app.py

from flask import Flask, jsonify, request
from transformers import pipeline, AutoTokenizer
import joblib
import json

# Load the model
model = joblib.load("iris_svm.joblib")

# Load the configuration file
with open("config.json", "r") as f:
    config = json.load(f)

# Get the input features and target variable
features = config["features"]
target = config["targets"][0]
target_mapping = config["target_mapping"]

# Initialize the tokenizer
tokenizer = AutoTokenizer.from_pretrained("bert-base-uncased")

# Initialize the Flask app
app = Flask(__name__)

# Define the prediction route
@app.route("/predict", methods=["POST"])
def predict():
    # Get the input data
    input_data = request.json

    # Construct the input text for the pipeline
    input_text = f"SepalLengthCm: {input_data['SepalLengthCm']}, SepalWidthCm: {input_data['SepalWidthCm']}, PetalLengthCm: {input_data['PetalLengthCm']}, PetalWidthCm: {input_data['PetalWidthCm']}"

    # Tokenize the input text
    tokenized_input = tokenizer(input_text, return_tensors="pt")

    # Make a prediction using the pipeline
    classifier = pipeline("text-classification", model=model, tokenizer=tokenizer, device=0)
    predicted_class_id = classifier(tokenized_input)[0]['label']

    # Convert the predicted class ID to a class name
    predicted_class_name = list(target_mapping.keys())[list(target_mapping.values()).index(predicted_class_id)]

    # Return the predicted class name as a JSON response
    return jsonify({"predicted_class": predicted_class_name})

# Run the Flask app
if __name__ == "__main__":
    app.run(debug=True)