Update + add C++ example

lectures/3/3.md

@@ -234,7 +234,7 @@ _class: titlepage
 
 # Modules: reusable code in Python
 
-In Python, the ability to reuse code is facilitated by modules. A module is a file with a `.py` extension that contains functions and variables. There are various methods to write modules, including using languages like C to create compiled modules.
+In Python, the ability to reuse code is facilitated by modules. A module is a file with a `.py` extension that contains functions and variables. There are various methods to write modules, including using languages like C or C++ to create compiled modules.
 
 When importing a module, to enhance import performance, Python creates byte-compiled files (`__pycache__/filename.pyc`). These files, platform-independent and located in the same directory as the corresponding `.py` files, speed up subsequent imports by storing preprocessed code.
 
@@ -476,14 +476,6 @@ In addition to the standard library, there is an active collection of hundreds o
 _class: titlepage
 -->
 
-# **Zen of Python**:<br>"Explicit is better than implicit."<br>Run `import this` in Python to learn more.
-
----
-
-<!--
-_class: titlepage
--->
-
 # Error handling
 
 ---
@@ -500,7 +492,7 @@ print("Another line") # Code fails before getting to this line.
 
 ---
 
-# `try-except`
+# `try`-`except`
 
 ```python
 try:
@@ -553,7 +545,7 @@ my_list[5]
 > IndexError: list index out of range
 
 ```python
-my_tuple = (1,2,3)
+my_tuple = (1, 2, 3)
 my_tuple[0] = 0
 ```
 > TypeError: 'tuple' object does not support item assignment
@@ -585,6 +577,14 @@ Finally, we can even define our own exception types by inheriting from the `Exce
 _class: titlepage
 -->
 
+# **Zen of Python**:<br>"Explicit is better than implicit."<br>Run `import this` in Python to learn more.
+
+---
+
+<!--
+_class: titlepage
+-->
+
 # Dependency management
 
 ---
@@ -593,7 +593,7 @@ _class: titlepage
 
 1. **Code longevity:** How do you ensure that your code will still function as expected in one year, or even five? Consider the implications of using libraries such as Numpy, TensorFlow, or packages from sources like GitHub.
 
-2. **Consistent results:** How can you guarantee that both your current and future collaborators are able to achieve the same computational results as you?
+2. **Consistent results:** How can you guarantee that both your current and future collaborators are able to achieve the same numerical results as you?
 
 3. **Easy installation:** What steps can you take to simplify the process for collaborators to set up your code with all required dependencies?
 
@@ -618,7 +618,7 @@ _class: titlepage
 # PyPI (The Python Package Index)
 
 - **Installation tool**: pip
-- **Usage summary**: PyPI is primarily used for Python-only packages or Python interfaces to external libraries. It also hosts packages that include bundled external libraries, such as numpy.
+- **Usage summary**: PyPI is primarily used for Python-only packages or Python interfaces to external libraries. It also hosts packages that include bundled external libraries, such as NumPy.
 - **Number of packages**: Extensive, with long-term support for older versions.
 - **Handling libraries**: Dependencies on external libraries require either inclusion in the package or installation through other means (e.g., OS installers or manual setup).
 - **Pros**:
@@ -682,7 +682,7 @@ For effective management and replication of Python environments, recording depen
 
 #### `requirements.txt`
 
-This is a straightforward text file used by pip with virtual environments. It lists the packages your project depends on. Here’s a basic example:
+This is a straightforward text file used by pip with virtual environments. It lists the packages your project depends on. Here’s a basic example, obtained by running `python -m pip freeze > requirements.txt`:
 
 ```
 numpy
@@ -697,7 +697,7 @@ scipy
 
 #### `environment.yml`
 
-Used by Conda, this YAML file provides a structured format to specify the name of the environment, the channels from which packages should be sourced, and the dependencies themselves. Example:
+Used by Conda, this YAML file provides a structured format to specify the name of the environment, the channels from which packages should be sourced, and the dependencies themselves. Example (`conda env export --name my_env > environment.yml`):
 
 ```yaml
 name: my-environment
@@ -787,7 +787,7 @@ Docker is a platform for developing, shipping, and running applications inside c
 # Pull the image.
 docker pull python:latest
 
-# Create a container.
+# Create a container with shared folder.
 docker run --name my_container -v /path/to/host/folder:/shared-folder -it -d python:latest
 
 # Enable the container.
@@ -813,7 +813,7 @@ If the status of the container is `Up`, you can stop it with
 docker stop my_container
 ```
 
