{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import os\n",
    "import sumolib\n",
    "import random\n",
    "from tqdm import tqdm\n",
    "from datetime import datetime"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# A. 이동류 매칭"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "def make_match1():\n",
    "    '''\n",
    "    신호 DB에는 매 초마다 이동류정보가 업데이트 된다. 그리고 이 이동류정보를 매 5초마다 불러와서 사용하게 된다.\n",
    "    '../Data/tables/move/'에는 5초마다의 이동류정보가 저장되어 있다.\n",
    "\n",
    "    return : 통합된 이동류정보\n",
    "     - 모든 inter_no(교차로번호)에 대한 A, B링 현시별 이동류정보\n",
    "\n",
    "    match1을 만드는 데 시간이 소요되므로 한 번 만들어서 저장해두고 저장해둔 것을 쓴다.\n",
    "    '''\n",
    "    # [이동류번호] 불러오기 (약 1분의 소요시간)\n",
    "    path_move = '../Data/tables/move/'\n",
    "    csv_moves = os.listdir(path_move)\n",
    "    moves = [pd.read_csv(path_move + csv_move, index_col=0) for csv_move in tqdm(csv_moves)]\n",
    "    match1 = pd.concat(moves).drop_duplicates().sort_values(by=['inter_no','phas_A','phas_B']).reset_index(drop=True)\n",
    "    match1.to_csv('../Intermediates/match1.csv')\n",
    "    return match1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "def make_match2(match1):\n",
    "    '''\n",
    "    match1을 계층화함.\n",
    "     - match1의 컬럼 : inter_no, phas_A, phas_B, move_A, move_B\n",
    "     - match2의 컬럼 : inter_no, phase_no, ring_type, move_no\n",
    "    '''\n",
    "    # 계층화 (inter_no, phas_A, phas_B, move_A, move_B) -> ('inter_no', 'phase_no', 'ring_type', 'move_no')\n",
    "    matchA = match1[['inter_no', 'phas_A', 'move_A']].copy()\n",
    "    matchA.columns = ['inter_no', 'phase_no', 'move_no']\n",
    "    matchA['ring_type'] = 'A'\n",
    "    matchB = match1[['inter_no', 'phas_B', 'move_B']].copy()\n",
    "    matchB.columns = ['inter_no', 'phase_no', 'move_no']\n",
    "    matchB['ring_type'] = 'B'\n",
    "    match2 = pd.concat([matchA, matchB]).drop_duplicates()\n",
    "    match2 = match2[['inter_no', 'phase_no', 'ring_type', 'move_no']]\n",
    "    match2 = match2.sort_values(by=list(match2.columns))\n",
    "    return match2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "def make_match3(match2):\n",
    "    '''\n",
    "    각 movement들에 방향(진입방향, 진출방향)을 매칭시켜 추가함.\n",
    "     - match2의 컬럼 : inter_no, phase_no, ring_type, move_no\n",
    "     - match3의 컬럼 : inter_no, phase_no, ring_type, move_no, inc_dir, out_dir\n",
    "\n",
    "    nema : \n",
    "     - 컬럼 : move_no, inc_dir, out_dir\n",
    "     - 모든 종류의 이동류번호에 대하여 진입방향과 진출방향을 매칭시키는 테이블\n",
    "     - 이동류번호 : 1 ~ 16, 17, 18, 21\n",
    "     - 진입, 진출방향(8방위) : 동, 서, 남, 북, 북동, 북서, 남동, 남서\n",
    "    '''\n",
    "    # nema 정보 불러오기 및 병합\n",
    "    nema = pd.read_csv('../Data/tables/nema.csv', encoding='cp949')\n",
    "    match3 = pd.merge(match2, nema, how='left', on='move_no').drop_duplicates()\n",
    "    return match3"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "def make_match4(match3):\n",
    "    '''\n",
    "    방위각 정보를 매칭시켜 추가함.\n",
    "     - match3의 컬럼 : inter_no, phase_no, ring_type, move_no, inc_dir, out_dir\n",
    "     - match4의 컬럼 : inter_no, phase_no, ring_type, move_no, inc_dir, out_dir, inc_angle, out_angle\n",
    "\n",
    "    angle_original : \n",
    "     - 컬럼 : inter_no, angle_Aj, angle_Bj (j : 1 ~ 8)\n",
    "     - 모든 종류의 이동류번호에 대하여 진입방향과 진출방향을 매칭시키는 테이블\n",
    "     - 이동류번호 : 1 ~ 16, 17, 18, 21\n",
    "     - 진입, 진출방향(8방위) : 동, 서, 남, 북, 북동, 북서, 남동, 남서\n",
    "    '''\n",
    "\n",
    "    # 방위각 정보 불러오기\n",
    "    dtype_dict = {f'angle_{alph}{j}':'str' for alph in ['A', 'B'] for j in range(1,9)}\n",
    "    angle_original = pd.read_csv('../Data/tables/angle.csv', index_col=0, dtype = dtype_dict)\n",
    "\n",
    "    # 계층화\n",
    "    angle = []\n",
    "    for i, row in angle_original.iterrows():\n",
    "        angle_codes = row[[f'angle_{alph}{j}' for alph in ['A', 'B'] for j in range(1,9)]]\n",
    "        new = pd.DataFrame({'inter_no':[row.inter_no] * 16, 'phase_no':list(range(1, 9))*2, 'ring_type':['A'] * 8 + ['B'] * 8, 'angle_code':angle_codes.to_list()})\n",
