lint

Author: facundoq

.github/workflows/pipy_uv.yml .github/workflows/pipy.yml.github/workflows/pipy_uv.yml renamed to .github/workflows/pipy.yml

@@ -1,4 +1,4 @@
-name: Publish Python Package
+name: PiPy
 
 on:
   release:

.github/workflows/lint_test.yml .github/workflows/tests.yml.github/workflows/lint_test.yml renamed to .github/workflows/tests.yml

@@ -1,4 +1,4 @@
-name: Linter And Tests
+name: tests
 
 on: [push, pull_request]
 

.gitignore

@@ -97,6 +97,10 @@ venv.bak/
 # Rope project settings
 .ropeproject
 
+
+
+
+
 # mkdocs documentation
 /site
 
@@ -107,5 +111,5 @@ venv.bak/
 .idea/
 .vscode/
 
-
-
+# ruff
+.ruff_cache/

README.md

@@ -1,7 +1,9 @@
 # ✴ Transformational Measures 📏
 
 
-![](https://img.shields.io/pypi/v/tmeasures) | ![](https://img.shields.io/codecov/c/github/facundoq/tmeasures) | ![GitHub Actions Workflow Status](https://img.shields.io/github/actions/workflow/status/facundoq/tmeasures/pipy_uv.yml) | [![Tests](https://github.com/facundoq/tmeasures/actions/workflows/lint_test.yml/badge.svg)](https://github.com/facundoq/tmeasures/actions/workflows/lint_test.yml)
+![](https://img.shields.io/pypi/v/tmeasures) 
+| ![](https://img.shields.io/codecov/c/github/facundoq/tmeasures) 
+| [![Tests](https://github.com/facundoq/tmeasures/actions/workflows/lint_test.yml/badge.svg)](https://github.com/facundoq/tmeasures/actions/workflows/lint_test.yml)
 
 
 The Transformational Measures (`tmeasures`) library allows neural network designers to evaluate the invariance, equivariance and other properties of their models with respect to a set of transformations. Support for Pytorch. 

pyproject.toml

@@ -131,6 +131,7 @@ executor.type = "uv"
 
 [tool.poe.tasks]
 test = "uv run pytest"
+cov-test = "uv run pytest --cov=tmeasures --cov-report=xml"
 lint = "uvx ruff check --fix tmeasures "
 format = "uvx ruff format tmeasures"
 build = "uv build"

readme.rst

@@ -9,7 +9,6 @@
 
 class ActivationsIterator(abc.ABC):
     """
-    .. 
         Iterate over the ST matrix of activations of a network, varying the samples and transformations.
         The iteration is in both orders:
         1) transformations_first:  column indexes first, then row indices indices

setup.py.old

@@ -62,7 +62,6 @@ def __init__(self, id:int, name:str, threshold:float,sign:int):
     def eval(self,q:Queue,inner_q:Queue):
         running_mean = RunningMeanWelford()
         # activation_sum=0
-        n=0
         for transformation in self.queue_as_generator(q):
             for activations in self.queue_as_generator(inner_q):
                 if self.sign != 1:

tmeasures/np/activations_iterator.py

@@ -39,7 +39,8 @@ def eval(self,activations_iterator: ActivationsIterator,verbose=False)->MeasureR
         for transformation, samples_activations_iterator in tqdm(activations_iterator.transformations_first(),disable=not verbose):
             for x, batch_activations in samples_activations_iterator:
                 for j, activations in enumerate(batch_activations):
-                    if self.sign != 1: activations *= self.sign
+                    if self.sign != 1:
+                        activations *= self.sign
                     running_means[j].update_all(activations)
 
         stds  = [running_mean.std() for running_mean in running_means]