third script done

Dockerfile

@@ -12,4 +12,5 @@ RUN conda install -y \
     notebook=7.0.6 \
     pytest=7.4.3 \
     responses=0.24.1 \
-    click=8.0.3
+    click=8.0.3 \
+    vl-convert-python=1.1.0

README.md

@@ -83,13 +83,25 @@ enter the following commands in the terminal in the project root:
 
 ```
 # download and extract data
-python scripts/download_data.py --url="https://archive.ics.uci.edu/static/public/15/breast+cancer+wisconsin+original.zip" --write-to="data/raw"
+python scripts/download_data.py --url="https://archive.ics.uci.edu/static/public/15/breast+cancer+wisconsin+original.zip" \
+   --write-to="data/raw"
 
 # split data into train and test sets, preprocess data for eda 
 # and save preprocessor
-python scripts/split_n_preprocess.py --raw-data=data/raw/wdbc.data --data-to=data/processed --preprocessor-to=results/models --seed=522
-
-
+python scripts/split_n_preprocess.py --raw-data=data/raw/wdbc.data \
+   --data-to=data/processed \
+   --preprocessor-to=results/models \
+   --seed=522
+
+# perform eda and save plots
+
+# train model, create visualize tuning, and save plot and model
+python scripts/fit_breast_cancer_classifier.py --training-data=data/processed/cancer_train.csv \
+   --preprocessor=results/models/cancer_preprocessor.pickle \
+   --columns-to-drop=data/processed/columns_to_drop.csv \
+   --pipeline-to=results/models \
+   --plot-to=results/figures \
+   --seed=523
 ```
 
 #### Clean up

scripts/fit_breast_cancer_classifier.py

@@ -4,22 +4,22 @@
 
 import click
 import os
+import altair as alt
 import numpy as np
 import pandas as pd
 import pickle
-from sklearn.model_selection import train_test_split
 from sklearn import set_config
-from sklearn.model_selection import train_test_split
-from sklearn.preprocessing import StandardScaler
-from sklearn.compose import make_column_transformer, make_column_selector
-
+from sklearn.neighbors import KNeighborsClassifier
+from sklearn.pipeline import make_pipeline
+from sklearn.model_selection import GridSearchCV
+from sklearn.metrics import fbeta_score, make_scorer
 
 @click.command()
 @click.option('--training-data', type=str, help="Path to training data")
 @click.option('--preprocessor', type=str, help="Path to preprocessor object")
 @click.option('--columns-to-drop', type=str, help="Optional: columns to drop")
 @click.option('--pipeline-to', type=str, help="Path to directory where the pipeline object will be written to")
-@click.option('--plot-to', type=int, help="Path to directory where the plot will be written to")
+@click.option('--plot-to', type=str, help="Path to directory where the plot will be written to")
 @click.option('--seed', type=int, help="Random seed", default=123)
 
 def main(training_data, preprocessor, columns_to_drop, pipeline_to, plot_to, seed):
@@ -30,10 +30,11 @@ def main(training_data, preprocessor, columns_to_drop, pipeline_to, plot_to, see
 
     # read in data & preprocessor
     cancer_train = pd.read_csv(training_data)
-    pickle.load(open(preprocessor, "rb"))
+    cancer_preprocessor = pickle.load(open(preprocessor, "rb"))
 
     if columns_to_drop:
-        to_drop = pd.read_csv(columns_to_drop).['feats_to_drop'].tolist()
+        to_drop = pd.read_csv(columns_to_drop).feats_to_drop.tolist()
+        print(to_drop)
         cancer_train = cancer_train.drop(columns=to_drop)
 
     # tune model (here, find K for k-nn using 30 fold cv)