app.py

from fastapi import FastAPI
from pydantic import BaseModel
import joblib
import nltk
import os
import numpy as np

# Set NLTK data directory
nltk_data_path = "/app/nltk_data"
os.environ["NLTK_DATA"] = nltk_data_path
nltk.data.path.append(nltk_data_path)

# Load ML model and transformers
model = joblib.load("model.joblib")
tfidf_vectorizer = joblib.load("tfidf_vectorizer.joblib")
le = joblib.load("labelencoder.joblib")

class TextInput(BaseModel):
    text: str

app = FastAPI()

@app.post("/predict")
def predict(input: TextInput):
    processed_text = preprocess_text(input.text)
    text_tfidf = tfidf_vectorizer.transform([processed_text]).toarray()
    y_pred = model.predict(text_tfidf)[0]
    category_name = le.inverse_transform([y_pred])[0]
    probabilities = model.predict_proba(text_tfidf)[0]

    if np.max(probabilities>0.5) :
      return category_name  
    else :
      return "Quite ambigous but maybe "+category_name 
    

def preprocess_text(text):
    import re
    from nltk.stem import WordNetLemmatizer
    from nltk.corpus import stopwords
    
    lemmatizer = WordNetLemmatizer()
    stop_words = set(stopwords.words('english'))
    
    text = re.sub('[^a-zA-Z]', ' ', text).lower()
    words = [lemmatizer.lemmatize(word) for word in text.split() if word not in stop_words]
    
    return ' '.join(words)