src/document_processor.py

#!/usr/bin/env python3
"""
Dynamic RAG Database Updater
Processes PDFs in memory and updates the vector database.
No PDFs, OCR text, or intermediate files are stored.
"""

import numpy as np
from pathlib import Path
from typing import List, Dict
import pickle
from datetime import datetime

# PDF processing
from pdf2image import convert_from_path

# OCR
import pytesseract

# Embeddings
from sentence_transformers import SentenceTransformer

# FAISS
import faiss


class DynamicRAGUpdater:
    """
    Handles dynamic updates to RAG database:
    1. PDF upload (temporary path)
    2. OCR extraction (in memory)
    3. Chunking
    4. Embedding generation
    5. FAISS + metadata update ONLY
    """

    def __init__(
        self,
        vector_db_path: str,
        embedding_model: str = "microsoft/BiomedNLP-BiomedBERT-base-uncased-abstract-fulltext",
    ):
        self.vector_db_path = Path(vector_db_path)

        print("Using Tesseract OCR (in-memory only)")

        self.embedding_model = SentenceTransformer(
            embedding_model,
            device="cpu",
        )

        self.embedding_dim = self.embedding_model.get_sentence_embedding_dimension()
        self.load_database()

    # --------------------------------------------------
    # Load / Save database
    # --------------------------------------------------
    def load_database(self):
        index_file = self.vector_db_path / "faiss.index"
        metadata_file = self.vector_db_path / "metadata.pkl"

        self.faiss_index = faiss.read_index(str(index_file))

        with open(metadata_file, "rb") as f:
            data = pickle.load(f)
            self.chunks = data["chunks"]
            self.chunk_id_to_idx = data["chunk_id_to_idx"]

    def save_database(self):
        faiss.write_index(self.faiss_index, str(self.vector_db_path / "faiss.index"))

        with open(self.vector_db_path / "metadata.pkl", "wb") as f:
            pickle.dump(
                {
                    "chunks": self.chunks,
                    "chunk_id_to_idx": self.chunk_id_to_idx,
                    "embedding_dim": self.embedding_dim,
                    "model": "microsoft/BiomedNLP-BiomedBERT-base-uncased-abstract-fulltext",
                },
                f,
            )

    # --------------------------------------------------
    # OCR (in memory)
    # --------------------------------------------------
    def extract_text_from_pdf(self, pdf_path: str) -> str:
        try:
            images = convert_from_path(pdf_path, dpi=300)
        except Exception as e:
            raise RuntimeError(
                "PDF conversion failed. Ensure Poppler is installed."
            ) from e

        pages = []
        for page_num, image in enumerate(images, 1):
            text = pytesseract.image_to_string(
                image,
                lang="eng",
                config="--oem 3 --psm 6",
            )

            pages.append(
                f"\n{'=' * 40}\nPAGE {page_num}\n{'=' * 40}\n{text}"
            )

        return "\n".join(pages)

    # --------------------------------------------------
    # Chunking
    # --------------------------------------------------
    def chunk_text(self, text: str, chunk_size: int = 512) -> List[str]:
        sentences = text.split(". ")
        chunks, current, length = [], [], 0

        for s in sentences:
            s = s.strip()
            if not s:
                continue

            s += ". "
            if length + len(s) > chunk_size and current:
                chunks.append("".join(current))
                current = [s]
                length = len(s)
            else:
                current.append(s)
                length += len(s)

        if current:
            chunks.append("".join(current))

        return chunks

    # --------------------------------------------------
    # Embeddings
    # --------------------------------------------------
    def generate_embeddings(self, chunks: List[str]) -> np.ndarray:
        return self.embedding_model.encode(
            chunks,
            batch_size=32,
            convert_to_numpy=True,
            show_progress_bar=True,
        )

    # --------------------------------------------------
    # FAISS update
    # --------------------------------------------------
    def add_to_database(
        self,
        embeddings: np.ndarray,
        chunks: List[str],
        filename: str,
    ) -> int:
        start_idx = self.faiss_index.ntotal
        self.faiss_index.add(embeddings.astype("float32"))

        for i, text in enumerate(chunks):
            meta = {
                "chunk_id": start_idx + i,
                "text": text,
                "filename": filename,
                "upload_date": datetime.now().isoformat(),
                "source": "user_upload",
            }
            self.chunks.append(meta)
            self.chunk_id_to_idx[f"{filename}_{i}"] = start_idx + i

        return len(embeddings)

    # --------------------------------------------------
    # Full pipeline (NO FILE STORAGE)
    # --------------------------------------------------
    def process_and_add_pdf(self, pdf_path: str) -> Dict:
        start = datetime.now()
        filename = Path(pdf_path).stem

        # All steps are in memory
        text = self.extract_text_from_pdf(pdf_path)
        chunks = self.chunk_text(text)
        embeddings = self.generate_embeddings(chunks)

        vectors_added = self.add_to_database(embeddings, chunks, filename)
        self.save_database()

        return {
            "filename": filename,
            "num_chunks": len(chunks),
            "vectors_added": vectors_added,
            "total_vectors": self.faiss_index.ntotal,
            "processing_time_seconds": (datetime.now() - start).total_seconds(),
            "timestamp": datetime.now().isoformat(),
        }