    "        angle.append(new)\n",
    "    angle = pd.concat(angle)\n",
    "    angle = angle.dropna().reset_index(drop=True)\n",
    "\n",
    "    # 병합\n",
    "    six_chars = angle.angle_code.apply(lambda x:len(x)==6)\n",
    "    angle.loc[six_chars,'inc_angle'] = angle.angle_code.apply(lambda x:x[:3])\n",
    "    angle.loc[six_chars,'out_angle'] = angle.angle_code.apply(lambda x:x[3:])\n",
    "    angle = angle.drop('angle_code', axis=1)\n",
    "    match4 = pd.merge(match3, angle, how='left', left_on=['inter_no', 'phase_no', 'ring_type'],\n",
    "                    right_on=['inter_no', 'phase_no', 'ring_type']).drop_duplicates()\n",
    "    return match4"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "def make_match5(match4):\n",
    "    '''\n",
    "    진입엣지id, 진출엣지id, 노드id를 추가함 (주교차로).\n",
    "     - match4의 컬럼 : inter_no, phase_no, ring_type, move_no, inc_dir, out_dir, inc_angle, out_angle\n",
    "     - match5의 컬럼 : inter_no, phase_no, ring_type, move_no, inc_dir, out_dir, inc_angle, out_angle, inc_edge, out_edge, node_id\n",
    "    \n",
    "    사용된 데이터 : \n",
    "    (1) net\n",
    "     - 성남시 정자동 부근의 샘플 네트워크\n",
    "    (2) inter_node\n",
    "     - 교차로번호와 노드id를 매칭시키는 테이블.\n",
    "     - parent/child 정보도 포함되어 있음\n",
    "     - 컬럼 : inter_no, node_id, inter_type\n",
    "    (3) inter_info\n",
    "     - 교차로 정보. 여기에서는 위도와 경도가 쓰임.\n",
    "     - 컬럼 : inter_no, inter_name, inter_lat, inter_lon, group_no, main_phase_no\n",
    "\n",
    "    진입엣지id, 진출엣지id를 얻는 과정 :\n",
    "     - match5 = match4.copy()의 각 열을 순회하면서 아래 과정을 반복함.\n",
    "     * 진입에 대해서만 서술하겠지만 진출도 마찬가지로 설명될 수 있음\n",
    "     - 해당 행의 교차로정보로부터 노드ID를 얻어내고, 해당 노드에 대한 모든 진출엣지id를 inc_edges에 저장.\n",
    "     * inc_edge(진입엣지) : incoming edge, out_edge(진출엣지) : outgoing_edge\n",
    "     - inc_edges의 모든 진입엣지에 대하여 진입방향(inc_dires, 2차원 단위벡터)을 얻어냄.\n",
    "     - 해당 행의 진입각으로부터 그에 대응되는 진입각방향(단위벡터)를 얻어냄.\n",
    "     - 주어진 진입각방향에 대하여 내적이 가장 작은 진입방향에 대한 진입엣지를 inc_edge_id로 지정함.\n",
    "    '''\n",
    "\n",
    "    # 네트워크 불러오기 \n",
    "    net = sumolib.net.readNet('../Data/networks/sn.net.xml')\n",
    "    # 교차로-노드 매칭 정보 불러오기\n",
    "    inter_node = pd.read_csv('../Data/tables/inter_node.csv', index_col=0)\n",
    "    # 교차로정보(위, 경도) 불러오기\n",
    "    inter_info = pd.read_csv('../Data/tables/inter_info.csv', index_col=0)\n",
    "\n",
    "    # parent node만 가져옴.\n",
    "    inter_node1 = inter_node[inter_node.inter_type == 'parent'].drop('inter_type', axis=1)\n",
    "    inter_info1 = inter_info[['inter_no', 'inter_lat', 'inter_lon']]\n",
    "    inter = pd.merge(inter_node1, inter_info1, how='left', left_on=['inter_no'],\n",
    "                    right_on=['inter_no']).drop_duplicates()\n",
    "\n",
    "    inter2node = dict(zip(inter['inter_no'], inter['node_id']))\n",
    "\n",
    "    match5 = match4.copy()\n",
    "    # 진입진출ID 매칭\n",
    "    for index, row in match5.iterrows():\n",
    "        node_id = inter2node[row.inter_no]\n",
    "        node = net.getNode(node_id)\n",
    "        # 교차로의 모든 (from / to) edges\n",
    "        inc_edges = [edge for edge in node.getIncoming() if edge.getFunction() == ''] # incoming edges\n",
    "        out_edges = [edge for edge in node.getOutgoing() if edge.getFunction() == ''] # outgoing edges\n",
    "        # 교차로의 모든 (from / to) directions\n",
    "        inc_dirs = []\n",
    "        for inc_edge in inc_edges:\n",
    "            start = inc_edge.getShape()[-2]\n",
    "            end = inc_edge.getShape()[-1]\n",
    "            inc_dir = np.array(end) - np.array(start)\n",
    "            inc_dir = inc_dir / (inc_dir ** 2).sum() ** 0.5\n",
    "            inc_dirs.append(inc_dir)\n",
    "        out_dirs = []\n",
    "        for out_edge in out_edges:\n",
    "            start = out_edge.getShape()[0]\n",
    "            end = out_edge.getShape()[1]\n",
    "            out_dir = np.array(end) - np.array(start)\n",
