evograd-diff 0.1.0__py3-none-any.whl

evograd_diff-0.1.0.dist-info/METADATA

@@ -0,0 +1,439 @@
+Metadata-Version: 2.4
+Name: evograd-diff
+Version: 0.1.0
+Summary: PyTorch-based framework for differentiable evolutionary computation and swarm intelligence
+Project-URL: Homepage, https://github.com/andreatangherloni/EvoGrad
+Project-URL: Repository, https://github.com/andreatangherloni/EvoGrad
+Project-URL: Issues, https://github.com/andreatangherloni/EvoGrad/issues
+Author-email: Andrea Tangherloni <andrea.tangherloni@unibocconi.it>
+License: Apache-2.0
+License-File: LICENSE
+Keywords: CMA-ES,differentiable programming,differential evolution,evolutionary computation,genetic algorithm,metaheuristics,optimization,particle swarm optimization,pytorch
+Classifier: Development Status :: 3 - Alpha
+Classifier: Intended Audience :: Science/Research
+Classifier: License :: OSI Approved :: Apache Software License
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
+Classifier: Topic :: Scientific/Engineering :: Mathematics
+Requires-Python: >=3.12
+Requires-Dist: matplotlib>=3.10.8
+Requires-Dist: numpy>=2.4.4
+Requires-Dist: pandas>=3.0.2
+Requires-Dist: scipy>=1.17.1
+Requires-Dist: seaborn>=0.13.2
+Requires-Dist: torch>=2.11.0
+Description-Content-Type: text/markdown
+
+# EvoGrad: Metaheuristics in a Differentiable Wonderland
+
+<p align="center">
+  <img src="https://img.shields.io/badge/python-3.9+-blue.svg" alt="Python 3.9+">
+  <img src="https://img.shields.io/badge/pytorch-2.0+-orange.svg" alt="PyTorch 2.0+">
+  <img src="https://img.shields.io/badge/license-Apache%20License%202.0-blue.svg" alt="License: Apache-2.0">
+ 
+</p>
+
+**EvoGrad** is a PyTorch-based framework for differentiable Evolutionary Computation and Swarm Intelligence. It bridges classical population-based optimisation with modern differentiable programming by enabling gradient flow through evolutionary operators.
+
+## 🌟 Key Features
+
+- **Fully Differentiable**: All operators support gradient computation via reparameterisation tricks (Gumbel-Softmax, Binary-Concrete, pathwise gradients)
+- **GPU Accelerated**: Native PyTorch implementation for seamless CPU/GPU/MPS execution
+- **Modular Design**: Dependency injection pattern inspired by [pymoo](https://pymoo.org/) for flexible operator composition
+- **Learnable Hyperparameters**: Automatically tune algorithm parameters via backpropagation
+- **Four Algorithms**: GA, DE, PSO, and CMA-ES with multiple variants
+
+## 📦 Installation
+
+```bash
+# Clone the repository
+git clone https://github.com/YOUR_USERNAME/evograd.git
+cd evograd
+
+# Install dependencies
+pip install torch numpy
+
+# Install in development mode
+pip install -e .
