app/main.py

from fastapi import FastAPI, HTTPException
from pydantic import BaseModel
import pickle
import pandas as pd
import numpy as np
import xgboost
from xgboost import XGBRegressor

# Load the pre-trained pipeline
pipe = pickle.load(open('app/pipe.pkl', 'rb'))

# Define the list of teams and cities
teams = ['Australia', 'India', 'Bangladesh', 'New Zealand', 'South Africa', 'England',
         'West Indies', 'Afghanistan', 'Pakistan', 'Sri Lanka']

cities = ['Colombo', 'Mirpur', 'Johannesburg', 'Dubai', 'Auckland', 'Cape Town', 'London',
          'Pallekele', 'Barbados', 'Sydney', 'Melbourne', 'Durban', 'St Lucia', 'Wellington',
          'Lauderhill', 'Hamilton', 'Centurion', 'Manchester', 'Abu Dhabi', 'Mumbai',
          'Nottingham', 'Southampton', 'Mount Maunganui', 'Chittagong', 'Kolkata', 'Lahore',
          'Delhi', 'Nagpur', 'Chandigarh', 'Adelaide', 'Bangalore', 'St Kitts', 'Cardiff',
          'Christchurch', 'Trinidad']

# Initialize FastAPI
app = FastAPI()

# Define the request model
class ScorePredictionRequest(BaseModel):
    batting_team: str
    bowling_team: str
    city: str
    current_score: int
    overs: float
    wickets: int
    last_five: int

# Define the response model
class ScorePredictionResponse(BaseModel):
    predicted_score: int

# Ensure input values are valid
def validate_input(data: ScorePredictionRequest):
    if data.batting_team not in teams:
        raise HTTPException(status_code=400, detail="Invalid batting team")
    if data.bowling_team not in teams:
        raise HTTPException(status_code=400, detail="Invalid bowling team")
    if data.city not in cities:
        raise HTTPException(status_code=400, detail="Invalid city")
    if not (0 <= data.overs <= 20):
        raise HTTPException(status_code=400, detail="Overs must be between 0 and 20")
    if not (0 <= data.wickets <= 10):
        raise HTTPException(status_code=400, detail="Wickets must be between 0 and 10")
    if data.current_score < 0:
        raise HTTPException(status_code=400, detail="Current score must be non-negative")
    if data.last_five < 0:
        raise HTTPException(status_code=400, detail="Runs scored in last 5 overs must be non-negative")

# Define the prediction endpoint
@app.post("/predict", response_model=ScorePredictionResponse)
def predict_score(data: ScorePredictionRequest):
    # Validate input
    validate_input(data)

    # Calculate additional features
    balls_left = 120 - (data.overs * 6)
    wickets_left = 10 - data.wickets
    crr = data.current_score / data.overs if data.overs > 0 else 0

    # Create input dataframe for the model
    input_df = pd.DataFrame(
        {'batting_team': [data.batting_team], 'bowling_team': [data.bowling_team], 'city': [data.city],
         'current_score': [data.current_score], 'balls_left': [balls_left], 'wickets_left': [wickets_left],
         'crr': [crr], 'last_five': [data.last_five]}
    )

    # Predict the score
    result = pipe.predict(input_df)
    predicted_score = int(result[0])

    # Return the prediction
    return ScorePredictionResponse(predicted_score=predicted_score)

if __name__ == "__main__":
    import uvicorn
    uvicorn.run(app, host="0.0.0.0", port=8000)