    "            out_dir = out_dir / (out_dir ** 2).sum() ** 0.5\n",
    "            out_dirs.append(out_dir)\n",
    "        # 진입각, 진출각 불러오기\n",
    "        if not pd.isna(row.inc_angle):\n",
    "            inc_angle = int(row.inc_angle)\n",
    "            out_angle = int(row.out_angle)\n",
    "            # 방위각을 일반각으로 가공, 라디안 변환, 단위벡터로 변환\n",
    "            inc_angle = (-90 - inc_angle) % 360\n",
    "            inc_angle = inc_angle * np.pi / 180.\n",
    "            inc_dir_true = np.array([np.cos(inc_angle), np.sin(inc_angle)])\n",
    "            out_angle = (90 - out_angle) % 360\n",
    "            out_angle = out_angle * np.pi / 180.\n",
    "            out_dir_true = np.array([np.cos(out_angle), np.sin(out_angle)])\n",
    "            # 매칭 엣지 반환\n",
    "            inc_index = np.array([np.dot(inc_dir, inc_dir_true) for inc_dir in inc_dirs]).argmax()\n",
    "            out_index = np.array([np.dot(out_dir, out_dir_true) for out_dir in out_dirs]).argmax()\n",
    "            inc_edge_id = inc_edges[inc_index].getID()\n",
    "            out_edge_id = out_edges[out_index].getID()\n",
    "            match5.at[index, 'inc_edge'] = inc_edge_id\n",
    "            match5.at[index, 'out_edge'] = out_edge_id\n",
    "    match5['node_id'] = match5['inter_no'].map(inter2node)\n",
    "    match5 = match5.sort_values(by=['inter_no','phase_no','ring_type']).reset_index(drop=True)\n",
    "    return match5"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "def make_match6(match5):\n",
    "    '''\n",
    "    진입엣지id, 진출엣지id, 노드id를 추가함 (부교차로).\n",
    "     - match6의 컬럼 : inter_no, phase_no, ring_type, move_no, inc_dir, out_dir, inc_angle, out_angle, inc_edge, out_edge, node_id\n",
    "    \n",
    "    사용된 데이터 : \n",
    "    (1) inter_node\n",
    "     - 교차로번호와 노드id를 매칭시키는 테이블.\n",
    "     - parent/child 정보도 포함되어 있음\n",
    "     - 컬럼 : inter_no, node_id, inter_type\n",
    "    (2) uturn (유턴정보)\n",
    "     - 컬럼 : parent_id, child_id, direction, condition, inc_edge, out_edge\n",
    "     - parent_id, child_id : 주교차로id, 유턴교차로id\n",
    "     - direction : 주교차로에 대한 유턴노드의 상대적인 위치(방향)\n",
    "     - condition : 좌회전시, 직진시, 직좌시, 보행신호시 중 하나\n",
    "     - inc_edge, out_edge : 유턴에 대한 진입진출엣지\n",
    "    (3) coord (연동교차로정보)\n",
    "     - 컬럼 : parent_id, child_id, phase_no, ring_type, inc_edge, out_edge\n",
    "     - parent_id, child_id : 주교차로id, 연동교차로id\n",
    "     - 나머지 컬럼 : 각 (현시, 링)별 진입진출엣지\n",
    "\n",
    "    설명 :\n",
    "     - match5는 주교차로에 대해서만 진입엣지id, 진출엣지id, 노드id를 추가했었음.\n",
    "       여기에서 uturn, coord를 사용해서 부교차로들(유턴교차로, 연동교차로)에 대해서도 해당 값들을 부여함.\n",
    "    유턴교차로 :\n",
    "     - directions를 정북기준 시계방향의 8방위로 정함.\n",
    "        - 이를 통해 진입방향이 주어진 경우에 좌회전, 직진, 보행 등에 대한 (진입방향, 진출방향)을 얻어낼 수 있음.\n",
    "        - 예) 진입방향(direction)이 '북'일 때, \n",
    "        - 직진 : (북, 남)\n",
    "            * 남 : directions[(ind + 4) % len(directions)]\n",
    "        - 좌회전 : (북, 동)\n",
    "            * 동 : directions[(ind + 2) % len(directions)]\n",
    "        - 보행 : (서, 동)\n",
    "            * 서 : directions[(ind - 2) % len(directions)]\n",
    "     - uturn의 각 행을 순회하면서 아래 과정을 반복함\n",
    "        - match5에서 parent_id에 해당하는 행들을 가져옴(cmatch).\n",
    "        - condition 별로 진입방향, 진출방향A, 진출방향B 정함.\n",
    "            - 상술한 directions를 활용하여 정함.\n",
    "        - (진입방향, 진출방향A, 진출방향B)을 고려하여 (현시, 링) 별로 진입엣지id, 진출엣지id를 정함.\n",
    "            - ex) cmatch.loc[(cmatch.inc_dir==inc_dire) & (cmatch.out_dir==out_dire_A), ['inc_edge', 'out_edge']] = [inc_edge_id, out_edge_id]\n",
    "     - 순회하면서 만든 cmatch를 cmatchs라는 리스트에 저장함.\n",
    "\n",
    "    연동교차로 :\n",
    "     - 연동교차로의 경우 coord에 (현시, 링)별 진입엣지ID, 진출엣지ID가 명시되어 있음.\n",
    "     - 'inc_dir', 'out_dir', 'inc_angle','out_angle'와 같은 열들은 np.nan을 지정해놓음.\n",
    "     - 이 열들은, 사실상 다음 스텝부터는 사용되지 않는 열들이기 때문에 np.nan으로 지정해놓아도 문제없음.\n",
    "\n",
    "    match6 :\n",
    "     - 이렇게 얻은 match5, cmatchs, coord를 모두 pd.concat하여 match6을 얻어냄.\n",