+```
+
+## 🚀 Quick Start
+
+```python
+import torch
+from evograd.core import Problem, minimize
+from evograd.algorithms import GA, DE, PSO, CMAES
+
+# Define an optimisation problem
+problem = Problem(
+    objective=lambda x: (x**2).sum(dim=-1),  # Sphere function
+    n_var=30,
+    xl=-100.0,
+    xu=100.0,
+)
+
+# Run with Genetic Algorithm
+ga = GA(pop_size=100, differentiable=True)
+result = minimize(problem, ga, max_evals=10000, seed=42)
+print(f"GA Best: {result.best_fitness:.6f}")
+
+# Run with Differential Evolution
+de = DE(pop_size=100, variant="DE/rand/1/bin", adaptive=True)
+result = minimize(problem, de, max_evals=10000, seed=42)
+print(f"DE Best: {result.best_fitness:.6f}")
+
+# Run with Particle Swarm Optimisation
+pso = PSO(pop_size=100, adaptive=True, differentiable=True)
+result = minimize(problem, pso, max_evals=10000, seed=42)
+print(f"PSO Best: {result.best_fitness:.6f}")
+
+# Run with CMA-ES
+cmaes = CMAES(sigma=0.5, adaptive=True)
+result = minimize(problem, cmaes, max_evals=10000, seed=42)
+print(f"CMA-ES Best: {result.best_fitness:.6f}")
+```
+
+## 🔧 Algorithms and Operating Modes
+
+### Genetic Algorithm (GA)
+
+The GA uses **operator-level differentiability**. Each operator (selection, crossover, mutation, survival) can independently be set to differentiable mode:
+
+```python
+from evograd.algorithms import GA
+from evograd.operators import (
+    RouletteSelection,
+    SBXCrossover,
+    PolynomialMutation,
+    MergeSurvival,
+)
+
+# Classical GA (no gradients)
+ga = GA(pop_size=100, differentiable=False)
+
+# Fully differentiable GA with custom operators
+ga = GA(
+    pop_size=100,
+    selection=RouletteSelection(differentiable=True, learn_temperature=True),
+    crossover=SBXCrossover(differentiable=True, learn_eta=True, learn_prob=True),
+    mutation=PolynomialMutation(differentiable=True, learn_eta=True, learn_prob=True),
+    survival=MergeSurvival(selection=RouletteSelection(differentiable=True)),
+    differentiable=True,  # Makes population learnable
+)
+```
+
+| Parameter | Effect |
+|-----------|--------|
+| `differentiable=False` | Classical GA with discrete operators |
+| `differentiable=True` | Population is an `nn.Parameter` (learnable via backprop) |
+| Operator `differentiable=True` | Operator uses Gumbel-Softmax/Binary-Concrete for gradient flow |
+| Operator `learn_*=True` | Operator hyperparameters become learnable `nn.Parameter` |
+
+### Differential Evolution (DE)
+
+DE uses **algorithm-level flags** for adaptive hyperparameters and differentiable population:
+
+```python
+from evograd.algorithms import DE, de_rand_1_bin, de_best_1_bin
+
+# Classical DE
+de = DE(pop_size=100, variant="DE/rand/1/bin", F=0.5, CR=0.9)
+
+# Adaptive DE (learnable F, CR, selection temperature)
+de = DE(pop_size=100, variant="DE/best/1/bin", adaptive=True)
+
+# Differentiable population
+de = DE(pop_size=100, variant="DE/rand/1/bin", differentiable=True)
+
+# Both adaptive and differentiable
+de = DE(pop_size=100, variant="DE/current-to-best/1/bin", adaptive=True, differentiable=True)
+```
+
+| `adaptive` | `differentiable` | Effect |
+|------------|------------------|--------|
+| False | False | Classical DE |
+| True | False | F, CR, temperatures learnable via backprop |
+| False | True | Population learnable via backprop |
+| True | True | Both hyperparameters and population learnable |
+
+**Supported Variants:**
+- `DE/rand/1/bin`, `DE/rand/1/exp`, `DE/rand/2/bin`, `DE/rand/2/exp`
+- `DE/best/1/bin`, `DE/best/1/exp`, `DE/best/2/bin`, `DE/best/2/exp`
+- `DE/current-to-best/1/bin`, `DE/current-to-best/1/exp`
+- `DE/current-to-rand/1`
+
+### Particle Swarm Optimisation (PSO)
+
+PSO uses the same **algorithm-level flags** as DE:
+
+```python
+from evograd.algorithms import PSO, pso_constriction, pso_adaptive
+
+# Classical PSO
+pso = PSO(pop_size=100, inertia=0.7, c1=1.5, c2=1.5)
+
+# Adaptive PSO (learnable inertia, c1, c2)
+pso = PSO(pop_size=100, adaptive=True)
