Fix np array int labels

[ternaus]

Fixes: #2324

Summary by Sourcery

Bug Fixes:

• Fix issue where integer labels in numpy arrays were being converted to floats.

[sourcery-ai]

Reviewer's Guide by Sourcery

This pull request refactors the label encoding logic to correctly handle numpy arrays and preserve the original data types of labels. It introduces a LabelManager class to manage label encoding and decoding, replacing the previous LabelEncoder implementation within the DataProcessor class. This change ensures that labels, whether numerical or string-based, are processed without losing their original type or data structure.

Sequence diagram for label processing flow

sequenceDiagram
    participant DP as DataProcessor
    participant LM as LabelManager
    participant LE as LabelEncoder

    DP->>LM: process_field(data_name, label_field, data)
    activate LM
    LM->>LM: _analyze_input(data)

    alt is numerical data
        LM->>LM: Convert to float32 array
    else non-numerical data
        LM->>LE: Create new encoder
        LM->>LE: fit_transform(data)
    end

    LM-->>DP: Return encoded data
    deactivate LM

    Note over DP,LM: Later when restoring...

    DP->>LM: restore_field(data_name, label_field, encoded_data)
    activate LM

    alt is numerical data
        LM->>LM: Restore original dtype
    else non-numerical data
        LM->>LE: inverse_transform(encoded_data)
    end

    LM->>LM: _restore_type(decoded_data)
    LM-->>DP: Return restored data
    deactivate LM

Loading

Class diagram showing the new label management structure

classDiagram
    class LabelManager {
      -metadata: dict
      +process_field(data_name: str, label_field: str, field_data: Any): np.ndarray
      +restore_field(data_name: str, label_field: str, encoded_data: np.ndarray): Any
      -_analyze_input(field_data: Any): LabelMetadata
      -_encode_data(field_data: Any, metadata: LabelMetadata): np.ndarray
      -_decode_data(encoded_data: np.ndarray, metadata: LabelMetadata): np.ndarray
      -_restore_type(decoded_data: np.ndarray, metadata: LabelMetadata): Any
      +handle_empty_data(): list
    }

    class LabelMetadata {
      +input_type: type
      +is_numerical: bool
      +dtype: np.dtype
      +encoder: LabelEncoder
    }

    class LabelEncoder {
      -classes_: dict
      -inverse_classes_: dict
      -num_classes: int
      -is_numerical: bool
      +fit(y: Sequence|np.ndarray): LabelEncoder
      +transform(y: Sequence|np.ndarray): np.ndarray
      +fit_transform(y: Sequence|np.ndarray): np.ndarray
      +inverse_transform(y: Sequence|np.ndarray): np.ndarray
    }

    class DataProcessor {
      -label_manager: LabelManager
    }

    DataProcessor --> LabelManager
    LabelManager --> LabelMetadata
    LabelMetadata --> LabelEncoder

Loading

File-Level Changes

Change	Details	Files
Introduced LabelManager class to handle label encoding and decoding.	Created a new LabelManager class to encapsulate label processing logic. Added LabelMetadata dataclass to store metadata about label fields. Implemented methods for analyzing, encoding, decoding, and restoring label fields.	albumentations/core/label_manager.py
Removed LabelEncoder from DataProcessor and integrated LabelManager.	Removed the LabelEncoder class from albumentations/core/utils.py. Replaced the label_encoders and is_numerical_label dictionaries with a label_manager instance. Modified _process_label_fields to use label_manager for encoding labels. Modified _decode_label_field to use label_manager for decoding labels. Updated _handle_empty_data_array to use label_manager for handling empty labels.	albumentations/core/utils.py
Preserved original data types of labels.	Modified LabelManager to preserve the original type (list or numpy array) and dtype of labels. Updated tests to verify that label types and dtypes are preserved after transformations. Ensured that string labels are preserved exactly. Added a test case for empty labels.	albumentations/core/label_manager.py albumentations/core/utils.py tests/test_core.py
Updated CoarseDropout to use LabelManager.	Modified get_boxes_from_bboxes in CoarseDropout to use label_manager to access the label encoder.	albumentations/augmentations/dropout/coarse_dropout.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 @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: 1 issue 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 and I'll use the feedback to improve your reviews.}