import json
from datetime import datetime

from app import config
from app.logger import log


def get_latest_cdc_event(snowflake, playlist_id):
    log.info("Getting latest CDC event")

    # Get latest CDC event log from Snowflake DB
    data_frame = snowflake.execute_query('spotify_playlists_processing_log', playlist_id=playlist_id)

    if data_frame.empty:
        log.warning(f"No CDC events found for playlist {playlist_id}")
        return None

    cdc_event = data_frame.iloc[0]
    return cdc_event


def get_latest_cdc_event_timestamp(snowflake, playlist_id):
    log.info("Getting latest CDC event timestamp")
    cdc_event = get_latest_cdc_event(snowflake, playlist_id)
    if cdc_event is None:
        raise ValueError(f"No CDC event found for playlist {playlist_id}")
    latest_cdc_timestamp = json.loads(cdc_event['UPDATE_SENT'])[0]
    return datetime.strptime(latest_cdc_timestamp.replace(" Z", " +0000"), "%a, %d %b %Y %H:%M:%S %z")


PLAYLIST_ID_CHUNK_SIZE = 1000


def get_ingested_playlist_ids(snowflake, store_id, playlist_ids):
    log.info(f"Getting playlist ids ingested for store {store_id} from v_playlist_metadata")
    ingested_ids = set()
    for i in range(0, len(playlist_ids), PLAYLIST_ID_CHUNK_SIZE):
        chunk = playlist_ids[i:i + PLAYLIST_ID_CHUNK_SIZE]
        placeholders = ", ".join(["%s"] * len(chunk))
        data_frame = snowflake.execute_query(
            'playlist_metadata_ids', params=chunk, store_id=str(store_id), placeholders=placeholders, env=config.ENV)
        # Normalize to str since callers diff this set against plain CSV-loaded strings.
        ingested_ids.update(str(pid).strip() for pid in data_frame['STORE_PLAYLIST_ID'])
    return ingested_ids