search

Dana-Farber-AIOS / pathml

Tools for computational pathology
https://pathml.org
GNU General Public License v2.0
392 stars 84 forks source link

PathML 3.0.0: New Features and User Experience Improvements #429

Closed VarunUllanat closed 2 weeks ago

VarunUllanat commented 8 months ago

This release introduces powerful new features and updates designed to streamline computational pathology workflows and expand the toolkit available to researchers. Here's what's new:

New Features:

  1. Graph API Integration: Build sophisticated graph-based computational pathology pipelines. This feature supports entity-based graph construction for cells and tissues, enabling advanced analyses and machine learning (ML) model training with Graph neural networks. Inference API with ONNX Support: Leverage the flexibility of the Open Neural Network Exchange (ONNX) format with
  2. PathML's new Inference API. This addition makes it easier for researchers to integrate any neural network architecture into their analysis pipelines, enhancing model-based inference capabilities.
  3. Model Zoo Expansion: Access cutting-edge models like Mesmer and HoVer-Net directly within your workflows, courtesy of our hugging face repository, with more models coming soon.
  4. Tile-Stitching API: Simplify preprocessing with the new tile-stitching utility. By integrating QuPath's capabilities via JPype, this release has made it easier for Python users to process diverse TIFF inputs, supporting spectrally unmixed image processing with efficiency.

Improvements:

  1. Preprocessing Transforms Update: This release has streamlined PathML's preprocessing transforms, reducing dependencies on external packages such as spams and DeepCell. This update ensures smoother operation and easier installation.
  2. Enhanced Documentation and Resources: Get up and running with PathML faster thanks to updated documentation, example notebooks, and README instructions. Our new docker images also simplify the setup process.
  3. GitHub Actions for Windows and macOS: We've implemented new GitHub Actions workflows to automate testing and ensure the compatibility of PathML across Windows and macOS platforms.
codecov-commenter commented 7 months ago

:warning: Please install the 'codecov app svg image' to ensure uploads and comments are reliably processed by Codecov.

Codecov Report

Attention: Patch coverage is 95.79465% with 77 lines in your changes missing coverage. Please review.

Project coverage is 94.48%. Comparing base (f86c5c6) to head (58c8f6a).

Files with missing lines Patch % Lines
pathml/preprocessing/workflows.py 0.00% 31 Missing :warning:
pathml/graph/preprocessing.py 95.22% 17 Missing :warning:
pathml/inference/mesmer_utils.py 91.34% 9 Missing :warning:
pathml/ml/utils.py 88.37% 5 Missing :warning:
pathml/ml/models/hovernet.py 98.87% 4 Missing :warning:
pathml/datasets/datasets.py 98.46% 3 Missing :warning:
pathml/datasets/utils.py 96.55% 3 Missing :warning:
pathml/preprocessing/tilestitcher.py 98.55% 3 Missing :warning:
pathml/core/slide_data.py 75.00% 1 Missing :warning:
pathml/preprocessing/transforms.py 98.14% 1 Missing :warning:

:exclamation: Your organization needs to install the Codecov GitHub app to enable full functionality.

Additional details and impacted files ```diff @@ Coverage Diff @@ ## master #429 +/- ## ========================================== + Coverage 86.37% 94.48% +8.10% ========================================== Files 27 39 +12 Lines 2503 3826 +1323 ========================================== + Hits 2162 3615 +1453 + Misses 341 211 -130 ```

:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.