Merge branch 'main' of github.com:UBC-CS/cpsc330-2024W1

.github/workflows/deploy.yml

@@ -26,6 +26,20 @@ jobs:
         path: ${{ env.pythonLocation }}
         key: ${{ env.pythonLocation }}-${{ hashFiles('setup.py') }}-${{ hashFiles('requirements.txt') }}
 
+    - id: install-graphviz-linux
+      name: Install Graphviz on Linux
+      # if: runner.os == 'Linux'
+      # shell: bash
+      run: |
+        # Install Graphviz on Linux
+        sudo apt update
+        sudo apt install -qy gsfonts
+        sudo apt -qq list fonts-liberation gsfonts
+        sudo apt install -qy graphviz
+        sudo apt -y autoremove --purge
+        sudo apt -y autoclean
+        sudo apt clean
+
     - name: Install dependencies
       run: |
         pip install -r requirements.txt

README.md

@@ -1,3 +1,5 @@
+[![deploy-book](https://github.com/UBC-CS/cpsc330-2024W1/actions/workflows/deploy.yml/badge.svg?branch=main)](https://github.com/UBC-CS/cpsc330-2024W1/actions/workflows/deploy.yml)
+
 # UBC CPSC 330: Applied Machine Learning (2024W1)
 
 This is the course homepage for CPSC 330: Applied Machine Learning at the University of British Columbia. You are looking at the current version (Sep-Dec 2024).
@@ -12,25 +14,24 @@ This is the course homepage for CPSC 330: Applied Machine Learning at the Univer
 - Devyani McLaren (cpsc330-admin@cs.ubc.ca), please reach out to Devyani for: admin questions, extensions, academic concessions etc. 
 
 ### TAs
-- Akash Adhikary	akash7adhikary@gmail.com
-- Amirali Goodarzvand Chegini	amiralic@student.ubc.ca
-- Aryan Ballani	aryanballani070603@gmail.com 
-- Atabak Eghbal	eghbal.atabak@gmail.com
-- Derrick Cheng	dcheng04@students.cs.ubc.ca
-- Frederick Sunstrum	fr.sunstrum@gmail.com
-- Hongkai Liu	liuhongkai112233@163.com
-- Noah Marusenko	noah.marusenko@icloud.com
-- Jialin (Mike) Lu	mike020830@gmail.com
-- Kimia Rostin	krostin@student.ubc.ca
-- Mahsa Zarei	mahsazarei19@gmail.com
-- Mike Ju	feng0025@student.ubc.ca
-- Mishaal Kazmi	mkazmi@cs.ubc.ca
-- Rubia Reis Guerra	rubiarg@cs.ubc.ca
-- Shadab Shaikh	shadabs3@cs.ubc.ca
-- Sohbat Sandhu	sohbat@student.ubc.ca
-- Sparsh Trivedy	sparsh01@student.ubc.ca
-- Stash Currie	stashubc@student.ubc.ca
-- Tianyu (Niki) Duan	dty200@student.ubc.ca
+- Akash Adhikary	
+- Amirali Goodarzvand Chegini	
+- Aryan Ballani	
+- Atabak Eghbal	
+- Derrick Cheng	
+- Frederick Sunstrum
+- Hongkai Liu
+- Noah Marusenko
+- Jialin (Mike) Lu
+- Kimia Rostin	
+- Mahsa Zarei	
+- Mike Ju	
+- Mishaal Kazmi	
+- Rubia Reis Guerra	
+- Shadab Shaikh	
+- Sohbat Sandhu	
+- Stash Currie	
+- Tianyu (Niki) Duan	
 
 ## License
 
@@ -57,7 +58,7 @@ Usually the homework assignments will be due on Mondays (except next week) and w
 | Assessment     | Due date              | Where to find?                                                                     | Where to submit?                                      |
 |----------------|-----------------------|------------------------------------------------------------------------------------|-------------------------------------------------------|
 | hw1            | Sept 10, 11:59 pm     | [GitHub repo](https://github.com/new?template_name=hw1&template_owner=ubc-cpsc330) | [Gradescope](https://www.gradescope.ca/courses/18608) |
-| hw2            | Sept 16, 11:59 pm     | GitHub repo[](https://github.com/new?template_name=hw2&template_owner=ubc-cpsc330) | [Gradescope](https://www.gradescope.ca/courses/18608) |
+| hw2            | Sept 16, 11:59 pm     | [GitHub repo](https://github.com/new?template_name=hw2&template_owner=ubc-cpsc330) | [Gradescope](https://www.gradescope.ca/courses/18608) |
 | Syllabus quiz  | Sept 19, 11:59 pm     | [PrairieLearn](https://us.prairielearn.com/pl/course_instance/163899/assessment/2451488) | [PrairieLearn](https://us.prairielearn.com/pl/course_instance/163899/assessment/2451488)       |
 | hw3            | Oct 01, 11:59 pm      | GitHub repo[](https://github.com/new?template_name=hw3&template_owner=ubc-cpsc330) | [Gradescope](https://www.gradescope.ca/courses/18608) |
 | hw4            | Oct 07, 11:59 pm      | GitHub repo[](https://github.com/new?template_name=hw4&template_owner=ubc-cpsc330) | [Gradescope](https://www.gradescope.ca/courses/18608) |