    "    '''\n",
    "\n",
    "    inter_node = pd.read_csv('../Data/tables/inter_node.csv', index_col=0)\n",
    "    node2inter = dict(zip(inter_node['node_id'], inter_node['inter_no']))\n",
    "\n",
    "    uturn = pd.read_csv('../Data/tables/child_uturn.csv')\n",
    "    coord = pd.read_csv('../Data/tables/child_coord.csv')\n",
    "    child_ids = inter_node[inter_node.inter_type=='child'].node_id.unique()\n",
    "    ch2pa = {} # child to parent\n",
    "    for child_id in child_ids:\n",
    "        parent_no = inter_node[inter_node.node_id==child_id].inter_no.iloc[0]\n",
    "        sub_inter_node = inter_node[inter_node.inter_no==parent_no]\n",
    "        ch2pa[child_id] = sub_inter_node[sub_inter_node.inter_type=='parent'].iloc[0].node_id\n",
    "    directions = ['북', '북동', '동', '남동', '남', '남서', '서', '북서'] # 정북기준 시계방향으로 8방향\n",
    "\n",
    "    # 각 uturn node에 대하여 (inc_edge_id, out_edge_id) 부여\n",
    "    cmatches = []\n",
    "    for _, row in uturn.iterrows():\n",
    "        child_id = row.child_id\n",
    "        parent_id = row.parent_id\n",
    "        direction = row.direction\n",
    "        condition = row.condition\n",
    "        inc_edge_id = row.inc_edge\n",
    "        out_edge_id = row.out_edge\n",
    "        # match5에서 parent_id에 해당하는 행들을 가져옴\n",
    "        cmatch = match5.copy()[match5.node_id==parent_id] # match dataframe for a child node\n",
    "        cmatch = cmatch.sort_values(by=['phase_no', 'ring_type']).reset_index(drop=True)\n",
    "        cmatch['node_id'] = child_id\n",
    "        cmatch[['inc_edge', 'out_edge']] = np.nan\n",
    "\n",
    "        # condition 별로 inc_dire, out_dire_A, out_dire_B를 정함\n",
    "        ind = directions.index(direction)\n",
    "        if condition == \"좌회전시\":\n",
    "            inc_dire = direction\n",
    "            out_dire_A = out_dire_B = directions[(ind + 2) % len(directions)]\n",
    "        elif condition == \"직진시\":\n",
    "            inc_dire = direction\n",
    "            out_dire_A = out_dire_B = directions[(ind + 4) % len(directions)]\n",
    "        elif condition == \"보행신호시\":\n",
    "            inc_dire = directions[(ind + 2) % len(directions)]\n",
    "            out_dire_A = directions[(ind - 2) % len(directions)]\n",
    "            out_dire_B = directions[(ind - 2) % len(directions)]\n",
    "\n",
    "        # (inc_dire, out_dire_A, out_dire_B) 별로 inc_edge_id, out_edge_id를 정함\n",
    "        if condition == '보행신호시':\n",
    "            cmatch.loc[(cmatch.inc_dir==inc_dire) & (cmatch.out_dir==out_dire_A), ['inc_edge', 'out_edge']] = [inc_edge_id, out_edge_id]\n",
    "            cmatch.loc[(cmatch.inc_dir==inc_dire) & (cmatch.out_dir==out_dire_B), ['inc_edge', 'out_edge']] = [inc_edge_id, out_edge_id]\n",
    "            # 이동류번호가 17(보행신호)이면서 유턴노드방향으로 가는 신호가 없으면 (inc_edge_id, out_edge_id)를 부여한다.\n",
    "            cmatch.loc[(cmatch.move_no==17) & (cmatch.out_dir!=direction), ['inc_edge', 'out_edge']] = [inc_edge_id, out_edge_id]\n",
    "        else: # '직진시', '좌회전시'\n",
    "            cmatch.loc[(cmatch.inc_dir==inc_dire) & (cmatch.out_dir==out_dire_A), ['inc_edge', 'out_edge']] = [inc_edge_id, out_edge_id]\n",
    "            cmatch.loc[(cmatch.inc_dir==inc_dire) & (cmatch.out_dir==out_dire_B), ['inc_edge', 'out_edge']] = [inc_edge_id, out_edge_id]\n",
    "        # 유턴신호의 이동류번호를 19로 부여한다.\n",
    "        cmatch.loc[(cmatch.inc_dir==inc_dire) & (cmatch.out_dir==out_dire_A), 'move_no'] = 19\n",
    "        cmatch.loc[(cmatch.inc_dir==inc_dire) & (cmatch.out_dir==out_dire_B), 'move_no'] = 19\n",
    "        cmatches.append(cmatch)\n",
    "\n",
    "    # 각 coordination node에 대하여 (inc_edge_id, out_edge_id) 부여\n",
    "    coord['inter_no'] = coord['parent_id'].map(node2inter)\n",
    "    coord = coord.rename(columns={'child_id':'node_id'})\n",
    "    coord[['inc_dir', 'out_dir', 'inc_angle','out_angle']] = np.nan\n",
    "    coord['move_no'] = 20\n",
