actions

README.md

@@ -1,17 +1,14 @@
 # Testing pydantic v2 
 
-This repository contains a [talk](https://sunnivin.github.io/a-new-core/talk/slides.html) and some demos (`demo\v1\`, `demo\v2\`) for exploring the new features of Pydantic v2. 
+This repository contains a [talk](https://sunnivin.github.io/a-new-core/talk/slides.html) and a demo (`demo`)for exploring optimization in Python presented at the Oslo Python MeetUp 24.10.2024.
 
-Compare the how to use v2 by looking at the different scripts with the same names in folder `demo/v1` and `demo/v2`.
 
 
 # Credit 
 
 The material in the slides and demo are based on the content from the following sources: 
 
 - [Curve fitting with python](https://machinelearningmastery.com/curve-fitting-with-python/) (accessed 21.10.2024)
-- Terrence Dorsey and Samuel Colvin [Pydantic v2 Pre Release](https://docs.pydantic.dev/latest/blog/pydantic-v2-alpha/) (accessed 22.08.2023)
-- Prashanth Rao, The Data Quarry, [Obtain a 5x speedup for free by upgrading to Pydantic v2](https://thedataquarry.com/posts/why-pydantic-v2-matters/) (accessed 22.08.2023)
-- Yaakov Bressler in [Medium](https://blog.det.life/dont-write-another-line-of-code-until-you-see-these-pydantic-v2-breakthrough-features-5cdc65e6b448) (accessed 22.08.2023)
-- Samuel Colvin at [PyCon US 2023](https://www.youtube.com/watch?v=pWZw7hYoRVU) (accessed 22.08.2023)
-- Configuring [mypy with Pydantic](https://docs.pydantic.dev/latest/integrations/mypy/#configuring-the-plugin) (accessed 22.08.2023)
+- H.-P. Halvorsen [Optimization in python](https://www.halvorsen.blog/documents/programming/python/resources/powerpoints/Optimization%20in%20Python.pdf) (accessed 21.10.2024)
+- Documentation of [SciPy.Optimize](https://docs.scipy.org/doc/scipy/reference/generated/scipy.optimize.least_squares.html) (accessed 21.10.2024)
+- Documentation of [Lmfit](https://lmfit.github.io/lmfit-py/index.html) (accessed 21.10.2024)

demo/demo/example_with_bounds.py

@@ -0,0 +1,83 @@
+"""
+Fit Using Bounds
+================
+
+A major advantage of using lmfit is that one can specify boundaries on fitting
+parameters, even if the underlying algorithm in SciPy does not support this.
+For more information on how this is implemented, please refer to:
+https://lmfit.github.io/lmfit-py/bounds.html
+
+The example below shows how to set boundaries using the ``min`` and ``max``
+attributes to fitting parameters.
+
+"""
+
+import matplotlib.pyplot as plt
+from numpy import exp, linspace, pi, random, sign, sin
+
+from lmfit import create_params, minimize
+from lmfit.printfuncs import report_fit
+
+import matplotlib
+
+matplotlib.use("Agg")  # Use a non-interactive backend
+
+###############################################################################
+# create the 'true' Parameter values and residual function:
+p_true = create_params(amp=14.0, period=5.4321, shift=0.12345, decay=0.010)
+
+
+def residual(pars, x, data=None):
+    argu = (x * pars["decay"]) ** 2
+    shift = pars["shift"]
+    if abs(shift) > pi / 2:
+        shift = shift - sign(shift) * pi
+    model = pars["amp"] * sin(shift + x / pars["period"]) * exp(-argu)
+    if data is None:
+        return model
+    return model - data
+
+
+###############################################################################
+# Generate synthetic data and initialize fitting Parameters:
+random.seed(0)
+x = linspace(0, 250, 1500)
+noise = random.normal(scale=0.8, size=x.size)
+data = residual(p_true, x) + noise
+
+fit_params = create_params(
+    amp=dict(value=13, max=20, min=0),
+    period=dict(value=2, max=10),
+    shift=dict(value=0, max=pi / 2.0, min=-pi / 2.0),
+    decay=dict(value=0.02, max=0.1, min=0),
+)
+
+
+fit_params = create_params(
+    amp=dict(
+        value=13,
+    ),
+    period=dict(
+        value=2,
+    ),
+    shift=dict(
+        value=0,
+    ),
+    decay=dict(value=0.02),
+)
+
+###############################################################################
+# Perform the fit and show the results:
+out = minimize(residual, fit_params, args=(x,), kws={"data": data})
+fit = residual(out.params, x)
+
+###############################################################################
+report_fit(out, modelpars=p_true, correl_mode="table")
+
+###############################################################################
+plt.plot(x, data, "o", label="data")
+plt.plot(x, fit, label="best fit")
+plt.plot(x, p_true, label="initial guess")
+plt.legend()
+
+plt.savefig("with_bounds.png")

