#!/usr/bin/env python3

import sys
import typing

import pandas


def _merge_adjacent(
    recordings: pandas.DataFrame, remove_tags: typing.Set[str]
) -> typing.Iterator[pandas.Series]:
    previous_recording: typing.Optional[pandas.Series] = None
    for _, next_recording in recordings.sort_values("start").iterrows():
        next_recording["tags"] = set(next_recording["tags"]).difference(remove_tags)
        if previous_recording is not None:
            if next_recording["start"] <= previous_recording["end"]:
                previous_recording["end"] = next_recording["end"]
                previous_recording["tags"].update(next_recording["tags"])
            else:
                yield previous_recording
                previous_recording = next_recording
        else:
            previous_recording = next_recording
    if previous_recording is not None:
        yield previous_recording


def _main():
    pandas.set_option("display.width", None)
    pandas.set_option("display.max_rows", None)
    # https://timewarrior.net/docs/api/
    config_text, recordings_json = sys.stdin.read().split("\n\n", maxsplit=1)
    config = dict(o.split(": ", maxsplit=1) for o in config_text.splitlines())
    selected_tags = set(config["temp.report.tags"].split(","))
    print(selected_tags)
    recordings = pandas.DataFrame(
        _merge_adjacent(
            pandas.read_json(recordings_json, convert_dates=["start", "end"]).set_index(
                "id", verify_integrity=True
            ),
            remove_tags=selected_tags,
        )
    ).assign(
        start=lambda df: df["start"]
        .dt.tz_convert("Europe/Vienna")
        .dt.strftime("%Y-%m-%d %H:%M"),
        end=lambda df: df["end"].dt.tz_convert("Europe/Vienna").dt.strftime("%H:%M"),
    )
    print(recordings)


if __name__ == "__main__":
    _main()