Metadata-Version: 2.4
Name: qlty
Version: 1.3.3
Summary: Unstitch and stitch back pytorch tensors
Author-email: "Petrus H. Zwart" <PHZwart@lbl.gov>
License: BSD
Project-URL: Homepage, https://github.com/phzwart/qlty
Project-URL: Documentation, https://qlty.readthedocs.io/
Project-URL: Repository, https://github.com/phzwart/qlty
Project-URL: Issues, https://github.com/phzwart/qlty/issues
Keywords: qlty,pytorch,tensor,stitch,unstitch,patch,image
Classifier: Development Status :: 4 - Beta
Classifier: Intended Audience :: Developers
Classifier: Intended Audience :: Science/Research
Classifier: License :: OSI Approved :: BSD License
Classifier: Natural Language :: English
Classifier: Programming Language :: Python :: 3
Classifier: Programming Language :: Python :: 3.8
Classifier: Programming Language :: Python :: 3.9
Classifier: Programming Language :: Python :: 3.10
Classifier: Programming Language :: Python :: 3.11
Classifier: Topic :: Scientific/Engineering
Classifier: Topic :: Software Development :: Libraries :: Python Modules
Requires-Python: >=3.8
Description-Content-Type: text/x-rst
License-File: LICENSE
License-File: AUTHORS.rst
Requires-Dist: numpy>=1.20.1
Requires-Dist: torch>=1.11
Requires-Dist: einops>=0.3.0
Requires-Dist: scipy>=1.8.1
Requires-Dist: zarr>=2.11.1
Requires-Dist: numba>=0.56.0
Requires-Dist: dask>=2021.12.0
Requires-Dist: umap-learn>=0.5.0
Requires-Dist: scikit-learn>=1.0.0
Provides-Extra: docs
Requires-Dist: sphinx; extra == "docs"
Requires-Dist: sphinx_rtd_theme; extra == "docs"
Provides-Extra: tests
Requires-Dist: pytest>=8.0.0; extra == "tests"
Requires-Dist: coverage>=7.0.0; extra == "tests"
Requires-Dist: tifffile>=2021.0.0; extra == "tests"
Dynamic: description
Dynamic: description-content-type
Dynamic: license-file

====
qlty
====

.. image:: https://img.shields.io/pypi/v/qlty.svg
        :target: https://pypi.python.org/pypi/qlty

.. image:: https://img.shields.io/travis/phzwart/qlty.svg
        :target: https://travis-ci.com/phzwart/qlty

.. image:: https://readthedocs.org/projects/qlty/badge/?version=latest
        :target: https://qlty.readthedocs.io/en/latest/?version=latest
        :alt: Documentation Status

qlty is a Python library designed to handle large 2D or 3D tensors efficiently by splitting them into smaller, manageable chunks. This library is particularly useful for processing large datasets that do not fit into memory, enabling chunked processing for machine learning workflows.

Features
--------

* Efficient tensor splitting and stitching
* Intelligent border handling to minimize artifacts
* Support for both in-memory and disk-cached processing
* 2D and 3D tensor support
* **2.5D Quilt** - Convert 3D volumetric data to multi-channel 2D by slicing Z dimension
* **Backend System** - Unified interface for multiple data sources (torch.Tensor, Zarr, HDF5, memory-mapped)
* **Image Stack Utilities** - Convert image file stacks to efficient Zarr format
* **False Color Visualization** - UMAP-based false-color visualization of 2D images
* Sparse data handling utilities
* Patch pair extraction helpers for partially overlapping regions in 2D and 3D
* Pre-tokenization utilities (``pretokenizer_2d``) for preparing patches for sequence models
* Numba acceleration for 2D stitching and batch token processing with parallel execution

Quick Start
-----------

Installation::

    pip install qlty torch zarr numpy einops dask numba

Basic Usage::

    import torch
    from qlty import NCYXQuilt

    # Create a quilt object
    quilt = NCYXQuilt(
        Y=128, X=128,
        window=(32, 32),
        step=(16, 16),
        border=(5, 5),
        border_weight=0.1
    )

    # Split data into patches
    data = torch.randn(10, 3, 128, 128)
    patches = quilt.unstitch(data)

    # Process patches (e.g., with a neural network)
    processed = your_model(patches)

    # Stitch back together
    reconstructed, weights = quilt.stitch(processed)

Documentation
-------------

Full documentation is available at https://qlty.readthedocs.io

* `Installation Guide <installation.html>`_
* `Usage Guide <usage.html>`_
* `Examples <examples.html>`_
* `API Reference <api.html>`_
* `Troubleshooting <troubleshooting.html>`_

Modules
-------

In-Memory Classes
~~~~~~~~~~~~~~~~~

* **NCYXQuilt**: 2D tensor splitting and stitching (shape: N, C, Y, X)
* **NCZYXQuilt**: 3D tensor splitting and stitching (shape: N, C, Z, Y, X)

Disk-Cached Classes
~~~~~~~~~~~~~~~~~~~

* **LargeNCYXQuilt**: 2D with on-disk caching using Zarr
* **LargeNCZYXQuilt**: 3D with on-disk caching using Zarr

License
-------

* Free software: BSD license

Credits
-------

This package was created with Cookiecutter_ and the `audreyr/cookiecutter-pypackage`_ project template.