+
+# Per-particle adaptive coefficients
+pso = PSO(pop_size=100, adaptive=True, per_particle_coeffs=True)
+
+# Constriction factor PSO
+pso = pso_constriction(pop_size=100)
+
+# Fully differentiable
+pso = PSO(pop_size=100, adaptive=True, differentiable=True)
+```
+
+| `adaptive` | `differentiable` | Effect |
+|------------|------------------|--------|
+| False | False | Classical PSO |
+| True | False | Inertia, c1, c2 learnable via backprop |
+| False | True | Particle positions learnable via backprop |
+| True | True | Both coefficients and positions learnable |
+
+### CMA-ES
+
+CMA-ES supports **adaptive coefficients** and **restart strategies** (IPOP/BIPOP):
+
+```python
+from evograd.algorithms import CMAES, cmaes_ipop, cmaes_bipop
+
+# Classical CMA-ES
+cmaes = CMAES(pop_size=50, sigma=0.5)
+
+# Adaptive CMA-ES (learnable cc, cs, c1, cmu, damps)
+cmaes = CMAES(pop_size=50, sigma=0.5, adaptive=True)
+
+# Differentiable mean
+cmaes = CMAES(pop_size=50, sigma=0.5, differentiable=True)
+
+# IPOP-CMA-ES (increasing population restarts)
+cmaes = cmaes_ipop(restarts=9, incpopsize=2)
+
+# BIPOP-CMA-ES (alternating small/large populations)
+cmaes = cmaes_bipop(restarts=9)
+```
+
+| `adaptive` | `differentiable` | Effect |
+|------------|------------------|--------|
+| False | False | Classical CMA-ES |
+| True | False | Adaptation coefficients learnable via backprop |
+| False | True | Distribution mean μ learnable via backprop |
+| True | True | Both coefficients and mean learnable |
+
+**Restart Strategies:**
+- **IPOP**: Restart with doubled population after convergence
+- **BIPOP**: Alternate between small (focused) and large (exploratory) populations
+
+## 📚 Operators Library
+
+EvoGrad provides a comprehensive library of evolutionary operators:
+
+### Selection
+| Operator | Description | Differentiable |
+|----------|-------------|----------------|
+| `RandomSelection` | Uniform random selection | ✗ |
+| `RouletteSelection` | Fitness-proportionate (Gumbel-Softmax) | ✓ |
+| `TournamentSelection` | Tournament with soft winner (Gumbel-Softmax) | ✓ |
+| `RankSelection` | Rank-based probabilities | ✓ |
+| `StochasticUniversalSampling` | SUS with soft selection | ✓ |
+
+### Crossover
+| Operator | Description | Differentiable |
+|----------|-------------|----------------|
+| `SBXCrossover` | Simulated Binary Crossover | ✓ |
+| `BinomialCrossover` | DE-style binomial | ✓ |
+| `ExponentialCrossover` | DE-style exponential | ✓ |
+| `BlendCrossover` | BLX-α crossover | ✓ |
+| `ArithmeticCrossover` | Weighted average | ✓ |
+| `UniformCrossover` | Gene-wise uniform swap | ✓ |
+| `SimulatedBinaryCrossover` | Alias for SBX | ✓ |
+
+### Mutation
+| Operator | Description | Differentiable |
+|----------|-------------|----------------|
+| `PolynomialMutation` | Polynomial bounded mutation | ✓ |
+| `GaussianMutation` | Additive Gaussian noise | ✓ |
+| `UniformMutation` | Uniform random replacement | ✓ |
+| `AdaptiveMutation` | Self-adaptive mutation rates | ✓ |
+
+### Survival
+| Operator | Description |
+|----------|-------------|
+| `MergeSurvival` | (μ+λ) with optional elitism |
+| `ReplacementSurvival` | (μ,λ) generational replacement |
+| `AgingSurvival` | Age-based replacement |
+| `FitnessSurvival` | Pure fitness-based truncation |
+
+### Repair
+| Operator | Description |
+|----------|-------------|
+| `BoundsRepair` | Clamp to bounds |
+| `ReflectionRepair` | Bounce off boundaries |
+| `WrapRepair` | Toroidal wrap-around |
+| `RandomRepair` | Random resampling |
+
+## 🎯 Advanced Usage
+
+### Training Neural Networks with EvoGrad
+
+```python
+import torch
+import torch.nn as nn
+from evograd.algorithms import CMAES
+from evograd.core import Problem
+from evograd.core.termination import MaxEvaluations
+
+
+# Define a simple MLP
+class MLP(nn.Module):
+    def __init__(self):
+        super().__init__()
+        self.net = nn.Sequential(
+            nn.Linear(10, 64),
