import asyncio
from pathlib import Path
from data_access import Listing
import pandas as pd
from rec.query import QueryParameters, filter_listings


async def export_to_csv(
    listings: list[Listing],
    output_file: Path,
    columns: list[str],
    query_parameters: QueryParameters | None = None,
) -> None:
    if query_parameters is not None:
        listings = await filter_listings(listings, query_parameters)
    ds = await asyncio.gather(*[listing.dict_nicely() for listing in listings])
    df = pd.DataFrame(ds)
    # read decisions on file
    decisions_path = "data/decisions.json"
    decisions = pd.read_json(decisions_path)
    df.loc[:, "decision"] = df.identifier.apply(lambda x: decisions.get(x))

    # remove all entries where we didnt calculate transit time (probably due to a too far distance)
    # df2 = df[df.travel_time_fastest.notna()]
    df2 = df

    # drop columns
    # dropcolumns = ['distance_per_transit', 'duration_static', 'distance']
    # s1 = df2['travel_time_fastest'].apply(pd.Series).drop(dropcolumns, axis=1)
    # s1 = df2

    # fill in gap values for service charge and lease left. This is for excel so we can use filters better there
    df2.loc[:, "service_charge"] = df2.service_charge.fillna(-1)
    df2.loc[:, "lease_left"] = df2.lease_left.fillna(-1)
    df2.loc[:, "sqm_ocr"] = df2.sqm_ocr.fillna(-1)

    df3 = df2
    # df3 = pd.concat([df2.drop(['travel_time_fastest', 'travel_time_second'], axis=1), s1], axis=1)
    # df3.loc[:, 'duration'] = (df3.loc[:, ['duration']].min(axis=1) / 60).round()
    df3.shape
    df4 = df3

    df5 = df4[columns]
    df6 = df5.sort_values(by=["price_per_sqm"], ascending=True)
    df6.to_csv(str(output_file), index=False)