12ongo commited on
Commit
96bf02b
·
verified ·
1 Parent(s): 076b640

Upload folder using huggingface_hub

Browse files
Files changed (3) hide show
  1. Dockerfile +12 -9
  2. app.py +90 -48
  3. requirements.txt +8 -0
Dockerfile CHANGED
@@ -1,16 +1,19 @@
1
- # Use a minimal base image with Python 3.9 installed
2
  FROM python:3.9-slim
3
 
4
- # Set the working directory inside the container to /app
 
 
 
5
  WORKDIR /app
6
 
7
- # Copy all files from the current directory on the host to the container's /app directory
8
  COPY . .
9
 
10
- # Install Python dependencies listed in requirements.txt
11
- RUN pip3 install -r requirements.txt
12
-
13
- # Define the command to run the Streamlit app on port 8501 and make it accessible externally
14
- CMD ["streamlit", "run", "app.py", "--server.port=8501", "--server.address=0.0.0.0", "--server.enableXsrfProtection=false"]
15
 
16
- # NOTE: Disable XSRF protection for easier external access in order to make batch predictions
 
 
 
 
 
 
1
  FROM python:3.9-slim
2
 
3
+ # Install wget
4
+ RUN apt-get update && apt-get install -y wget && rm -rf /var/lib/apt/lists/*
5
+
6
+ # Set the working directory inside the container
7
  WORKDIR /app
8
 
9
+ # Copy all files from the current directory to the container's working directory
10
  COPY . .
11
 
12
+ # Install dependencies from the requirements file without using cache to reduce image size
13
+ RUN pip install --no-cache-dir --upgrade -r requirements.txt
 
 
 
