real_multi_modality

[xingzhongyu]

No description provided.

[xingzhongyu]

Get train, valid and test data through a single h5ad

[xingzhongyu]

Get specific run_kwargs to search

[xingzhongyu]

see PyCQA/docformatter#287

[xingzhongyu]

Calculate the similarity of data sets and return the most similar data set in the atlas for the query data set

[xingzhongyu]

FilterCellsCommonMod: Filters single-cell data to retain only common cells between two modalities.

This transformation filters two single-cell modalities (mod1 and mod2) to retain only the cells
that are shared between them. Optionally, a solution dataset (sol) containing cell labels or annotations
can be filtered to match the common cells.

[xingzhongyu]

Transform the data set using the tfidf method

[xingzhongyu]

Record the correspondence between atlas datasets and algorithms for statistical algorithm results

[xingzhongyu]

Dropbox Data Upload Utility

A module providing utility functions for uploading single-cell data to Dropbox and generating shareable links.

Core Features

• upload_file_to_dropbox: Upload local files to Dropbox
• file_upload: Handle large files using chunked upload
• create_shared_link: Create or retrieve Dropbox shared links
• get_link: Upload files and get downloadable shared links
• get_ans: Generate dataset metadata dictionary

Technical Highlights

• Supports chunked upload for large files
• Handles authentication errors gracefully
• Provides dataset metadata management
• Optimized for single-cell data (AnnData format)

[xingzhongyu]

Cache some sweep data for the get_runs method

[xingzhongyu]

Merge the metadata and sweep data of the dataset

[xingzhongyu]

Introducing the atlas verification process

[xingzhongyu]

Calculate the similarity between query datasets and atlas under different tissues

[xingzhongyu]

Calculate optimal weights for similarity metrics in cell type annotation.

[xingzhongyu]

Process tissue similarity matrices across different datasets.

This script handles the processing of tissue similarity data stored in Excel files:

• Converts and unifies complex number representations
• Handles data across multiple tissues (blood, brain, heart, intestine, kidney, lung, pancreas)
• Combines data from multiple Excel sheets and files
• Provides different processing methods (by cell, row, or column)
• Exports processed data to both Excel and CSV formats

[xingzhongyu]

A comprehensive test suite for AnnDataSimilarity class that validates:

1. Similarity computation across different metrics (wasserstein, Hausdorff, chamfer, etc.)
2. Data preprocessing functionality including gene filtering
3. Cell sampling capabilities
4. Error handling for invalid methods

The tests use mock AnnData objects with random data to verify the correctness and robustness of similarity calculations between single-cell datasets.

[xingzhongyu]

Test Suite for Web Result Processing Module

This test suite covers the functionality of the get_result_web.py module, which handles web-based result processing and data management. The tests include:

Key Components Tested

• check_identical_strings: Validates string consistency checking
• spilt_web: Tests W&B URL parsing functionality
• check_exist: Verifies directory and file existence checks
• write_ans: Tests data writing and updating mechanisms

Test Coverage

• Input validation
• Error handling
• File I/O operations
• Data processing logic
• Integration with W&B API (using mocks)

Test Environment

• Uses pytest fixtures for environment setup
• Implements temporary directory handling
• Mocks external dependencies (wandb API, ATLAS_DIR, META_DIR)
• Includes both positive and negative test cases

The test suite ensures robust handling of result processing, data persistence, and configuration management.

[xingzhongyu]

Test Suite for MuData to AnnData Transformation

This test suite validates the functionality of the add_mod_and_transform decorator and its application on data transformation operations between MuData and AnnData objects.

Key Test Cases

• Initialization of decorated classes with and without modality parameters
• Data transformation on AnnData objects
• Selective modality transformation on MuData objects
• Error handling for invalid modalities
• Preservation of class metadata after decoration
• Handling of additional arguments during transformation

The test suite ensures that:

1. The decorator properly wraps class initialization and call methods
2. Data transformations are correctly applied to specified modalities
3. Unspecified modalities remain unchanged
4. Class metadata and documentation are preserved
5. Error cases are properly handled

All tests use mock data fixtures for consistent and isolated testing.

[xingzhongyu]

This document explains how to integrate new single-cell algorithms into the project's automatic search framework.

[xingzhongyu]

This document explains how to integrate new algorithms into the project's automatic search framework.

[xingzhongyu]

Test suite for the Atlas similarity calculation functionality.
This test verifies that the main function correctly returns:

1. The most similar dataset from the atlas
2. Its corresponding configuration settings
3. The similarity score

[xingzhongyu]

This script finds the most similar dataset from the atlas for a given user-uploaded dataset
It calculates similarity scores and returns the best matching dataset along with its configurations