    "    coord = coord[['inter_no', 'phase_no', 'ring_type', 'move_no', 'inc_dir', 'out_dir', 'inc_angle','out_angle', 'inc_edge', 'out_edge', 'node_id']]\n",
    "    \n",
    "    # display(coord)\n",
    "    cmatches = pd.concat(cmatches)\n",
    "    match6 = pd.concat([match5, cmatches, coord]).drop_duplicates().sort_values(by=['inter_no', 'node_id', 'phase_no', 'ring_type'])\n",
    "    match6.to_csv('../Intermediates/match6.csv')\n",
    "    return match6"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "def make_matching(match6):\n",
    "    '''\n",
    "    이동류 매칭 : 각 교차로에 대하여, 가능한 모든 이동류 (1~18, 21)에 대한 진입·진출엣지ID를 지정한다.\n",
    "    모든 이동류에 대해 지정하므로, 시차제시 이전과 다른 이동류가 등장하더라도 항상 진입·진출 엣지 ID를 지정할 수 있다.    \n",
    "     - matching의 컬럼 : inter_no, move_no, inc_dir, out_dir, inc_edge, out_edge, node_id\n",
    "    \n",
    "    설명 : \n",
    "     - 필요한 리스트, 딕셔너리 등을 정의\n",
    "        (1) 가능한 (진입방향, 진출방향) 목록 [리스트]\n",
    "        (2) 각 교차로별 방향 목록 : pdires (possible directions) [딕셔너리]\n",
    "        (3) 각 (교차로, 진입방향) 별 진입id 목록 : inc2id (incoming direction to incoming edge_id) [딕셔너리]\n",
    "        (4) 각 (교차로, 진출방향) 별 진출id 목록 : out2id (outgoing direction to outgoing edge_id) [딕셔너리]\n",
    "        (5) 각 교차로별 가능한 (진입방향, 진출방향) 목록 : pflow (possible flows) [딕셔너리]\n",
    "     - matching은 빈 리스트로 지정.\n",
    "     - 모든 노드id에 대하여 다음 과정을 반복\n",
    "     - 해당 노드id에 대한 모든 가능한 (진입방향, 진출방향)에 대하여 다음 과정을 반복\n",
    "     - (노드id, 진입방향)으로부터 진입엣지id를 얻어냄. 마찬가지로 진출엣지id도 얻어냄\n",
    "     - 얻어낸 정보를 바탕으로 한 행(new_row)을 만들고 이것을 matching에 append\n",
    "     '''\n",
    "\n",
    "    match7 = match6.copy()\n",
    "    match7 = match7[['inter_no', 'move_no', 'inc_dir', 'out_dir', 'inc_edge', 'out_edge', 'node_id']]\n",
    "    inter_node = pd.read_csv('../Data/tables/inter_node.csv', index_col=0)\n",
    "    nema = pd.read_csv('../Data/tables/nema.csv', encoding='cp949')\n",
    "\n",
    "    parent_ids = sorted(inter_node[inter_node.inter_type=='parent'].node_id.unique())\n",
    "    child_ids = sorted(inter_node[inter_node.inter_type=='child'].node_id.unique())\n",
    "\n",
    "    # (1) 가능한 (진입방향, 진출방향) 목록 \n",
    "    flows = nema.dropna().apply(lambda row: (row['inc_dir'], row['out_dir']), axis=1).tolist()\n",
    "    # (2) 각 교차로별 방향 목록 : pdires (possible directions)\n",
    "    pdires = {}\n",
    "    for node_id in parent_ids:\n",
    "        dires = match7[match7.node_id == node_id][['inc_dir','out_dir']].values.flatten()\n",
    "        dires = {dire for dire in dires if type(dire)==str}\n",
    "        pdires[node_id] = dires\n",
    "    # (3) 각 (교차로, 진입방향) 별 진입id 목록 : inc2id (incoming direction to incoming edge_id)\n",
    "    inc2id = {}\n",
    "    for node_id in parent_ids:\n",
    "        for inc_dir in pdires[node_id]:\n",
    "            df = match7[(match7.node_id==node_id) & (match7.inc_dir==inc_dir)]\n",
    "            inc2id[(node_id, inc_dir)] = df.inc_edge.iloc[0]\n",
    "    # (4) 각 (교차로, 진출방향) 별 진출id 목록 : out2id (outgoing direction to outgoing edge_id)\n",
    "    out2id = {}\n",
    "    for node_id in parent_ids:\n",
    "        for out_dir in pdires[node_id]:\n",
    "            df = match7[(match7.node_id==node_id) & (match7.out_dir==out_dir)]\n",
    "            out2id[(node_id, out_dir)] = df.out_edge.iloc[0]\n",
    "    # (5) 각 교차로별 가능한 (진입방향, 진출방향) 목록 : pflow (possible flows)\n",
    "    pflow = {}\n",
    "    for node_id in parent_ids:\n",
    "        pflow[node_id] = [flow for flow in flows if set(flow).issubset(pdires[node_id])]\n",
    "    # (6) 가능한 이동류에 대하여 진입id, 진출id 배정 : matching\n",
    "    node2inter = dict(zip(match7['node_id'], match7['inter_no']))\n",
    "    dires_right = ['북', '서', '남', '동', '북'] # ex (북, 서), (서, 남) 등은 우회전 flow\n",
    "    matching = []\n",
    "    for node_id in parent_ids:\n",
    "        inter_no = node2inter[node_id]\n",
    "        # 좌회전과 직진(1 ~ 16)\n",
    "        for (inc_dir, out_dir) in pflow[node_id]:\n",
    "            move_no = nema[(nema.inc_dir==inc_dir) & (nema.out_dir==out_dir)].move_no.iloc[0]\n",
