modules/llm/summarizer/helpers/db_helper.py

import uuid
from config import SanatanConfig
from db import SanatanDatabase
from modules.llm.summarizer.models import ScriptureRequest
import logging

logging.basicConfig()
logger = logging.getLogger(__name__)
logger.setLevel(logging.INFO)


async def get_scripture_from_db(req: ScriptureRequest):
    """
    Return a scripture unit (page or verse, based on config),
    including all metadata fields separately.
    Used for page view to fetch by global index.
    """
    # ensure we have a valid request id (deviceId)
    request_id = req.request_id or f"auto-{uuid.uuid4()}"
    logger.info(
        f"get_scripture: received requestId={request_id}, scripture={req.scripture_name}, unit_index={req.unit_index}"
    )

    # find config entry for the scripture
    config = next(
        (s for s in SanatanConfig().scriptures if s["name"] == req.scripture_name), None
    )
    if not config:
        return {
            "error": f"Scripture '{req.scripture_name}' not found",
            "requestId": request_id,
        }

    # fetch the raw document from DB
    raw_doc = SanatanDatabase().fetch_document_by_index(
        collection_name=config["collection_name"],
        index=req.unit_index,
    )

    if not raw_doc or isinstance(raw_doc, str) or "error" in raw_doc:
        return {
            "error": f"No data available for unit {req.unit_index}",
            "requestId": request_id,
        }

    # canonicalize it
    canonical_doc = SanatanConfig().canonicalize_document(
        scripture_name=req.scripture_name,
        document_text=raw_doc.get("document", ""),
        metadata_doc=raw_doc,
    )

    # add unit index & total units (so Flutter can paginate)
    canonical_doc["total"] = SanatanDatabase().count(config["collection_name"])
    canonical_doc["requestId"] = request_id

    return canonical_doc