demo/demo/parameters_values_lmfit.py

@@ -0,0 +1,61 @@
+import numpy as np
+
+from lmfit import Minimizer, create_params, report_fit, minimize
+
+import matplotlib.pyplot as plt
+import matplotlib
+
+matplotlib.use("Agg")  # Use a non-interactive backend
+
+
+# create data to be fitted
+x = np.linspace(0, 15, 301)
+np.random.seed(2021)
+amp = 5.0
+omega = 2.0
+shift = -0.1
+decay = 0.025
+data = amp * np.sin(omega * x + shift) * np.exp(-x * x * decay) + np.random.normal(
+    size=x.size, scale=0.2
+)
+
+
+# define objective function: returns the array to be minimized
+def fcn2min(params, x, data):
+    """Model a decaying sine wave and subtract data."""
+    v = params.valuesdict()
+
+    model = v["amp"] * np.sin(x * v["omega"] + v["shift"]) * np.exp(-x * x * v["decay"])
+    return model - data
+
+
+# create a set of Parameters
+params = create_params(
+    amp=dict(value=10, min=0),
+    decay=0.1,
+    omega=3.0,
+    shift=dict(value=0.0, min=-np.pi / 2.0, max=np.pi / 2),
+)
+
+initial_model = (
+    params["amp"].value
+    * np.sin(x * params["omega"].value + params["shift"].value)
+    * np.exp(-x * x * params["decay"].value)
+)
+# do fit, here with the default leastsq algorithm
+minner = Minimizer(fcn2min, params, fcn_args=(x, data))
+result = minner.minimize(method="least_squares")
+
+# calculate final result
+final = data + result.residual
+
+# write error report
+report_fit(result)
+
+# try to plot results
+
+plt.plot(x, data, "+", label="data", color="red")
+plt.plot(x, initial_model, label="initial", linestyle="--", color="green")
+plt.plot(x, final, label="final", color="blue")
+plt.legend()
+plt.savefig("parameters_values_lmfit.png")

demo/demo/parameters_values_scipy_no_bounds.py

@@ -0,0 +1,46 @@
+import numpy as np
+import matplotlib.pyplot as plt
+from scipy.optimize import curve_fit
+import matplotlib
+
+matplotlib.use("Agg")  # Use a non-interactive backend
+
+# Create data to be fitted
+x = np.linspace(0, 15, 301)
+np.random.seed(2021)
+amp = 5.0
+omega = 2.0
+shift = -0.1
+decay = 0.025
+data = amp * np.sin(omega * x + shift) * np.exp(-x * x * decay) + np.random.normal(
+    size=x.size, scale=0.2
+)
+
+
+# Define the model function
+def model_func(x, amp, omega, shift, decay):
+    return amp * np.sin(omega * x + shift) * np.exp(-x * x * decay)
+
+
+# Initial guess for the parameters
+initial_guess = [10, 3.0, 0.0, 0.1]
+
+# Perform the curve fitting
+popt, pcov = curve_fit(model_func, x, data, p0=initial_guess)
+
+# Calculate the fitted data
+fitted_data = model_func(x, *popt)
+
+# Plot the results
+plt.figure(figsize=(10, 6))
+plt.plot(x, data, "+", label="Data", color="red")
+plt.plot(x, fitted_data, label="Fitted Curve", color="blue")
+plt.axhline(
+    0, color="gray", lw=0.5, ls="--"
+)  # Add a horizontal line at y=0 for reference
+plt.title("Curve Fitting with scipy.optimize.curve_fit")
+plt.xlabel("x")
+plt.ylabel("Data and Fitted Curve")
+plt.legend()
+plt.grid()
+plt.savefig("parameters_values_scipy.png")

