Bugfixes in textimage

[ternaus]

Also fixes: #1926

Summary by Sourcery

Fix a bug in fog augmentation by correcting the function call for center calculation. Enhance several image transformation classes with improved functionality, documentation, and examples. Update tests to cover new features and remove redundant tests.

Bug Fixes:

• Fix incorrect function call for calculating the center of an image in fog augmentation.

Enhancements:

• Improve the ToRGB transform by adding support for specifying the number of output channels and updating the documentation.
• Enhance the ToSepia transform with detailed documentation and examples.
• Refactor the Superpixels transform to improve performance and update documentation.
• Update the TemplateTransform to handle different types of font colors and improve blending logic.
• Refactor the UnsharpMask transform to improve edge detection and sharpening logic.

Tests:

• Add tests for the TextImage augmentation to ensure metadata handling and image transformation.
• Remove redundant test for TemplateTransform with incorrect size.

[sourcery-ai]

Reviewer's Guide by Sourcery

This pull request implements several bugfixes and improvements in the 'albumentations' library, focusing on the 'textimage' module and other related components. The changes include code refactoring, performance optimizations, and enhanced documentation for various image augmentation transforms.

File-Level Changes

Change	Details	Files
Refactored and improved the TextImage transform	Updated the TextImage class to use a tuple for stopwords instead of a list Modified the preprocess_metadata method to use numpy array for bbox checking Improved error handling and type checking in the draw_text_on_multi_channel_image function	albumentations/augmentations/text/transforms.py albumentations/augmentations/text/functional.py
Optimized and refactored geometric transformations	Moved center and center_bbox functions from functional.py to geometric/functional.py Updated imports and function calls related to geometric transformations in various files	albumentations/augmentations/geometric/functional.py albumentations/augmentations/geometric/rotate.py albumentations/augmentations/geometric/transforms.py albumentations/augmentations/domain_adaptation_functional.py
Enhanced documentation and type hints for various transforms	Added detailed docstrings for ToRGB, ToSepia, ToFloat, and Superpixels transforms Improved type hints and parameter descriptions for several transforms Added examples and mathematical formulations to the docstrings	albumentations/augmentations/transforms.py
Refactored and optimized image processing functions	Moved bbox_from_mask and mask_from_bbox functions to bbox_utils.py Optimized the superpixels function for better performance Improved the unsharp_mask function to handle single-channel images correctly	albumentations/augmentations/functional.py albumentations/core/bbox_utils.py
Updated and fixed various test cases	Modified test cases to accommodate changes in TextImage and TemplateTransform Updated test cases for geometric transformations Added new test cases for multi-channel image processing	tests/test_transforms.py tests/test_augmentations.py tests/test_core.py tests/functional/test_affine.py tests/test_text.py

---

Tips

• Trigger a new Sourcery review by commenting @sourcery-ai review on the pull request.
• Continue your discussion with Sourcery by replying directly to review comments.
• You can change your review settings at any time by accessing your dashboard:
  • Enable or disable the Sourcery-generated pull request summary or reviewer's guide;
  • Change the review language;
• You can always contact us if you have any questions or feedback.

[sourcery-ai]

Hey @ternaus - I've reviewed your changes and they look great!

Here's what I looked at during the review

• 🟡 General issues: 1 issue found
• 🟢 Security: all looks good
• 🟡 Testing: 5 issues found
• 🟢 Complexity: all looks good
• 🟢 Documentation: all looks good

---

Sourcery is free for open source - if you like our reviews please consider sharing them ✨

• X
• Mastodon
• LinkedIn
• Facebook

_{Help me be more useful! Please click 👍 or 👎 on each comment to tell me if it was helpful.}