+            nn.Tanh(),
+            nn.Linear(64, 1),
+        )
+    
+    def forward(self, x):
+        return self.net(x)
+
+# Flatten parameters for optimisation
+model = MLP()
+n_params = sum(p.numel() for p in model.parameters())
+
+def loss_fn(params):
+    # Reshape flat params back to model
+    idx = 0
+    for p in model.parameters():
+        numel = p.numel()
+        p.data.copy_(params[idx:idx+numel].view(p.shape))
+        idx += numel
+    
+    # Compute loss on dummy data
+    x = torch.randn(32, 10)
+    y = torch.randn(32, 1)
+    pred = model(x)
+    return ((pred - y)**2).mean()
+
+# Batch evaluation
+def batch_loss(pop):
+    return torch.stack([loss_fn(p) for p in pop])
+
+problem = Problem(
+    objective=batch_loss,
+    n_var=n_params,
+    xl=-10.0,
+    xu=10.0,
+)
+
+cmaes = CMAES(pop_size=50, sigma=1.0, adaptive=True)
+result = minimize(problem, cmaes, MaxEvaluations(10000))
+print(f"Final loss: {result.best_fitness:.6f}")
+```
+
+### Callbacks for Logging
+
+```python
+from evograd.core import minimize
+from evograd.utils import HistoryCallback, PrintCallback
+
+callbacks = [
+    PrintCallback(every=10),  # Print progress every 10 generations
+    HistoryCallback(),        # Record full history
+]
+
+result = minimize(problem, algorithm, termination=MaxEvaluations(10000), callback=callbacks)
+
+# Access history
+print(f"Fitness over time: {result.history['best_fitness']}")
+```
+
+## 🏗️ Architecture
+
+```
+evograd/
+├── algorithms/
+│   └── cmaes.py          # CMA-ES with IPOP/BIPOP
+│   ├── de.py             # Differential Evolution
+│   ├── ga.py             # Genetic Algorithm
+│   ├── pso.py            # Particle Swarm Optimisation
+├── core/
+│   ├── algorithm.py      # Base Algorithm class
+│   ├── maximize.py       # Optimisation loop (maximisation)
+│   ├── minimize.py       # Optimisation loop (minimisation)
+│   ├── problem.py        # Problem definition
+│   ├── result.py         # Result container
+│   └── termination.py    # Stopping criteria
+├── operators/
+│   ├── crossover.py      # Crossover operators
+│   ├── mutation.py       # Mutation operators
+│   ├── sampling.py       # Sampling operators
+│   ├── selection.py      # Selection operators
+│   ├── survival.py       # Survival/replacement
+│   └── repair.py         # Constraint handling
+└── utils/
+    ├── callbacks.py      # Logging utilities
+    ├── device.py         # Device management
+    └── duplicates.py     # Duplicate elimination
+```
+
+## 🔬 How It Works
+
+EvoGrad makes evolutionary algorithms differentiable through:
+
+1. **Reparameterisation Trick**: Convert random sampling into deterministic transformations of parameter-free noise:
+   ```
+   x = g_θ(ε), ε ~ p(ε)  →  ∇_θ L ≈ ∇_θ f(g_θ(ε))
+   ```
+
+2. **Gumbel-Softmax**: Differentiable approximation for categorical selection:
+   ```python
+   # Soft selection (differentiable)
+   probs = softmax((log_probs + gumbel_noise) / temperature)
+   selected = probs @ population  # Weighted combination
+   ```
+
+3. **Binary-Concrete**: Differentiable approximation for binary masks (mutation/crossover):
+   ```python
+   # Soft mask (differentiable)
+   mask = sigmoid((log(u) - log(1-u) + logits) / temperature)
+   # Straight-through estimator for hard decisions
+   hard_mask = (mask > 0.5).float() - mask.detach() + mask
+   ```
+
+4. **Pathwise Gradients**: For continuous distributions (Gaussian sampling in CMA-ES):
+   ```python
+   # x = μ + σ * L @ z, z ~ N(0, I)
+   z = torch.randn(pop_size, n_var)
+   x = mean + sigma * (L @ z.T).T  # Fully differentiable
+   ```
+
+## 📊 Benchmarks
+
+TODO
+
+## 📖 Citation
+
+TBA
+
+## 📄 License
+
+This project is licensed under the Apache-2.0 License - see the [LICENSE](LICENSE) file for details.
+
+## 🤝 Contributing
+
+Contributions are welcome! Please feel free to submit a Pull Request. For major changes, please open an issue first to discuss what you would like to change.
