Fix ElasticTransform blur kernel and add missing args

[liamjwang]

Hello, a few suggested changes to ElasticTransform.

I think the approximate argument was reversed by mistake.

Also, currently the blur step is skipped if the kernel size is (0, 0).
However, I think it was intended to let cv2.GaussianBlur automatically compute the kernel size.

With this change, the behavior matches the ElasticTransform output from older versions (e.g. 1.4.15).

Also added border_mode, fill, and fill_mask args.

Summary by Sourcery

Fixes issues in the ElasticTransform augmentation, including a reversed approximate argument, skipping the blur step when the kernel size is (0, 0), and missing border_mode, fill, and fill_mask arguments.

Bug Fixes:

• Fixes a bug where the blur step was skipped in ElasticTransform when the kernel size was (0, 0).
• Fixes a bug where the approximate argument was reversed in ElasticTransform.

Enhancements:

• Adds border_mode, fill, and fill_mask arguments to ElasticTransform to provide more control over the transformation.

[sourcery-ai]

Reviewer's Guide by Sourcery

This pull request fixes the ElasticTransform blur kernel, corrects the approximate argument, and adds border_mode, fill, and fill_mask arguments. The blur step in generate_displacement_fields was modified to allow cv2.GaussianBlur to automatically compute the kernel size when kernel_size is (0, 0). The approximate argument logic was corrected by swapping the values assigned to the kernel_size variable. The border_mode, fill, and fill_mask arguments were added to the InitSchema and ElasticTransform classes, including updates to the get_transform_init_args_names methods.

Sequence diagram for generate_displacement_fields with blur

sequenceDiagram
    participant fgeometric.generate_displacement_fields
    participant cv2.GaussianBlur

    fgeometric.generate_displacement_fields->>cv2.GaussianBlur: Apply Gaussian blur to displacement fields
    activate cv2.GaussianBlur
    cv2.GaussianBlur-->>fgeometric.generate_displacement_fields: Blurred displacement fields
    deactivate cv2.GaussianBlur
    fgeometric.generate_displacement_fields->>fgeometric.generate_displacement_fields: Scale by alpha

Loading

Updated class diagram for ElasticTransform

classDiagram
    class ElasticTransform {
        -alpha: float
        -sigma: float
        -approximate: bool
        -same_dxdy: bool
        -noise_distribution: str
        -keypoint_remapping_method: str
        -border_mode: int
        -fill: tuple[float, ...] | float
        -fill_mask: tuple[float, ...] | float
        +__init__(
            alpha: float,
            sigma: float,
            approximate: bool,
            same_dxdy: bool,
            interpolation: int,
            mask_interpolation: int,
            noise_distribution: str,
            keypoint_remapping_method: str,
            border_mode: int,
            fill: tuple[float, ...] | float,
            fill_mask: tuple[float, ...] | float,
            p: float
        )
        +get_params_dependent_on_data(data: dict[str, Any]) : dict[str, Any]
        +get_transform_init_args_names() : tuple[str, ...]
    }
    ElasticTransform --|> BaseDistortion
    note for ElasticTransform "Added border_mode, fill, and fill_mask attributes and constructor parameters."

Loading

File-Level Changes

Change	Details	Files
Added border_mode, fill, and fill_mask arguments to the InitSchema and ElasticTransform classes.	Added border_mode to the InitSchema class. Added fill to the InitSchema class. Added fill_mask to the InitSchema class. Added border_mode to the ElasticTransform class. Added fill to the ElasticTransform class. Added fill_mask to the ElasticTransform class. Added border_mode, fill, and fill_mask to the get_transform_init_args_names method in both classes.	albumentations/augmentations/geometric/transforms.py
Corrected the logic for the approximate argument in the ElasticTransform class.	The kernel_size is now set to (17, 17) when approximate is true, and (0, 0) when approximate is false.	albumentations/augmentations/geometric/transforms.py
Modified the Gaussian blur application in generate_displacement_fields to allow cv2.GaussianBlur to automatically compute the kernel size when kernel_size is (0, 0).	Removed the conditional check that skipped the blur step when kernel_size was (0, 0). The Gaussian blur is now always applied.	albumentations/augmentations/geometric/functional.py

---

Tips and commands

Interacting with Sourcery

• Trigger a new review: Comment @sourcery-ai review on the pull request.
• Continue discussions: Reply directly to Sourcery's review comments.
• Generate a GitHub issue from a review comment: Ask Sourcery to create an
  issue from a review comment by replying to it. You can also reply to a
  review comment with @sourcery-ai issue to create an issue from it.
• Generate a pull request title: Write @sourcery-ai anywhere in the pull
  request title to generate a title at any time. You can also comment
  @sourcery-ai title on the pull request to (re-)generate the title at any time.
• Generate a pull request summary: Write @sourcery-ai summary anywhere in
  the pull request body to generate a PR summary at any time exactly where you
  want it. You can also comment @sourcery-ai summary on the pull request to
  (re-)generate the summary at any time.
• Generate reviewer's guide: Comment @sourcery-ai guide on the pull
  request to (re-)generate the reviewer's guide at any time.
• Resolve all Sourcery comments: Comment @sourcery-ai resolve on the
  pull request to resolve all Sourcery comments. Useful if you've already
  addressed all the comments and don't want to see them anymore.
• Dismiss all Sourcery reviews: Comment @sourcery-ai dismiss on the pull
  request to dismiss all existing Sourcery reviews. Especially useful if you
  want to start fresh with a new review - don't forget to comment
  @sourcery-ai review to trigger a new review!
• Generate a plan of action for an issue: Comment @sourcery-ai plan on
  an issue to generate a plan of action for it.

Customizing Your Experience

Access your dashboard to:

• Enable or disable review features such as the Sourcery-generated pull request
  summary, the reviewer's guide, and others.
• Change the review language.
• Add, remove or edit custom review instructions.
• Adjust other review settings.

Getting Help

• Contact our support team for questions or feedback.
• Visit our documentation for detailed guides and information.
• Keep in touch with the Sourcery team by following us on X/Twitter, LinkedIn or GitHub.

[sourcery-ai]

Hey @liamjwang - I've reviewed your changes - here's some feedback:

Overall Comments:

• It might be worth adding a test case that specifically checks the behavior of ElasticTransform with the new border_mode, fill, and fill_mask parameters.

Here's what I looked at during the review

• 🟢 General issues: all looks good
• 🟢 Security: all looks good
• 🟢 Testing: all looks good
• 🟡 Complexity: 1 issue found
• 🟢 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 and I'll use the feedback to improve your reviews.}

[ternaus]

border_mode
fill
fill_mask

Do not really have material effect in Elastic Transform

[liamjwang]

Thank you!

[ternaus]

@liamjwang you still need to update the code for all checks to pass.

Please read Contributoru's guide - https://github.com/albumentations-team/albumentations/blob/main/CONTRIBUTING.md