.. _Cookiecutter: https://github.com/audreyr/cookiecutter
.. _`audreyr/cookiecutter-pypackage`: https://github.com/audreyr/cookiecutter-pypackage


=======
History
=======

1.3.3 (2025-11-23)
------------------

* **Progress Reporting** - Added comprehensive progress feedback to ``stack_files_to_zarr()``:
  * Progress bars using ``tqdm`` (if available) or periodic status messages
  * Shows overall progress for multiple stacks
  * Displays progress during image loading and zarr writing operations
  * Status messages showing stack information, worker count, and completion status
  * Helps monitor long-running operations on large image stacks

1.3.2 (2025-11-23)
------------------

* **Performance Optimization** - Enhanced multiprocessing in ``stack_files_to_zarr()``:
  * Implemented parallel load-and-write for large stacks (>10 images)
  * Workers now load images and write directly to zarr in parallel (not sequentially)
  * Dramatically reduces memory usage by avoiding loading all images into memory at once
  * Enables concurrent zarr writes using all available CPU cores
  * Optimized for very large stacks (e.g., 800+ images on 127-core systems)
  * Uses ``imap_unordered`` for better performance with many tasks

1.3.1 (2025-11-23)
------------------

* **Performance Improvement** - Added multiprocessing support to ``stack_files_to_zarr()``:
  * New ``num_workers`` parameter for parallel image loading
  * Auto-detects CPU count when ``num_workers=None`` (default)
  * Significantly improves performance for large image stacks by loading images in parallel
  * Maintains sequential zarr writing to preserve order

1.3.0 (2025-11-23)
------------------

* **Test Coverage Improvements** - Improved overall test coverage from 89% to 91%:
  * Fixed ``stack_to_zarr.py`` coverage from 38% to 94% by adding ``tifffile`` to test dependencies
  * Improved ``base.py`` coverage from 82% to 99%
  * Improved ``qlty2DLarge.py`` coverage to 100% (removed test function from production code)
  * Added comprehensive tests for ``qlty2DLarge`` and ``qlty3DLarge`` modules
  * Fixed CI coverage reporting to use ``coverage run`` directly for better accuracy
* **Code Quality** - Removed test/example functions from production modules
* **CI/CD** - Enhanced dependency verification and coverage reporting in CI workflows

1.2.3 (2025-11-23)
------------------

* **New: 2.5D Quilt Module** - Added ``NCZYX25DQuilt`` class for converting 3D volumetric data
  (N, C, Z, Y, X) into 2.5D multi-channel data by slicing the Z dimension into channels.
  Supports flexible channel specifications (identity, mean, std operations), selective z-slice
  processing, and two accumulation modes (2D planes or 3D stack).
* **New: Backend System** - Added comprehensive backend support for various data sources:
  * ``InMemoryBackend``: Wraps torch.Tensor for in-memory data
  * ``ZarrBackend``: On-demand loading from OME-Zarr files
  * ``HDF5Backend``: On-demand loading from HDF5 datasets
  * ``MemoryMappedBackend``: Memory-mapped numpy arrays
  * ``TensorLike3D``: Unified tensor-like interface for all backends
  * Convenience functions: ``from_zarr()``, ``from_hdf5()``, ``from_memmap()``
* **New: Image Stack to Zarr Utility** - Added ``stack_files_to_zarr()`` function in
  ``qlty.utils.stack_to_zarr`` for converting image file stacks (TIFF, PNG, etc.) into
  efficient Zarr format with automatic pattern matching, gap detection, and metadata storage.
* **New: False Color Visualization** - Added ``FalseColorGenerator`` class in ``qlty.utils.false_color_2D``
  for creating UMAP-based false-color visualizations of 2D images using patch-based dimensionality
  reduction.
* **Improved Test Coverage** - Added 65+ new tests across qlty2_5D, backends_2_5D, and stack_to_zarr
  modules, significantly improving coverage:
  * ``qlty2_5D.py``: 75% → 88% coverage
  * ``backends_2_5D.py``: 62% → 70% coverage
  * ``stack_to_zarr.py``: 38% → 94% coverage
* **CI Improvements** - Fixed coverage reporting in CI by using ``coverage run`` directly instead
  of pytest-cov to avoid torch import conflicts. Added coverage verification steps.

1.2.0 (2025-11-13)
------------------

* Added optional rotation-aware extraction for 2D patch pairs with matching overlap handling.
* Expanded tests and documentation to cover rotated patch workflows.

1.1.0 (2025-11-12)
------------------

* Restored Numba acceleration for 2D quilting via color-based parallel stitching that avoids write races.
* Expanded 3D patch-pair sampling tests to cover edge cases and fallback logic, driving coverage to 100%.
* Updated documentation to describe partially overlapping patch-pair utilities.
* Noted that NCZYXQuilt and the Large* variants still need analogous race-free acceleration.

0.1.0 (2021-10-20)
------------------

* First release on PyPI.

0.1.1 (some time ago)
---------------------

* Minor bug fixes

0.1.2. (2022-9-13)
------------------

* Support for N-channel 3D tensors
* On disc-accmulation for large datasets


0.1.3. (2022-9-13)
------------------

* Cleanup and ready to redistribute


0.1.4. (2023-8-28)
------------------

* Bug fix / behavoir change

0.1.5. (2023-12-28)
-------------------

* Changes to qlty3DLarge:
  bug fixes
  normalizing / averaging now done by dask

0.1.6. (2024-03-10)
-------------------
  bug fixes

0.1.7. (2024-03-12)
-------------------
*  bug fix in border tensor definition.

0.2.0. (2024-03-12)
-------------------
*  bug fixes
*  2DLarge, mimics 3DLarge
