Merge pull request #46 from v0lta/v0.1.5

merge v0.1.5

Author: v0lta

.flake8

@@ -36,4 +36,3 @@ import-order-style = pycharm
 application-import-names =
     ptwt
     tests
-format = ${cyan}%(path)s${reset}:${yellow_bold}%(row)d${reset}:${green_bold}%(col)d${reset}: ${red_bold}%(code)s${reset} %(text)s

.gitignore

@@ -5,6 +5,7 @@ htmlcov
 .noseids
 examples/ffhq_style_gan/
 examples/ffhq_style_gan.zip
+examples/data
 
 
 # Byte-compiled / optimized / DLL files

CONTRIBUTING.md

@@ -0,0 +1,25 @@
+## Contributing to the PyTorch-Wavelet-Toolbox
+
+Contributions to the PyTorch-Wavelet-Toolbox are always welcome!
+
+### Development workflow:
+We use nox to run our unit tests. Before creating your pull request follow these three steps.
+
+1. Make sure all unit tests are passing.
+Run:
+``` bash
+nox -s test
+```
+to check.
+
+2. Help yourself by running,
+``` bash
+nox -s format
+```
+to take care of linting issues, with an automatic fix.
+
+3. Afterward, run 
+``` bash
+nox -s lint
+```
+to learn where manual fixes are required for style compatability.

README.rst

@@ -26,21 +26,27 @@ Pytorch Wavelet Toolbox (`ptwt`)
     :target: https://github.com/psf/black
     :alt: Black code style
 
+.. image:: https://static.pepy.tech/personalized-badge/ptwt?period=total&units=international_system&left_color=grey&right_color=brightgreen&left_text=Downloads
+ :target: https://pepy.tech/project/ptwt
+
+
 
 
 Welcome to the PyTorch wavelet toolbox. This package implements:
 
 - the fast wavelet transform (fwt) via ``wavedec`` and its inverse by providing the ``waverec`` function,
 - the two-dimensional fwt is called ``wavedec2`` the synthesis counterpart ``waverec2``,
 - ``wavedec3`` and ``waverec3`` cover the three-dimensional analysis and synthesis case,
-- ``MatrixWavedec`` and ``MatrixWaverec`` provide sparse-matrix-based fast wavelet transforms with boundary filters,
-- 2d sparse-matrix transforms with separable & non-separable boundary filters are available (experimental),
+- ``fswavedec2``, ``fswavedec3``, ``fswaverec2`` and ``fswaverec3`` support separable transformations.
+- ``MatrixWavedec`` and ``MatrixWaverec`` implement sparse-matrix-based fast wavelet transforms with boundary filters,
+- 2d sparse-matrix transforms with separable & non-separable boundary filters are available,
+- ``MatrixWavedec3`` and ``MatrixWaverec3`` allow separable 3D-fwt's with boundary filters.
 - ``cwt`` computes a one-dimensional continuous forward transform,
 - single and two-dimensional wavelet packet forward and backward transforms are available via the ``WaveletPacket`` and ``WaveletPacket2D`` objects,
 - finally, this package provides adaptive wavelet support (experimental).
 
-This toolbox supports pywt-wavelets. Complete documentation is available:
-https://pytorch-wavelet-toolbox.readthedocs.io/
+This toolbox extends `PyWavelets <https://pywavelets.readthedocs.io/en/latest/>`_ . We additionally provide GPU and gradient support via a PyTorch backend.
+Complete documentation is available at: https://pytorch-wavelet-toolbox.readthedocs.io/
 
 
 **Installation**
@@ -101,7 +107,7 @@ To test an example run:
   import numpy as np
   import scipy.misc
 
-  face = np.transpose(scipy.misc.face(),
+  face = np.transpose(scipy.datasets.face(),
                           [2, 0, 1]).astype(np.float64)
   pytorch_face = torch.tensor(face)
   coefficients = ptwt.wavedec2(pytorch_face, pywt.Wavelet("haar"),
@@ -143,8 +149,8 @@ Reconsidering the 1d case, try:
 
 
 The process for the 2d transforms ``MatrixWavedec2``, ``MatrixWaverec2`` works similarly.
-By default, a non-separable transformation is used.
-To use a separable transformation, pass ``separable=True`` to ``MatrixWavedec2`` and ``MatrixWaverec2``.
+By default, a separable transformation is used.
+To use a non-separable transformation, pass ``separable=False`` to ``MatrixWavedec2`` and ``MatrixWaverec2``.
 Separable transformations use a 1d transformation along both axes, which might be faster since fewer matrix entries
 have to be orthogonalized.
 