+
+1. Fork the repository
+2. Create your feature branch (`git checkout -b feature/amazing-feature`)
+3. Commit your changes (`git commit -m 'Add amazing feature'`)
+4. Push to the branch (`git push origin feature/amazing-feature`)
+5. Open a Pull Request
+
+## 🙏 Acknowledgements
+
+- Inspired by [pymoo](https://pymoo.org/) for API design
+- Built with [PyTorch](https://pytorch.org/) for automatic differentiation

evograd_diff-0.1.0.dist-info/RECORD

@@ -0,0 +1,50 @@
+evograd/__init__.py,sha256=OVdolqUOenKujo0OUIL4Q0mNJCJqkyNRT9OWqJsh5kU,2141
+evograd/algorithms/__init__.py,sha256=tckBjddiPRHdTJIujb5q5ptWhKCCfHDeN267_B7QX-c,3190
+evograd/algorithms/cmaes.py,sha256=ToWdHWhm0uD71C8m1TQE4NtcT6JR9Hwf_I1-v63OO9s,47435
+evograd/algorithms/de.py,sha256=eGv0pEOOtGWXG-e4BJNY1Ed4rRC8MwWd-FoPeQuslQc,32996
+evograd/algorithms/ga.py,sha256=d_6DAjW7NPNefJWeoaQZY1Op74ijLZENzv51P2wPYfU,17646
+evograd/algorithms/pso.py,sha256=AKq8vT6Zf7DM0HQnQtksfMeJwwH5fOzFbdJsQXvDvyo,21567
+evograd/algorithms/shade.py,sha256=yEoYw0VgCB4uXlgshjTJdBBAb31DIjP74rvuswerl3w,41213
+evograd/benchmarks/run_benchmark_functions.py,sha256=zAo3lFGj5wgRxHDl-1HH4RznmQoPmS9a8rmKgyotdSA,44602
+evograd/benchmarks/functions/__init__.py,sha256=ZyMJwNUBJ9Adkt8uJU4lM2yYLJ3k-grwaqOOxkdOFTk,5046
+evograd/benchmarks/functions/base.py,sha256=AdGSBG3ZnKjuikVtLmqLKD6DSYiqqtVXbBRTtuMcpi8,6797
+evograd/benchmarks/functions/classical.py,sha256=GlfQ7OdprmSlo2yT0_kIEe3-iqJYofufI_pi7z4AaH0,17782
+evograd/benchmarks/functions/smoothed_funnel.py,sha256=RYRm3G0_f34jBUP5kfLms_WlVJkc7PDOUzTBT3dje9g,15635
+evograd/benchmarks/functions/transforms.py,sha256=yQIKMnXGgWa7eVoueT4GLLc8cfF8AqGsNb6fuxC5ZT4,13744
+evograd/benchmarks/functions/cec2017/__init__.py,sha256=XMtpn1fN-8VQUzeasCrSv5X25N-oLkgwnObeuWx5Bkg,5840
+evograd/benchmarks/functions/cec2017/basic.py,sha256=HHkHUuGxI7QQMMpCCuXol2Rjbzk-CGHBo2QyE_-h4KI,10962
+evograd/benchmarks/functions/cec2017/composition.py,sha256=OikeBgiJvGMf0nbj4YkJp9pPEfwjqVGZxWcZR8x_SoU,18161
+evograd/benchmarks/functions/cec2017/data.py,sha256=SSdacN0qIGaoUrOhtZK4pzcM0vs--24TaiS0C-tyeBs,10372
+evograd/benchmarks/functions/cec2017/hybrid.py,sha256=MpudxERgGc8R2SoYmgkMhv5zndjdVwzJyYI2hZ7Kb40,11783
+evograd/benchmarks/functions/cec2017/simple.py,sha256=xp1x22p02JRRdihZxzGJM-ClRPDGyVgMkppZ9ZwrP90,9272
+evograd/core/__init__.py,sha256=CZkCkmZJp8W3-b60g0IgiWDX1a5-DtFGhpXmJiWUOXA,1877
+evograd/core/algorithm.py,sha256=uK_WhcuXNisq4eUojimIo07nwDDOdQAIZE5pYPCoNRA,28007
+evograd/core/maximize.py,sha256=x39c2Fwot89AJkvOyxIaKwZuGZhjSHgmUUMdpOOT5TE,9754