    "            inc_edge = inc2id[(node_id, inc_dir)]\n",
    "            out_edge = out2id[(node_id, out_dir)]\n",
    "            new_row = pd.DataFrame({'inter_no':[inter_no], 'move_no':[move_no],\n",
    "                                    'inc_dir':[inc_dir], 'out_dir':[out_dir],\n",
    "                                    'inc_edge':[inc_edge], 'out_edge':[out_edge], 'node_id':[node_id]})\n",
    "            matching.append(new_row)\n",
    "        # 보행신호(17), 전적색(18)\n",
    "        new_row = pd.DataFrame({'inter_no':[inter_no] * 2, 'move_no':[17, 18],\n",
    "                                'inc_dir':[None]*2, 'out_dir':[None]*2,\n",
    "                                'inc_edge':[None]*2, 'out_edge':[None]*2, 'node_id':[node_id]*2})\n",
    "        matching.append(new_row)\n",
    "        # 신호우회전(21)\n",
    "        for d in range(len(dires_right)-1):\n",
    "            inc_dir = dires_right[d]\n",
    "            out_dir = dires_right[d+1]\n",
    "            if {inc_dir, out_dir}.issubset(pdires[node_id]):\n",
    "                inc_edge = inc2id[(node_id, inc_dir)]\n",
    "                out_edge = out2id[(node_id, out_dir)]\n",
    "                new_row = pd.DataFrame({'inter_no':[inter_no], 'move_no':[21],\n",
    "                                        'inc_dir':[inc_dir], 'out_dir':[out_dir],\n",
    "                                        'inc_edge':[inc_edge], 'out_edge':[out_edge], 'node_id':[node_id]})\n",
    "                matching.append(new_row)\n",
    "    matching.append(match7[match7.node_id.isin(child_ids)])\n",
    "    matching = pd.concat(matching)\n",
    "    matching = matching.dropna().sort_values(by=['inter_no', 'node_id', 'move_no']).reset_index(drop=True)\n",
    "    matching['move_no'] = matching['move_no'].astype(int)\n",
    "    matching.to_csv('../Intermediates/matching.csv')\n",
    "    return matching"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "match1 = pd.read_csv('../Intermediates/match1.csv', index_col=0)\n",
    "match2 = make_match2(match1)\n",
    "match3 = make_match3(match2)\n",
    "match4 = make_match4(match3)\n",
    "match5 = make_match5(match4)\n",
    "match6 = make_match6(match5)\n",
    "matching = make_matching(match6)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# B. 5초 간격으로 이동류번호 수집"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "# 5초 단위로 이동류번호 저장 및 신호이력에서 유닉스시각 가져와서 표시, 한시간동안의 데이터만 보관\n",
    "midnight = int(datetime(2024, 1, 5, 0, 0, 0).timestamp())\n",
    "next_day = int(datetime(2024, 1, 6, 0, 0, 0).timestamp())\n",
    "fsecs = range(midnight, next_day, 5) # fsecs : unix time by Five SECondS\n",
    "fmins = range(midnight, next_day, 300) # fmins : unix time by Five MINuteS"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "def save_movement():\n",
    "    # time2move = dict(zip(fsecs,moves)) # move : 어느 순간의 이동류정보\n",
    "    history = pd.read_csv('../Data/tables/history.csv', index_col=0)\n",
    "\n",
    "    time2movement = {} # movement : 어느 순간의, 그 순간으로부터 한시간 동안의 (교차로번호 + 현시별이동류번호 + 시작시간)\n",
    "    # - 아래 절차를 5초마다 반복\n",
    "    for fsec in tqdm(fsecs): # fsec : unix time by Five SECond\n",
    "        # 1. 상태 테이블 조회해서 전체 데이터중 필요데이터(교차로번호, A링 현시번호, A링 이동류번호, B링 현시번호, B링 이동류번호)만 수집 : A\n",
    "        # move = time2move[fsec]\n",
    "        move = pd.read_csv(f'../Data/tables/move/move_{fsec}.csv', index_col=0)\n",
    "        # 2. 이력 테이블 조회해서 교차로별로 유닉스시간 최대인 데이터(교차로변호, 종료유닉스타임)만 수집 : B\n",
    "        recent_histories = [group.iloc[-1:] for _, group in history[history['end_unix'] < fsec].groupby('inter_no')] # 교차로별로 유닉스시간이 최대인 행들\n",
    "        if not recent_histories:\n",
    "            rhistory = pd.DataFrame({'inter_no':[], 'end_unix':[]}) # recent history\n",
    "        else:\n",
    "            rhistory = pd.concat(recent_histories)\n",
    "        recent_unix = rhistory[['inter_no', 'end_unix']]\n",
    "        # 3. 상태 테이블 조회정보(A)와 이력 테이블 조회정보(B) 조인(키값 : 교차로번호) : C\n",
