import os
import sys
if '__file__' in globals():
    sys.path.append(os.path.dirname(os.path.abspath(__file__)))

import pandas as pd

import sumolib
from sumolib.net import readNet

def read_tllogic(path):
    df_tllogic = []
    for value in sumolib.xml.parse(path, "tlLogic"):
        tllogic_id = value.id
        program_id = value.programID
        for i, phase in enumerate(value.phase):
            row = {
                'tllogic_id': tllogic_id,
                'program_id': program_id,
                'index': i, 
                'duration': float(phase.duration),
                'state': phase.state,
            }
            df_tllogic.append(row)
    df_tllogic = pd.DataFrame(df_tllogic)
    return df_tllogic

def check_always_red(states):
    '''Checks whether all characters across multiple state strings at each position are 'r'.'''
    colors = ['r'] * len(states[0])
    for s in states:
        for i, char in enumerate(s.lower()):
            if char != 'r':
                colors[i] = 'G'
    return ''.join(colors)

def check_state_transitions(states):
    """Check state transitions for missing yellow signals and other specific transitions."""
    notes = []
    for i, state in enumerate(states[:-1]):
        next_state = states[i + 1]
        for j, (p0, p1) in enumerate(zip(state.lower(), next_state.lower())):
            if p0 == 'g' and p1 == 'r':
                notes.append(f'녹색 -> 적색, 황색신호 누락. {i}, {i + 1} 사이의 state에서 {j}번째 문자를 확인하세요.')
            if p0 == 'r' and p1 == 'y':
                notes.append(f'적색 -> 황색. {i}, {i + 1} 사이의 state에서 {j}번째 문자를 확인하세요.')
            if p0 == 'y' and p1 == 'g':
                notes.append(f'황색 -> 녹색. {i}, {i + 1} 사이의 state에서 {j}번째 문자를 확인하세요.')
    return notes

def main(path_tll_xml, path_network_xml, interval):
    df_tllogic = read_tllogic(path_tll_xml)
    net = readNet(path_network_xml, withInternal=True, withPrograms=True)   

    review_needed = []

    # duration의 합계 비교
    df_tmp = df_tllogic \
        .groupby(['tllogic_id', 'program_id']) \
        .agg(sum_duration=('duration', 'sum')) \
        .reset_index() \
        .query('sum_duration < @interval')
    if not df_tmp.empty:
        for tup in df_tmp.itertuples():
            sub = {'note': f'duration의 합계가 {interval}보다 작음. tllogic_id: {tup.tllogic_id}, program_id: {tup.program_id}'}
            review_needed.append(sub)

    for (tllogic_id, program_id), group in df_tllogic.groupby(['tllogic_id', 'program_id']):
        group_sorted = group.sort_values(by='index')
        states = list(group_sorted['state'])

        # 항상 적색인 경우
        merged_state = check_always_red(states)
        if 'r' in merged_state:
            sub = {'note': f'There is always a red signal. state: {merged_state}. tllogic_id: {tllogic_id}, program_id: {program_id}'}
            review_needed.append(sub)

        # 신호 색상 순서 확인
        for note in check_state_transitions(states):
            note = note + f' tllogic_id: {tllogic_id}, program_id: {program_id}'
            review_needed.append({'note': note})

    review_needed = pd.DataFrame(review_needed)

    return review_needed

if __name__ == '__main__':
    file_names = os.listdir(os.path.dirname(os.path.abspath(__file__)))
    path_tll_xml = file_names[3]
    path_network_xml = file_names[1]
    print(path_tll_xml, path_network_xml)
    interval = 300
    review_needed = main(path_tll_xml, path_network_xml, interval)
    review_needed.to_csv('review_needed.csv', index=False, encoding='cp949')