document-classifier-xgb

import joblib

# Load the model bundle
bundle = joblib.load("document_classifier_xgb.pkl")
model            = bundle["model"]
word_vectorizer  = bundle["word_vectorizer"]
char_vectorizer  = bundle["char_vectorizer"]
scaler           = bundle["scaler"]

from scipy.sparse import hstack, csr_matrix
import numpy as np

def predict(text: str) -> int:
    word_feat = word_vectorizer.transform([text])
    char_feat = char_vectorizer.transform([text])
    num_feat  = scaler.transform([[
        len(text),                          # char_count
        sum(c.isdigit() for c in text),     # digit_count
        sum(c.isupper() for c in text),     # uppercase_count
        text.count("$") + text.count("£"),  # currency_count
        text.count("\n"),                   # line_count
    ]])
    features = hstack([word_feat, char_feat, csr_matrix(num_feat)])
    return int(model.predict(features)[0])

label = predict("Invoice No. 12345  Total: $499.99  Date: 01/01/2024")
print("Predicted label:", label)