-
Notifications
You must be signed in to change notification settings - Fork 0
/
app.py
235 lines (203 loc) · 8.06 KB
/
app.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
from pathlib import Path
import joblib
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns
import streamlit as st
from sklearn.metrics import PredictionErrorDisplay
from sklearn.utils import shuffle
from millify import prettify
from training import load_model, get_estimator_name
st.set_page_config(layout="wide")
sns.set_context("paper")
@st.cache_data
def get_filepath(directory) -> list[str]:
file_paths = [file for file in directory.iterdir() if file.is_file()]
return file_paths
@st.cache_resource
def get_model_dict(filepaths) -> dict:
models = {}
for file in filepaths:
m = load_model(file)
models.update({get_estimator_name(m) : m})
return models
@st.cache_data
def check_nan(df) -> bool:
boolean = df.isna().values.any(axis=1)
return boolean
@st.cache_data
def get_nan_columns(df) -> list[str]:
nan_cols = [col for col in df.columns if df[col].isna().any()]
return nan_cols
def get_prediction_vs_actual(_model, X, y):
X, y = shuffle(X, y)
fig, ax = plt.subplots()
scatter_kwgs = {
"color": "steelblue",
"s": 15,
"edgecolor": "steelblue",
"alpha": 0.6
}
line_kwgs = {
"color": "red"
}
_ = PredictionErrorDisplay.from_estimator(
estimator=_model,
X=X,
y=y,
ax = ax,
subsample=3000,
scatter_kwargs=scatter_kwgs,
kind="actual_vs_predicted",
line_kwargs=line_kwgs,
)
ax.set_title("Prediction vs Actual Values")
return fig
def get_prediction_vs_residuals(_model, X, y):
X, y = shuffle(X, y)
fig, ax = plt.subplots()
scatter_kwgs = {
"color": "steelblue",
"s": 15,
"edgecolor": "steelblue",
"alpha": 0.6
}
_ = PredictionErrorDisplay.from_estimator(
estimator=_model,
X=X,
y=y,
ax = ax,
subsample=3000,
scatter_kwargs=scatter_kwgs
)
ax.set_title("Prediction vs Residuals")
return fig
model_directory = Path('./models')
paths_to_model = get_filepath(model_directory)
models_dict = get_model_dict(paths_to_model)
performance_df = pd.read_csv("./models_performance/performance_metrics.csv", index_col=0)
st.title("Diamond Price Prediction :gem:")
st.divider()
tab1, tab2 = st.tabs(["Prediction", "Model Performance"])
with st.sidebar:
selected_model = st.selectbox("Regression Models",
options=models_dict.keys(),
key="current_regressor")
with tab1:
st.markdown(
f"<h2 style='font-family: monospace; text-align: center; color: dodgerblue;'>{selected_model}</h2>",
unsafe_allow_html=True
)
with st.form("input_features"):
f_col1, f_col2, f_col3 = st.columns(3, gap='medium')
with f_col1:
carat = st.number_input(label="Carat",
min_value=np.nan,
key="carat",
max_value = 6.0
)
depth = st.number_input(label="depth",
min_value=np.nan,
key="depth",
max_value=100.0
)
table = st.number_input(label="table",
min_value=np.nan,
key='table',
max_value=100.0
)
with f_col2:
x = st.number_input(label="X",
min_value=np.nan,
key='x',
max_value=12.0
)
y = st.number_input(label="Y",
min_value=np.nan,
key='y',
max_value=70.0
)
z = st.number_input(label="Z",
min_value=np.nan,
key='z',
max_value=40.0
)
with f_col3:
cut = st.selectbox(label="Cut Quality",
options=['Fair', 'Very Good', 'Good', 'Premium', 'Ideal'],
index=None,
key='cut',
help="Cut Quality [Fair, Good, Very Good, Premium, Ideal(best)]")
color = st.selectbox(label="Color",
options=['J', 'I', 'H', 'G', 'F', 'E', 'D'],
index=None,
key='color',
help="Diamond Color:J(worst) to D(best)")
clarity = st.selectbox(label="Clarity",
options=['I1', 'SI2', 'SI1', 'VS2', 'VS1', 'VVS2', 'VVS1', 'IF'],
index=None,
key='clarity',
help="How clear the diamond is [I1 (worst), SI2, SI1, VS2, VS1, VVS2, VVS1, IF (best)]")
submit_button = st.form_submit_button("Predict")
features_dict = {
"carat": carat,
"cut": cut,
"color": color,
"clarity": clarity,
"depth": depth,
"table": table,
"x": x,
"y": y,
"z": z
}
input_df = pd.DataFrame(features_dict, index=[0])
if submit_button:
if check_nan(input_df):
nan_cols_ls = get_nan_columns(input_df)
nan_cols_str = ', '.join(nan_cols_ls)
st.error("InputValueError!", icon=":material/error:")
st.error(f"Input feature {nan_cols_str} contains unsupported/invalid value.", icon=":material/exclamation:")
st.warning("Prediction model expects valid numeric/categorical value.", icon=":material/data_alert:")
else:
prediction = np.round(models_dict[selected_model].predict(input_df), 2)
prediction_str = prediction.astype(np.str_)
p_col1, p_col2, p_col3 = st.columns(3, gap="large")
with p_col1.container(height=200, border=False):
st.header("Estimated Price:")
with p_col2.container(height=100, border=False):
st.header(f"${prettify(prediction_str[0])}")
with tab2:
st.markdown("<div style='margin-top: 30px;'></div>", unsafe_allow_html=True)
st.markdown(
f"<h3 style='font-family: monospace;text-align: center;'>Metrics of {selected_model}</h3>",
unsafe_allow_html=True)
mcol1, mcol2, mcol3, mcol4 = st.columns(4, gap="medium")
with mcol1.container(height=150):
st.subheader("R-squared")
st.subheader(f":gray-background[{performance_df.loc[selected_model, "R-squared"]}]")
with mcol2.container(height=150):
st.subheader("MAE")
st.subheader(f":gray-background[{performance_df.loc[selected_model, "Mean Absolute Error"]}]")
with mcol3.container(height=150):
st.subheader("MSE")
st.subheader(f":gray-background[{performance_df.loc[selected_model, "Mean Squared Error"]}]")
with mcol4.container(height=150):
st.subheader("RMSE")
st.subheader(f":gray-background[{performance_df.loc[selected_model, "Root Mean Squared Error"]}]")
st.markdown("<div style='margin-top: 50px;'></div>", unsafe_allow_html=True)
X_test = joblib.load("./datasets/X_test")
y_test = joblib.load("./datasets/y_test")
st.markdown(
f"<h3 style='font-family: monospace; text-align: center;'>Errors Viz of {selected_model} </h3>",
unsafe_allow_html=True)
fcol1, fcol2 = st.columns(2, gap="large")
with fcol1:
st.pyplot(
get_prediction_vs_actual(models_dict[selected_model], X_test, y_test)
)
with fcol2:
st.pyplot(
get_prediction_vs_residuals(models_dict[selected_model], X_test, y_test)
)
st.divider()