14
 
15
+ # Define the command to start the application using Gunicorn with 4 worker processes
16
+ # - `-w 4`: Uses 4 worker processes for handling requests
17
+ # - `-b 0.0.0.0:7860`: Binds the server to port 7860 on all network interfaces
18
+ # - `app:sales_predictor_api`: Runs the Flask app instance named 'sales_predictor_api' from the 'app.py' file
19
+ CMD ["gunicorn", "-w", "4", "-b", "0.0.0.0:7860", "app:sales_predictor_api"]
app.py CHANGED
@@ -1,48 +1,90 @@
1
- import streamlit as st
2
- import pandas as pd
3
- import requests
4
-
5
- # Set the title of the Streamlit app
6
- st.title("SuperKart Sales Prediction")
7
-
8
- # Section for online prediction
9
- st.subheader("Predict Single Product Sales")
10
-
11
- # Collect user input for product and store features
12
- product_weight = st.number_input("Product Weight", min_value=0.0, value=10.0)
13
- product_sugar_content = st.selectbox("Product Sugar Content", ["Low Sugar", "Regular", "No Sugar", "reg"]) # Include 'reg' based on EDA
14
- product_allocated_area = st.number_input("Product Allocated Area", min_value=0.0, value=0.05)
15
- product_type = st.selectbox("Product Type", ['Frozen Foods', 'Dairy', 'Canned', 'Baking Goods', 'Health and Hygiene', 'Snack Foods', 'Meat', 'Household', 'Hard Drinks', 'Fruits and Vegetables', 'Breads', 'Soft Drinks', 'Breakfast', 'Others', 'Starchy Foods', 'Seafood']) # Based on EDA
16
- product_mrp = st.number_input("Product MRP", min_value=0.0, value=100.0)
17
- store_establishment_year_before_2005 = st.selectbox("Store Established Before 2005?", ["True", "False"])
18
- store_size = st.selectbox("Store Size", ["Small", "Medium", "High"])
19
- store_location_city_type = st.selectbox("Store Location City Type", ["Tier 1", "Tier 2", "Tier 3"])
20
- store_type = st.selectbox("Store Type", ['Supermarket Type2', 'Departmental Store', 'Supermarket Type1', 'Food Mart']) # Based on EDA
21
-
22
-
23
- # Convert user input into a dictionary
24
- input_data = {
25
- 'Product_Weight': product_weight,
26
- 'Product_Sugar_Content': product_sugar_content,
27
- 'Product_Allocated_Area': product_allocated_area,
28
- 'Product_Type': product_type,
29
- 'Product_MRP': product_mrp,
30
- 'Store_Establishment_Year_Before_2005': store_establishment_year_before_2005 == "True", # Convert string to boolean
31
- 'Store_Size': store_size,
32
- 'Store_Location_City_Type': store_location_city_type,
33
- 'Store_Type': store_type
34
- }
35
-
36
- # Make prediction when the "Predict" button is clicked
37
- if st.button("Predict Sales"):
38
- # Replace <username>-<repo_id> with your actual Hugging Face username and backend space repo ID
39
- backend_url = "https://<username>-<repo_id>.hf.space/v1/predict" # Update endpoint
40
- response = requests.post(backend_url, json=input_data) # Send data as JSON
41
- if response.status_code == 200:
42
- prediction = response.json()['Predicted Sales Total (in dollars)']
43
- st.success(f"Predicted Sales Total (in dollars): {prediction}")
44
- else:
45
- st.error(f"Error making prediction. Status code: {response.status_code}")
46
- st.error(f"Response: {response.text}")
47
-
48
- # Removed batch prediction section as it's commented out in the backend Flask app
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # Import necessary libraries
2
+ import numpy as np
3
+ import joblib # For loading the serialized model
4
+ import pandas as pd # For data manipulation
5
+ from flask import Flask, request, jsonify # For creating the Flask API
6
+
7
+ # Initialize the Flask application
8
+ sales_predictor_api = Flask("SuperKart Sales Predictor")
9
+
10
+ # Load the trained machine learning model
11
+ model = joblib.load("sales_predictor.joblib")
12
+
13
+ # Define a route for the home page (GET request)
14
+ @sales_predictor_api.get('/')
15
+ def home():
16
+ """
17
+ This function handles GET requests to the root URL ('/') of the API.
18
+ It returns a simple welcome message.
19
+ """
20
+ return "Welcome to the SuperKart Sales Prediction API!"
21
+
22
+ # Define an endpoint for single product prediction (POST request)
23
+ @sales_predictor_api.post('/v1/predict')
24
+ def predict_sales():
25
+ """
26
+ This function handles POST requests to the '/v1/predict' endpoint.
27
+ It expects a JSON payload containing product and store details and returns
28
+ the predicted sales as a JSON response.
29
+ """
30
+ # Get the JSON data from the request body
31
+ product_store_data = request.get_json()
32
+
33
+ # Extract relevant features from the JSON data
34
+ sample = {
35
+ 'Product_Weight': product_store_data['Product_Weight'],
36
+ 'Product_Sugar_Content': product_store_data['Product_Sugar_Content'],
37
+ 'Product_Allocated_Area': product_store_data['Product_Allocated_Area'],
38
+ 'Product_Type': product_store_data['Product_Type'],
39
+ 'Product_MRP': product_store_data['Product_MRP'],
40
+ 'Store_Establishment_Year_Before_2005': product_store_data['Store_Establishment_Year_Before_2005'],
41
+ 'Store_Size': product_store_data['Store_Size'],
42
+ 'Store_Location_City_Type': product_store_data['Store_Location_City_Type'],
43
+ 'Store_Type': product_store_data['Store_Type']
44
+ }
45
+
46
+ # Convert the extracted data into a Pandas DataFrame
47
+ input_data = pd.DataFrame([sample])
48
+
49
+ # Make prediction
50
+ predicted_sales = model.predict(input_data)[0]
51
+
52
+ # Convert predicted_sales to Python float
53
+ predicted_sales = round(float(predicted_sales), 2)
54
+
55
+
56
+ # Return the predicted sales
57
+ return jsonify({'Predicted Sales Total (in dollars)': predicted_sales})
58
+
59
+
60
+ # Define an endpoint for batch prediction (POST request) - This endpoint is not functional with the current model
61
+ # as the model is trained on individual product-store combinations, not batch data with IDs.
62
+ # @sales_predictor_api.post('/v1/predictbatch')
63
+ # def predict_sales_batch():
64
+ # """
65
+ # This function handles POST requests to the '/v1/predictbatch' endpoint.
66
+ # It expects a CSV file containing product and store details for multiple entries
67
+ # and returns the predicted sales as a dictionary in the JSON response.
68
+ # """
69
+ # # Get the uploaded CSV file from the request
70
+ # file = request.files['file']
71
+
72
+ # # Read the CSV file into a Pandas DataFrame
73
+ # input_data = pd.read_csv(file)
74
+
75
+ # # Make predictions for all entries in the DataFrame
76
+ # predicted_sales = model.predict(input_data).tolist()
77
+
78
+ # # For batch prediction with IDs, you would need an ID column in the input data
79
+ # # and a way to associate predictions with IDs.
80
+ # # Since the current model predicts for individual entries based on features,
81
+ # # this batch endpoint is not directly applicable without further modifications
82
+ # # to handle input data with IDs and return predictions associated with those IDs.
83
+
84
+ # # Returning a placeholder response
85
+ # return jsonify({"message": "Batch prediction endpoint is not yet implemented for this model."})
86
+
87
+
88
+ # Run the Flask application in debug mode if this script is executed directly
89
+ if __name__ == '__main__':
90
+ sales_predictor_api.run(debug=True)
requirements.txt CHANGED
@@ -1,3 +1,11 @@
1
  pandas==2.2.2
 
 
 
 
 
 
 
2
  requests==2.28.1
 
3
  streamlit==1.43.2
 
1
  pandas==2.2.2
2
+ numpy==2.0.2
3
+ scikit-learn==1.6.1
4
+ xgboost==2.1.4
5
+ joblib==1.4.2
6
+ Werkzeug==2.2.2
7
+ flask==2.2.2
8
+ gunicorn==20.1.0
9
  requests==2.28.1
10
+ uvicorn[standard]
11
  streamlit==1.43.2