-Once you have created your container remember to **do not** use again the commad `run` but just `start`. Otherwise you will create every time a new container. If you want to remove a container you can run:
+Once you have created your container remember to **not** use again the commad `run` but just `start`. Otherwise you will create a new container every time. If you want to remove a container you can run:
 
 ```bash
 docker rm <name-of-the-container>
@@ -849,7 +849,7 @@ A Dockerfile is a text document that contains all the commands a user could call
 
 ---
 
-# Building a Docker image: example
+# Building a Docker image: example of `Dockerfile`
 
 ```Dockerfile
 # Start from a Python base image.
@@ -891,7 +891,7 @@ docker push my_image:version # If needed.
 _class: titlepage
 -->
 
-# Continuous Integration/Continuous Deployment
+# Continuous Integration/<br>Continuous Deployment
 
 ---
 
@@ -976,18 +976,19 @@ jobs:
   test-and-document:
     runs-on: ubuntu-latest
 
-    container:
-      image: python:3.8-slim
-      options: --user 1001:1001
-
     steps:
     - name: Checkout repository
       uses: actions/checkout@v3
 
+    - name: Set up Python
+        uses: actions/setup-python@v4
+        with:
+          python-version: "3.8"
+          
     - name: Set up Python environment
       run: |
         python -m venv venv
-        . venv/bin/activate
+        source venv/bin/activate
 
 ```
 
@@ -1003,12 +1004,12 @@ jobs:
 
     - name: Run tests
       run: |
-        . venv/bin/activate
+        source venv/bin/activate
         pytest
 
     - name: Generate documentation
       run: |
-        . venv/bin/activate
+        source venv/bin/activate
         cd docs
         sphinx-build -b html . _build/html
 
@@ -1022,7 +1023,7 @@ jobs:
 ---
 
 # Explanation of key components
-1. **Container image**: The workflow runs in a container based on the `python:3.8-slim` image. This ensures that Python and any necessary system dependencies are already installed.
+1. **Container image**: The workflow runs in a container based on the `python:3.8` image. This ensures that Python and any necessary system dependencies are already installed.
   
 2. **Setup Python environment**: A virtual environment is set up within the container to isolate our project dependencies.
   

lectures/3/c++_vs_py/CMakeLists.txt

@@ -0,0 +1,7 @@
+cmake_minimum_required(VERSION 3.5)
+project(example)
+
+find_package(pybind11 REQUIRED)
+include_directories(SYSTEM ${pybind11_INCLUDE_DIRS})
+
+pybind11_add_module(matrix_ops matrix_multiplication.cpp)

lectures/3/c++_vs_py/main.py

@@ -0,0 +1,34 @@
+import matrix_ops
+import time
+import numpy as np
+
+# Create random matrices.
+n = 500
+mat1 = np.random.rand(n, n)
+mat2 = np.random.rand(n, n)
+
+# Benchmark C++ implementation (using pybind11).
+start_time = time.time()
+result_cpp = matrix_ops.matrix_multiply(mat1, mat2)
+cpp_duration = time.time() - start_time
+print(f"C++ implementation took {cpp_duration:.4f} seconds.")
+
+def matrix_multiply_python(mat1, mat2):
+    rows = len(mat1)
+    cols = len(mat2[0])
+    inner_dim = len(mat2)
+
+    result = [[0 for _ in range(cols)] for _ in range(rows)]
+    for i in range(rows):
+        for j in range(cols):
+            for k in range(inner_dim):
+                result[i][j] += mat1[i][k] * mat2[k][j]
+    return result
+
+# Benchmark Python implementation.
+start_time = time.time()
+result_python = matrix_multiply_python(mat1, mat2)
+python_duration = time.time() - start_time
+print(f"Python implementation took {python_duration:.4f} seconds.")
+
+print(f"Speedup: {python_duration // cpp_duration}.")

lectures/3/c++_vs_py/matrix_multiplication.cpp

@@ -0,0 +1,31 @@
+#include <pybind11/numpy.h>
+#include <pybind11/pybind11.h>
+#include <pybind11/stl.h>
+
+#include <vector>
+
+std::vector<std::vector<double>>
+matrix_multiply(const std::vector<std::vector<double>> &mat1,
+                const std::vector<std::vector<double>> &mat2) {
+  const size_t rows = mat1.size();
+  const size_t cols = mat2[0].size();
+  const size_t inner_dim = mat2.size();
+
+  std::vector<std::vector<double>> result(rows, std::vector<double>(cols, 0));
+  for (size_t i = 0; i < rows; ++i) {
+    for (size_t j = 0; j < cols; ++j) {
+      for (size_t k = 0; k < inner_dim; ++k) {
+        result[i][j] += mat1[i][k] * mat2[k][j];
+      }
+    }
+  }
+
+  return result;
+}
+
+namespace py = pybind11;
+
+PYBIND11_MODULE(matrix_ops, m) {
+  m.def("matrix_multiply", &matrix_multiply,
+        "A function which multiplies two NumPy matrices.");
+}