import tensorflow as tf
from transformers import DistilBertTokenizer, TFDistilBertForSequenceClassification
from huggingface_hub import snapshot_download
import joblib
import os

import os
from transformers import TFDistilBertForSequenceClassification

print("📁 Current working directory:", os.getcwd())
print("📄 Checking model path:", os.listdir("nlp_intent/intent_model") if os.path.exists("nlp_intent/intent_model") else "❌ not found")

try:
    model = TFDistilBertForSequenceClassification.from_pretrained("nlp_intent/intent_model")
    print("✅ Model loaded successfully!")
except Exception as e:
    print("🚨 Model load failed:", e)




repo_path = snapshot_download(
    repo_id="samithcs/nlp_intent_model",
    repo_type="model"
)



MODEL_PATH = os.path.join(repo_path, "nlp_intent", "intent_model")
TOKENIZER_PATH = os.path.join(repo_path, "nlp_intent", "intent_tokenizer")
LABEL_PATH = os.path.join(repo_path, "nlp_intent", "label_encoder.joblib")


model = TFDistilBertForSequenceClassification.from_pretrained(MODEL_PATH)
tokenizer = DistilBertTokenizer.from_pretrained(TOKENIZER_PATH)
label_encoder = joblib.load(LABEL_PATH)

print("Model loaded successfully!")


def predict_intent(text: str) -> dict:
    inputs = tokenizer(text, return_tensors="tf", truncation=True, padding=True, max_length=128)
    outputs = model(inputs)
    predicted_class = tf.argmax(outputs.logits, axis=1).numpy()[0]
    intent = label_encoder.inverse_transform([predicted_class])[0]
    confidence = float(tf.nn.softmax(outputs.logits)[0][predicted_class].numpy())
    
    return {"intent": intent, "confidence": confidence}