    "        move = pd.merge(move, recent_unix, how='left', on='inter_no')\n",
    "        move['end_unix'] = move['end_unix'].fillna(0).astype(int)\n",
    "        move = move.drop_duplicates()\n",
    "        # 4. C데이터 프레임에 신규 컬럼(시작 유닉스타임) 생성 후 종료유닉스 타임 값 입력, 종료 유닉스 타임 컬럼 제거\n",
    "        move = move.rename(columns = {'end_unix':'start_unix'})\n",
    "        # 5. 이동류 이력정보 READ\n",
    "        #     - CSV 파일로 서버에 저장된 이동류정보를 읽어옴(파일이 없는 경우에는 데이터가 없는 프레임 D 생성)\n",
    "        try:\n",
    "            if isinstance(movement, pd.DataFrame): # movement가 존재할 경우 그걸 그대로 씀.\n",
    "                pass\n",
    "            else: \n",
    "                movement = pd.DataFrame()\n",
    "        except NameError: # movement가 존재하지 않는 경우 생성\n",
    "            movement = pd.DataFrame()\n",
    "        # 6. 이동류 이력정보 데이터테이블(D)에 C데이터 add\n",
    "        movement = pd.concat([movement, move])\n",
    "        # 7. D데이터 프레임에서 중복데이터 제거(교차로번호, 시작 유닉스타임, A링 현시번호, B링 현시번호 같은 행은 제거)\n",
    "        movement = movement.drop_duplicates(['inter_no','phas_A','phas_B','start_unix'])\n",
    "        # 8. D데이터 보관 시간 기준시간을 시작 유닉스 타임의 최대값 - 3600을 값으로 산출하고, 보관 시간 기준시간보다 작은 시작 유닉스 타임을 가진 행은 모두 제거(1시간 데이터만 보관)\n",
    "        movement = movement[movement.start_unix > fsec - 3600]\n",
    "        movement = movement.sort_values(by=['start_unix','inter_no','phas_A','phas_B']).reset_index(drop=True)\n",
    "\n",
    "        time2movement[fsec] = movement\n",
    "        movement.to_csv(f'../Intermediates/movement/movement_{fsec}.csv')\n",
    "# save_movement()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# C. 5분 간격으로 신호이력 수집 및 통합테이블 생성"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "m = 105\n",
    "midnight = int(datetime(2024, 1, 5, 0, 0, 0).timestamp())\n",
    "next_day = int(datetime(2024, 1, 6, 0, 0, 0).timestamp())\n",
    "fmins = range(midnight, next_day, 300)\n",
    "\n",
    "# 현재시각\n",
    "present_time = fmins[m]\n",
    "\n",
    "plan = pd.read_csv('../Data/tables/plan.csv', index_col=0)\n",
    "history = pd.read_csv('../Data/tables/history.csv', index_col=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "def make_splits(plan):\n",
    "    # split, isplit : A,B 분리 혹은 통합시 사용될 수 있는 딕셔너리 \n",
    "    splits = {} # splits maps (inter_no, start_hour, start_minute) to split \n",
    "    for i, row in plan.iterrows():\n",
    "        inter_no = row.inter_no\n",
    "        start_hour = row.start_hour\n",
    "        start_minute = row.start_minute\n",
    "        cycle = row.cycle\n",
    "        cums_A = row[[f'dura_A{j}' for j in range(1,9)]].cumsum()\n",
    "        cums_B = row[[f'dura_B{j}' for j in range(1,9)]].cumsum()\n",
    "        splits[(inter_no, start_hour, start_minute)] = {} # split maps (phas_A, phas_B) to k\n",
    "        k = 0\n",
    "        for t in range(cycle):\n",
    "            new_phas_A = len(cums_A[cums_A < t]) + 1\n",
    "            new_phas_B = len(cums_B[cums_B < t]) + 1\n",
    "            if k == 0 or ((new_phas_A, new_phas_B) != (phas_A, phas_B)):\n",
    "                k += 1\n",
    "            phas_A = new_phas_A\n",
    "            phas_B = new_phas_B\n",
    "            splits[(inter_no, start_hour, start_minute)][(phas_A, phas_B)] = k\n",
    "\n",
    "    isplits = {} # the inverse of splits\n",
    "    for i in splits:\n",
    "        isplits[i] = {splits[i][k]:k for k in splits[i]} # isplit maps k to (phas_A, phas_B)\n",
    "    return splits, isplits\n",
    "splits, isplits = make_splits(plan)\n",
    "\n",
    "def make_timetable(plan):\n",
    "    # timetable\n",
    "    timetable = plan[['start_hour', 'start_minute']].drop_duplicates()\n",
    "    timetable['start_seconds'] = midnight + timetable['start_hour'] * 3600 + timetable['start_minute'] * 60\n",
    "    return timetable\n",
    "timetable = make_timetable(plan)\n",
    "\n",
    "# inter2node\n",
    "inter_node = pd.read_csv('../Data/tables/inter_node.csv', index_col=0)\n",