_toc.yml

@@ -21,6 +21,8 @@ parts:
 - caption: Lectures
   chapters: 
   - file: lectures/notes/01_intro.ipynb
+  - file: lectures/notes/02_terminology-decision-trees.ipynb
+  - file: lectures/notes/03_ml-fundamentals.ipynb    
 - caption: Section slides
   chapters:
   - file: lectures/101-Giulia-lectures/README

lectures/code/plotting_functions.py

@@ -28,7 +28,7 @@ def plot_tree_decision_boundary(
 
 
 def plot_tree_decision_boundary_and_tree(
-    model, X, y, height=6, width=16, x_label="x-axis", y_label="y-axis", eps=None
+    model, X, y, height=6, width=16, fontsize = 9, x_label="x-axis", y_label="y-axis", eps=None
 ):
     fig, ax = plt.subplots(
         1,
@@ -38,7 +38,11 @@ def plot_tree_decision_boundary_and_tree(
         gridspec_kw={"width_ratios": [1.5, 2]},
     )
     plot_tree_decision_boundary(model, X, y, x_label, y_label, eps, ax=ax[0])
-    ax[1].imshow(tree_image(X.columns, model))
+    custom_plot_tree(model, 
+                 feature_names=X.columns.tolist(), 
+                 class_names=['A+', 'not A+'],
+                 impurity=False,
+                 fontsize=fontsize, ax=ax[1])
     ax[1].set_axis_off()
     plt.show()
 

lectures/code/utils.py

@@ -1,13 +1,8 @@
 import pandas as pd
 import numpy as np
-import re 
-from PIL import Image
-from torchvision.models import vgg16
-from torchvision import transforms
-import torch
+import re
 from sklearn.model_selection import cross_val_score, cross_validate, train_test_split
-
-from sklearn.tree import DecisionTreeClassifier, DecisionTreeRegressor, export_graphviz
+from sklearn.tree import DecisionTreeClassifier, DecisionTreeRegressor, export_graphviz, plot_tree
 
 import glob
 
@@ -18,46 +13,40 @@
 
 plt.rcParams["font.size"] = 16
 
-def display_tree(feature_names, tree, counts=False):
-    """ For binary classification only """
-    dot = export_graphviz(
-        tree,
-        out_file=None,
-        feature_names=feature_names,
-        class_names=tree.classes_.astype(str),
-        impurity=False,
-    )    
-    # adapted from https://stackoverflow.com/questions/44821349/python-graphviz-remove-legend-on-nodes-of-decisiontreeclassifier
-    # dot = re.sub('(\\\\nsamples = [0-9]+)(\\\\nvalue = \[[0-9]+, [0-9]+\])(\\\\nclass = [A-Za-z0-9]+)', '', dot)
-    if counts: 
-        dot = re.sub("(samples = [0-9]+)\\\\n", "", dot)
-        dot = re.sub("value", "counts", dot)
-    else:
-        dot = re.sub("(\\\\nsamples = [0-9]+)(\\\\nvalue = \[[0-9]+, [0-9]+\])", "", dot)
-        dot = re.sub("(samples = [0-9]+)(\\\\nvalue = \[[0-9]+, [0-9]+\])\\\\n", "", dot)
-
-    return graphviz.Source(dot)
-
-
-def tree_image(feature_names, tree):
-    """ For binary classification only """
-    dot = export_graphviz(
-        tree,
-        out_file=None,
-        feature_names=feature_names,
-        class_names=tree.classes_.astype(str),
-        impurity=False,
-    )    
-    # adapted from https://stackoverflow.com/questions/44821349/python-graphviz-remove-legend-on-nodes-of-decisiontreeclassifier
-    # dot = re.sub('(\\\\nsamples = [0-9]+)(\\\\nvalue = \[[0-9]+, [0-9]+\])(\\\\nclass = [A-Za-z0-9]+)', '', dot)
-    #dot = re.sub("(\\\\nsamples = [0-9]+)(\\\\nvalue = \[[0-9]+, [0-9]+\])", "", dot)
-    #dot = re.sub("(samples = [0-9]+)(\\\\nvalue = \[[0-9]+, [0-9]+\])\\\\n", "", dot)    
-    dot = re.sub("(samples = [0-9]+)\\\\n", "", dot)
-    dot = re.sub("value", "counts", dot)        
-    graph = graphviz.Source(dot, format="png")
-    fout = "tmp"
-    graph.render(fout)
-    return imread(fout + ".png")
+# Custom function to customize the tree plot and hide values and samples
+def custom_plot_tree(tree_model, feature_names=None, class_names=None, **kwargs):
+    """
+    Customizes and displays a tree plot for a scikit-learn Decision Tree Classifier.
+
+    Parameters:
+    - tree (sklearn.tree.DecisionTreeClassifier): The trained Decision Tree Classifier to visualize.
+    - width: width of the matplotlib plot in inches 
+    - height: height of the matplotlib plot in inches 
+    - feature_names (list or None): A list of feature names to label the tree nodes with feature names.
+                                    If None, generic feature names will be used.
+    - class_names (list or None): A list of class names to label the tree nodes with class names.
+                                  If None, generic class names will be used.
+    - **kwargs: Additional keyword arguments to be passed to the `sklearn.tree.plot_tree` function.
+
+    Returns:
+    - None: The function displays the customized tree plot using Matplotlib.
+    
+    This function customizes the appearance of a Decision Tree plot generated by the scikit-learn
+    `plot_tree` function. It hides both the samples and values in each node of the tree plot
+    for improved visualization.
+    """    
+    plot_tree(tree_model, 
+              feature_names=feature_names, 
+              class_names=class_names, 
+              filled=True, 
+              **kwargs)
+
+    # Customize the appearance of the text elements for each node
+    for text in plt.gca().texts:
+        new_text = re.sub('samples = \d+\n', '', text.get_text()) # Hide samples
+        text.set_text(new_text) 
+
+    plt.show()
 
 def cross_validate_std(*args, **kwargs):
     """Like cross_validate, except also gives the standard deviation of the score"""