demo/demo/rosenbrock.py

@@ -0,0 +1,74 @@
+import numpy as np
+import lmfit
+import matplotlib.pyplot as plt
+import matplotlib
+
+matplotlib.use("Agg")  # Use a non-interactive backend
+
+
+# Define Rosenbrock's Banana Function
+def rosenbrock(x, y, a=1, b=100):
+    return (a - x) ** 2 + b * (y - x**2) ** 2
+
+
+# Generate synthetic data points
+def generate_data(num_points=100, noise_level=0.1):
+    x_true = np.linspace(-2, 2, num_points)
+    y_true = x_true**2  # We expect y = x^2 for Rosenbrock
+    noise = np.random.normal(0, noise_level, size=y_true.shape)  # Add noise
+    y_observed = y_true + noise
+    return x_true, y_observed
+
+
+# Create a dataset
+x_data, y_data = generate_data(num_points=100, noise_level=1.0)
+
+
+# Define the objective function
+def objective(params):
+    # Get the parameters
+    a = params["a"]
+    b = params["b"]
+
+    # Model predictions
+    y_pred = a - x_data**2  # Compute y values based on Rosenbrock's function model
+    return y_data - y_pred  # Return residuals
+
+
+# Initial guess for the parameters
+initial_params = lmfit.Parameters()
+initial_params.add("a", value=1.0)  # Initial guess for parameter a
+initial_params.add("b", value=100.0)  # Initial guess for parameter b
+
+# Perform the optimization
+result = lmfit.minimize(objective, initial_params)
+
+# Print the results
+print(result.message)
+print(
+    f"Optimized parameters: a = {result.params['a'].value}, b = {result.params['b'].value}"
+)
+print(
+    f"Minimum value of Rosenbrock's function (sum of squares of residuals): {result.chisqr}"
+)
+
+# Optional: Plot the data and the fitted curve
+plt.figure(figsize=(10, 6))
+plt.scatter(x_data, y_data, label="Observed Data", color="red", alpha=0.6)
+plt.plot(
+    x_data,
+    rosenbrock(
+        result.params["a"].value,
+        x_data**2,
+        a=result.params["a"].value,
+        b=result.params["b"].value,
+    ),
+    label="Fitted Curve",
+    color="blue",
+)
+plt.title("Rosenbrock's Banana Function Optimization")
+plt.xlabel("x")
+plt.ylabel("y")
+plt.legend()
+plt.grid()
+plt.savefig("rosenbrock.png")

talk/.github/workflows/build-sldes.yaml

@@ -0,0 +1,50 @@
+name: Build and Deploy Slides
+
+on:
+  push:
+    branches:
+      - main  # Triggers on pushes to the main branch
+
+jobs:
+  build:
+    runs-on: ubuntu-latest
+
+    steps:
+    - name: Check out the repository
+      uses: actions/checkout@v2
+
+    - name: Set up Docker
+      run: |
+        echo "Docker is already available on GitHub-hosted runners."
+
+    - name: Run Marp to generate HTML slides
+      run: |
+        docker run --rm --init \
+          -v ${{ github.workspace }}:/home/marp/app \
+          -e LANG=${{ secrets.LANG }} \
+          -p 8080:8080 -p 37717:37717 \
+          marpteam/marp-cli:v3.2.0 --theme ngi-theme.css --html .
+
+    - name: Commit and Push HTML slides to gh-pages
+      run: |
+        # Install Git if it's not already available
+        sudo apt-get update
+        sudo apt-get install -y git
+
+        # Configure git user
+        git config --global user.name "GitHub Action"
+        git config --global user.email "[email protected]"
+
+        # Create gh-pages branch if it doesn't exist
+        git checkout -b gh-pages
+
+        # Remove old slides and add new ones
+        rm -rf ./*
+        cp -r /home/marp/app/*.html .
+
+        # Add, commit, and push the changes
+        git add .
+        git commit -m "Update slides"
+        git push -f origin gh-pages
+      env:
+        GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}