app.py

import fitz  # PyMuPDF
import re
from datasets import Dataset
from transformers import AutoModelForCausalLM, TrainingArguments, Trainer
import gradio as gr
from transformers import pipeline


def extract_text_from_pdf(pdf_path):
    """Extract text from a PDF file"""
    doc = fitz.open(pdf_path)
    text = ""
    for page in doc:
        text += page.get_text("text") + "\n"
    return text

pdf_text = extract_text_from_pdf("new-american-standard-bible.pdf")
#print(pdf_text[:1000])  # Preview first 1000 characters

def preprocess_text(text):
    """Clean and preprocess text"""
    text = re.sub(r'\s+', ' ', text)  # Remove extra whitespace
    text = text.strip()
    return text

clean_text = preprocess_text(pdf_text)
#print(clean_text[:1000])  # Preview cleaned text

# Create a dataset from text
data = {"text": [clean_text]}  # Single text entry
dataset = Dataset.from_dict(data)

# Tokenize text
from transformers import AutoTokenizer

model_name = "distilbert-base-uncased"
tokenizer = AutoTokenizer.from_pretrained(model_name)

def tokenize_function(examples):
    tokens = tokenizer(examples["text"], truncation=True, padding="max_length", max_length=512)
    tokens["labels"] = tokens["input_ids"].copy() # Use input as labels for unsupervised learning
    return tokens

tokenized_datasets = dataset.map(tokenize_function, batched=True)

model = AutoModelForCausalLM.from_pretrained(model_name)  # Adjust for task

training_args = TrainingArguments(
    output_dir="./results",
    evaluation_strategy="epoch",
    learning_rate=2e-5,
    per_device_train_batch_size=8,
    per_device_eval_batch_size=8,
    num_train_epochs=3,
    weight_decay=0.01,
    save_strategy="epoch",
)

trainer = Trainer(
    model=model,
    args=training_args,
    train_dataset=tokenized_datasets,
    eval_dataset=tokenized_datasets,
    tokenizer=tokenizer,
)

trainer.train()

model.save_pretrained("./distilbert-base-uncased-fine_tuned_model-NASB")
tokenizer.save_pretrained("./distilbert-base-uncased-fine_tuned_model-NASB")

classifier = pipeline("text-classification", model="./distilbert-base-uncased-fine_tuned_model-NASB")

def chatbot_response(text):
    return classifier(text)

iface = gr.Interface(fn=chatbot_response, inputs="text", outputs="text")
iface.launch()