Nathanotal's picture
df
eea8dd3
import gradio as gr
from PIL import Image
import hopsworks
# If You want to inspect the results for the synthetic data set SYNTHETIC = TRUE
SYNTHETIC = False
latestSurvivorImage = 'latest_survivor_pred'
latestSurvivorPred = 'latest_survivor_label_pred'
latestSurvivorLabel = 'latest_survivor_label_actual'
recentHistory = 'df_recent_titanic'
confusionMatrix = 'confusion_matrix'
if SYNTHETIC:
latestSurvivorImage += '_synthetic'
latestSurvivorPred += '_synthetic'
latestSurvivorLabel += '_synthetic'
recentHistory += '_synthetic'
confusionMatrix += '_synthetic'
latestSurvivorImage += '.png'
latestSurvivorPred += '.png'
latestSurvivorLabel += '.png'
recentHistory += '.png'
confusionMatrix += '.png'
with gr.Blocks() as demo:
# Login to hopsworks
project = hopsworks.login()
fs = project.get_feature_store()
# Download all the necessary files
dataset_api = project.get_dataset_api()
print('Downloading...')
dataset_api.download(f"Resources/images/{latestSurvivorImage}")
dataset_api.download(
f"Resources/images/{latestSurvivorPred}")
dataset_api.download(
f"Resources/images/{latestSurvivorLabel}")
dataset_api.download(f"Resources/images/{recentHistory}")
dataset_api.download(f"Resources/images/{confusionMatrix}")
# Arrange the images
with gr.Column():
gr.Label("Today's passenger")
input_img = gr.Image(f"{latestSurvivorImage}",
elem_id="passenger-img").style(
height='100', rounded=False)
with gr.Row():
with gr.Column():
gr.Label("Today's predicted survival")
input_img = gr.Image(
f"{latestSurvivorPred}", elem_id="predicted-img").style(
height='100', rounded=False)
with gr.Column():
gr.Label("Today's actual survival")
input_img = gr.Image(
f"{latestSurvivorLabel}", elem_id="actual-img").style(
height='100', rounded=False)
with gr.Row():
with gr.Column():
gr.Label("Recent Prediction History")
input_img = gr.Image(
f"{recentHistory}", elem_id="recent-predictions")
with gr.Column():
gr.Label(
"Confusion Maxtrix with Historical Prediction Performance")
input_img = gr.Image(f"{confusionMatrix}",
elem_id="confusion-matrix")
demo.launch()