensions at ub.traverses() to take more than extensions (593b69c by yada).

v0.4.1 - 2024-05-18

Compare with v0.4.0

Bug Fixes

  • ub.traverses(): traversing a s3 directory will not return the dir itself (f0d85db by yada).

v0.4.0 - 2024-05-08

Compare with v0.3.20

Bug Fixes

  • resizer hangs when handling large number of images to be resized (9627894 by yada).

v0.3.20 - 2024-03-13

Compare with v0.3.19

Features

  • adding load feather support on uniloader (fee98ff by yada).

v0.3.19 - 2024-03-09

Compare with v0.3.18

Bug Fixes

  • incorrect behavior signature on ub.loads() on s3 uri (ca00b40 by yada).

v0.3.18 - 2024-03-06

Compare with v0.3.17

v0.3.17 - 2024-03-06

Compare with v0.3.16

Bug Fixes

  • unclosed image at uniresizer (b554630 by yada).

v0.3.16 - 2024-03-05

Compare with v0.3.15

Bug Fixes

  • adding back max_workers in uni_resizer (7c533f6 by yada).

v0.3.15 - 2024-02-12

Compare with v0.3.14

Features

  • adding debug_print argument to unibox.traverses() (66d1c02 by yada).

v0.3.14 - 2023-12-18

Compare with v0.3.13

Bug Fixes

  • using unibox.loads() with s3 (2ba1a57 by yada).

v0.3.13 - 2023-12-14

Compare with v0.3.12

Bug Fixes

  • include ipykernel version to avoid tqdm issues (a081b70 by yada).
  • traverses() with folder: incomplete s3 uri (3b4d25b by yada).
  • adding unit for traverse s3 (c7d00d7 by yada).

v0.3.12 - 2023-12-12

Compare with v0.3.11

Bug Fixes

  • traverses(s3): allowing traverses() to return dir info (0af5fdb by yada).

v0.3.11 - 2023-12-11

Compare with v0.3.10

Features

  • unibox.peeks(): adding list peek support & proper command use (c346f99 by yada).

v0.3.10 - 2023-12-11

Compare with v0.3.9

Features

  • adding support for s3 dir in unibox.traverses() (b90a97a by yada).
  • adding unipeeker and unibox.peeks() method for previewing data (7b7a7cd by yada).
  • adding traverse() in s3 client (3d786b3 by yada).

v0.3.9 - 2023-12-10

Compare with v0.3.8

Bug Fixes

  • unibox.loads(): add ability to properly load files from url (32c5593 by yada).

v0.3.8 - 2023-11-21

Compare with v0.3.6

Features

  • support saving string as a txt file (5748e96 by yada).
  • adding url support for unibox.loads() (9cc351b by yada).

v0.3.6 - 2023-11-12

Compare with v0.3.5

Features

  • support s3 uri in unibox.saves() (7e0db63 by yada).
  • support s3 uri in unibox.loads() (7850ed4 by yada).

Bug Fixes

  • incorrect filename when using unibox.saves() on s3 (27a4121 by yada).

v0.3.5 - 2023-11-02

Compare with v0.3.4

Features

  • s3client | bump version to 0.3.5 (a0db95a by yada).

v0.3.4 - 2023-11-01

Compare with v0.3.3

Bug Fixes

  • merges at init.py (ca658ed by yada).

v0.3.3 - 2023-11-01

Compare with v0.3.2

v0.3.2 - 2023-11-01

Compare with v0.3.1

Features

  • UniMerger: unibox.merges(data1, data2) (d156e29 by yada).

v0.3.1 - 2023-09-17

Compare with v0.3.0

v0.3.0 - 2023-09-03

Compare with v0.2.14

Features

  • image resizer: adding debug prints (bcb55df by yada).
  • image resizer: adding ability to skip existing images (c745646 by yada).
  • adding resizer-next (2ee732c by yada).

Bug Fixes

  • find existing images and remove them from jobs list (4ba16d9 by yada).
  • adding lower() to suffix before doing checks (cc17d5c by yada).

v0.2.14 - 2023-09-02

Compare with v0.2.13

Features

  • adding traverses() method for unibox (fe997a2 by yada).

v0.2.13 - 2023-08-26

Compare with v0.2.12

v0.2.12 - 2023-08-23

Compare with v0.2.11

Features

  • adding wip file renameer (148280c by yada).

v0.2.11 - 2023-08-21

Compare with v0.2.10

Bug Fixes

  • typing alias issue in python38 (6179f3e by yada).

v0.2.10 - 2023-08-20

Compare with v0.2.9

Features

  • reducing minimum python dependency from 3.10 to 3.8 (bcf96e6 by yada).

v0.2.9 - 2023-08-19

Compare with v0.2.8

v0.2.8 - 2023-08-18

Compare with v0.2.7

v0.2.7 - 2023-08-17

Compare with v0.2.6

v0.2.6 - 2023-08-15

Compare with v0.2.5

v0.2.5 - 2023-08-15

Compare with v0.2.3

Features

  • updating UniTraverser for stateful calls and filepath store (0e385e9 by yada).
  • adding UniTraverser class: code that traverses trhough directory (2e6ebff by yada).

v0.2.3 - 2023-08-14

Compare with v0.2.2

Features

  • remove pandas / pyarrow dep version (fe83df4 by yada).

v0.2.2 - 2023-08-13

Compare with v0.1.4.3

Features

  • adding unisaver (62d2eb2 by yada).
  • adding UniSaver and unibox.saves() method; bump version number to 0.2.0 (4eec7b2 by yada).

v0.1.4.3 - 2023-07-16

Compare with v0.1.4

Features

  • updating loads() for jsonl files (ccf54f3 by yada).

v0.1.4 - 2023-07-14

Compare with v0.1.3.5

Features

  • optimizing UniLoader class for csv & parquet (383215f by yada).

v0.1.3.5 - 2023-07-14

Compare with v0.1.3.4

Features

  • adding file mover; update image resizer (4d46099 by yada).

v0.1.3.4 - 2023-07-10

Compare with v0.1.3.3

Features

  • using ProcessPool instead of ThreadPool; before: 80it/s -> now: 105it/s (0180ad1 by yada).

v0.1.3.3 - 2023-07-10

Compare with v0.1.3.2

Bug Fixes

  • missing _resize (4d5fa90 by yada).

v0.1.3.2 - 2023-07-10

Compare with v0.1.3

Features

  • updating version number (3c25239 by yada).

Bug Fixes

  • not resizing image when min_size > actual size (85acf5d by yada).

v0.1.3 - 2023-07-10

Compare with 0.1.21

Features

  • updating cli & click requirement version (cd324bd by yada).
  • adding image resizer; refactor dir (8586022 by yada).

0.1.21 - 2023-07-06

Compare with v0.1.2.1

v0.1.2.1 - 2023-07-06

Compare with v0.1.2

v0.1.2 - 2023-07-06

Compare with v0.1

Bug Fixes

  • update version number (d5dc53a by yada).

v0.1 - 2023-07-06

Compare with first commit

Features

  • adding pipeline logger & loader (56043e4 by yada).
  • adding basic functionality (42d83cb by yada).