+evograd/core/minimize.py,sha256=hNmn-SElGucDYHkNpqR6u4keaq3SLlRqbDtqBsVk_Pk,26954
+evograd/core/problem.py,sha256=GYINuJfdR4ZjXQAV_rUKhg8jbusmjv3qVkPxAVUnlIQ,14858
+evograd/core/result.py,sha256=LOyaar71CN4ma4-Q5Sb9HgyyIq7uDzsHWO-4IZTEH4w,18819
+evograd/core/termination.py,sha256=WzNrSyjxEEWl3IxUyKlRhd_J7WvcbMv3TlKMLHant3Q,18862
+evograd/operators/__init__.py,sha256=lGqBT-lE0KRrZ3ID6astqdIQ09xWA_rPGDRer-A1m2w,4626
+evograd/operators/crossover.py,sha256=wCzp0eFYv_ZhGiPtsEEf5st6P_ADLGuIwEnLib3kJRs,37222
+evograd/operators/mutation.py,sha256=fdMob0pngQZWN3wfBe_GKn0r4jobHfBjq0hrVo0JUp4,36223
+evograd/operators/relaxations.py,sha256=r3ZDxD2YjR-Bire7bUdGz6EeNFUI4vFDK2GeP_VG6Do,5361
+evograd/operators/repair.py,sha256=OFfN9SWx6j9S1-R0QbpQ71yTskjNcj99QnNw5R7m0G0,16950
+evograd/operators/sampling.py,sha256=YEmM8XE-oyyImGvrhYHQ59assEvvJoZ0PuemsgMb8Hw,18865
+evograd/operators/selection.py,sha256=H38qFAYWcyz5iVc3sTv8vM7T5cxdTaWEfPRIXr4C0gM,32665
+evograd/operators/survival.py,sha256=r56ZctWSXUHHviXR-Gscev1mNOyWVW9R1iOPPRJN5W4,36437
+evograd/tests/__init__.py,sha256=uC0DSDJQ8vpVxa47IAB7nEOxq4BaJICMk3PXlND0VFY,198
+evograd/tests/run_all.py,sha256=QVMnQKc_trP-KSYfwCyYk8Gfi3sjzRXfcmdBK9XSZow,1947
+evograd/tests/test_core.py,sha256=Qe4zsOh_KaF0fBSsqNc7igp_pI3movN9egIwWM70xrw,16871
+evograd/tests/test_ga.py,sha256=5ipLeXk_bI42iz0SuZ1604xmnlJmv4YRpdyAepzzDDM,17157
+evograd/tests/test_operators.py,sha256=E5oRZXGU5L7zFDiJ5vUc-va9UoIQ1qMJITqfJY0VTzI,21993
+evograd/tests/test_per_individual.py,sha256=JRVis_dI3uuNHMGL14Vt-E-i6IOOBnOzR7lS0ztHawY,12907
+evograd/tests/test_utils.py,sha256=SYUBt1SXZwNQXXXxIXHmC0ZSngbByHsDIRoGiouBPA4,10167
+evograd/utils/__init__.py,sha256=DdQY1bRvpCJ-04fSa6GHK5p0LvFYXrCBfLb6zsakzIQ,2124
+evograd/utils/callbacks.py,sha256=wMytUrw2QTJ_Xv0mHE56e-rZ46tKqYLxSO_xse5BzIQ,33302
+evograd/utils/device.py,sha256=jESrKGJSVFu_uxNTuzpRSwTP0rNbncUZsxqH1zjscSw,14323
+evograd/utils/duplicates.py,sha256=uDEZG9u-_XUrswvhRwwULQrlFd7ENgEL-zIJORCfMXA,13305
+evograd/benchmarks/functions/cec2017/data.pkl,sha256=X5eOxmIOZZ6pM-IcTEOYtHuBFGfT7tPcsgRWbDllkks,24612828
+evograd_diff-0.1.0.dist-info/METADATA,sha256=JfwSzpcKUDrzuZfgOT--T_Yl8FqHbTJlzyCnyJU1_AM,14831
+evograd_diff-0.1.0.dist-info/WHEEL,sha256=QccIxa26bgl1E6uMy58deGWi-0aeIkkangHcxk2kWfw,87
+evograd_diff-0.1.0.dist-info/licenses/LICENSE,sha256=xx0jnfkXJvxRnG63LTGOxlggYnIysveWIZ6H3PNdCrQ,11357
+evograd_diff-0.1.0.dist-info/RECORD,,

evograd_diff-0.1.0.dist-info/WHEEL

@@ -0,0 +1,4 @@
+Wheel-Version: 1.0
+Generator: hatchling 1.29.0
+Root-Is-Purelib: true
+Tag: py3-none-any