@@ -162,15 +168,18 @@ See https://github.com/v0lta/PyTorch-Wavelet-Toolbox/tree/main/examples/network_
 
 **Testing**
 
-The ``tests`` folder contains multiple tests to allow independent verification of this toolbox. After cloning the
-repository, and moving into the main directory, and installing ``nox`` with ``pip install nox`` run:
+The ``tests`` folder contains multiple tests to allow independent verification of this toolbox.
+The GitHub workflow executes a subset of all tests for efficiency reasons. 
+After cloning the repository, moving into the main directory, and installing ``nox`` with ``pip install nox`` run
 
 .. code-block:: sh
 
   $ nox --session test
 
 
 
+to run all existing tests.
+
 Citation
 """"""""
 

docs/ptwt.rst

@@ -27,6 +27,13 @@ ptwt.conv\_transform\_3 module
    :undoc-members:
    :show-inheritance:
 
+ptwt.separable\_conv\_transform module
+--------------------------------------
+
+.. automodule:: ptwt.separable_conv_transform
+   :members:
+   :undoc-members:
+   :show-inheritance:
 
 ptwt.continuous\_transform module
 ---------------------------------
@@ -60,6 +67,15 @@ ptwt.matmul\_transform\_2 module
    :undoc-members:
    :show-inheritance:
 
+ptwt.matmul\_transform\_3 module
+--------------------------------
+
+.. automodule:: ptwt.matmul_transform_3
+   :members:
+   :undoc-members:
+   :show-inheritance:
+
+
 ptwt.sparse\_math module
 ------------------------
 

examples/deepfake_analysis/README.md

@@ -25,7 +25,13 @@ Recreating these experiments requires roughly 8GB of free disc space and 10GB of
 To reproduce these plots:
 1. Download [ffhq_style_gan.zip](https://drive.google.com/uc?id=1MOHKuEVqURfCKAN9dwp1o2tuR19OTQCF&export=download) and
 2. Extract the image pairs here.
-3. Run `python packet_plot.py`
+3. Check the file structure. In `ffhq_style_gan` the folder structure should be:
+```
+source_data
+ ├── A_ffhq
+ ├── B_stylegan
+```
+4. Run `python packet_plot.py`
 
 You can read more about GAN-detection in the paper [Wavelet-Packets for Deepfake Image Analysis and Detection](https://rdcu.be/cUIRt). 
 A complete project building gan-detectors on top of wavelets is available at:

examples/wavelet_packet_chirp_analysis/chirp.png

@@ -7,27 +7,40 @@
 # use from src.ptwt.packets if you cloned the repo instead of using pip.
 from ptwt import WaveletPacket
 
-t = np.linspace(0, 10, 1500)
-w = scipy.signal.chirp(t, f0=1, f1=50, t1=10, method="linear")
+fs = 1000
+t = np.linspace(0, 2, int(2//(1/fs)))
+w = np.sin(256*np.pi*t**2)
 
-wavelet = pywt.Wavelet("db3")
+wavelet = pywt.Wavelet("sym8")
 wp = WaveletPacket(
-    data=torch.from_numpy(w.astype(np.float32)), wavelet=wavelet, mode="reflect"
+    data=torch.from_numpy(w.astype(np.float32)), wavelet=wavelet, mode="boundary"
 )
-nodes = wp.get_level(5)
+level = 5
+nodes = wp.get_level(level)
 np_lst = []
 for node in nodes:
     np_lst.append(wp[node])
 viz = np.stack(np_lst).squeeze()
 
+n = list(range(int(np.power(2, level))))
+freqs = (fs/2)*(n/(np.power(2, level)))
+
+xticks = list(range(viz.shape[-1]))[::6]
+xlabels = np.round(np.linspace(min(t), max(t), viz.shape[-1]), 2)[::6]
+
 fig, axs = plt.subplots(2)
 axs[0].plot(t, w)
-axs[0].set_title("Linear Chirp, f(0)=1, f(10)=50")
-axs[0].set_xlabel("t [s]")
+axs[0].set_title("Analyzed signal")
+axs[0].set_xlabel("time [s]")
+axs[0].set_ylabel("magnitude")
 
-axs[1].set_title("Wavelet analysis")
+axs[1].set_title("Wavelet packet analysis")
 axs[1].imshow(np.abs(viz))
-axs[1].set_xlabel("time")
-axs[1].set_ylabel("frequency")
+axs[1].set_xlabel("time [s]")
+axs[1].set_xticks(xticks)
+axs[1].set_xticklabels(xlabels)
+axs[1].set_ylabel("frequency [Hz]")
+axs[1].set_yticks(n[::4])
+axs[1].set_yticklabels(freqs[::4])
 axs[1].invert_yaxis()
 plt.show()

examples/wavelet_packet_chirp_analysis/chirp_analysis.py

@@ -21,9 +21,8 @@ def run_test_fast(session):
 @nox.session(name="lint")
 def lint(session):
     """Check code conventions."""
-    session.install("flake8==4.0.1")
+    session.install("flake8")
     session.install(
-        "flake8-colors",
         "flake8-black",
         "flake8-docstrings",
         "flake8-bugbear",
@@ -32,7 +31,6 @@ def lint(session):
         "pydocstyle",
         "darglint",
     )
-    session.install("flake8-bandit==2.1.2", "bandit==1.7.2")
     session.run("flake8", "src", "tests", "noxfile.py")
 
 
@@ -48,8 +46,7 @@ def mypy(session):
         "--ignore-missing-imports",
         "--strict",
         "--no-warn-return-any",
-        "--implicit-reexport",
-        "--allow-untyped-calls",
+        "--explicit-package-bases",
         "src",
     )
 

noxfile.py

@@ -41,9 +41,12 @@ classifiers =
 install_requires =
     PyWavelets
     torch
-    scipy
+    scipy>=1.10
+    pooch
     matplotlib
     numpy
+    pytest
+    nox
 
 packages = find:
 package_dir =
@@ -58,4 +61,4 @@ where = src
 ##########################
 [darglint]
 docstring_style = google
-strictness = short
+strictness = short