    "inter_node = inter_node[inter_node.inter_type=='parent']\n",
    "inter2node = dict(zip(inter_node['inter_no'], inter_node['node_id']))\n",
    "\n",
    "hours = np.array(range(midnight - 7200, next_day + 1, 3600)) # 정각에 해당하는 시각들 목록"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "def calculate_DS(rhist, curr_unix, hours, timetable):\n",
    "    program_starts = np.array(timetable.start_seconds)\n",
    "    idx = (program_starts <= present_time).sum() - 1\n",
    "    program_start = program_starts[idx]\n",
    "    if list(hours[hours <= curr_unix]):\n",
    "        ghour_lt_curr_unix = hours[hours <= curr_unix].max() # the greatest hour less than or equal to curr_unix\n",
    "    else:\n",
    "        ghour_lt_curr_unix = program_start\n",
    "    start_unixes = rhist.start_unix.unique()\n",
    "    start_unixes_lt_ghour = np.sort(start_unixes[start_unixes < ghour_lt_curr_unix]) # start unixes less than ghour_lt_curr_unix\n",
    "    # 기준유닉스(base_unix) : curr_unix보다 작은 hour 중에서 가장 큰 값으로부터 다섯 번째로 작은 start_unix\n",
    "    if len(start_unixes_lt_ghour) > 5:\n",
    "        base_unix = start_unixes_lt_ghour[-5]\n",
    "    # start_unixes_lt_ghour의 길이가 5 미만일 경우에는 맨 앞 start_unix로 base_unix를 지정\n",
    "    else:\n",
    "        base_unix = rhist.start_unix.min()\n",
    "    D_n = curr_unix - base_unix\n",
    "    S_n_durs = rhist[(rhist.start_unix > base_unix) & (rhist.start_unix <= curr_unix)] \\\n",
    "        [[f'dura_{alph}{j}' for alph in ['A', 'B'] for j in range(1,9)]]\n",
    "    S_n = S_n_durs.values.sum() // 2\n",
    "    return D_n, S_n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "def load_prow(timetable, inter_no, time):\n",
    "    '''\n",
    "    load planned row\n",
    "    '''\n",
    "    # 프로그램 시작시각\n",
    "    program_starts = np.array(timetable.start_seconds)\n",
    "    idx = (program_starts <= time).sum() - 1\n",
    "    program_start = program_starts[idx]\n",
    "\n",
    "    # 최근 프로그램 시작시각에 대한 신호계획\n",
    "    start_hour = timetable.iloc[idx].start_hour\n",
    "    start_minute = timetable.iloc[idx].start_minute\n",
    "    prow = plan[(plan.inter_no==inter_no) & (plan.start_hour==start_hour) & (plan.start_minute==start_minute)] # planned row\n",
    "    return program_start, prow\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "def make_rhistory(history, present_time):\n",
    "    # 1. 조회시점의 유닉스 타임 이전의 신호이력 수집\n",
    "    rhistory = history.copy() # recent history\n",
    "    rhistory = rhistory[(rhistory.end_unix <= present_time) & (rhistory.end_unix > present_time - 9000)] # 두 시간 반 전부터 현재까지의 신호이력을 가져옴. 9000 = 3600 * 2.5\n",
    "\n",
    "    # rhistory에 모든 교차로번호가 존재하지 않으면 해당 교차로번호에 대한 신호이력을 추가함 (at 최근 프로그램 시작시각)\n",
    "    whole_inter_nos = sorted(history.inter_no.unique())\n",
    "    recent_inter_nos = sorted(rhistory.inter_no.unique())\n",
    "    if not whole_inter_nos==recent_inter_nos:\n",
    "        for inter_no in set(whole_inter_nos) - set(recent_inter_nos):\n",
    "            program_start, prow = load_prow(timetable, inter_no, present_time - 9000)\n",
    "            cycle = prow.cycle.iloc[0]\n",
    "            row1 = prow.drop(['start_hour', 'start_minute'], axis=1).copy()\n",
    "            row2 = prow.drop(['start_hour', 'start_minute'], axis=1).copy()\n",
    "            row3 = prow.drop(['start_hour', 'start_minute'], axis=1).copy()\n",
    "            # prow에서 필요한 부분을 rhistory에 추가\n",
    "            row1['end_unix'] = program_start\n",
    "            row2['end_unix'] = program_start + cycle\n",
    "            row3['end_unix'] = present_time + 180\n",
    "            rhistory = pd.concat([rhistory, row1, row2, row3]).reset_index(drop=True)\n",
    "\n",
    "    for inter_no in set(whole_inter_nos):\n",
    "        program_start, prow = load_prow(timetable, inter_no, present_time)\n",
    "        cycle = prow.cycle.iloc[0]\n",
    "        row3 = prow.drop(['start_hour', 'start_minute'], axis=1).copy()\n",
    "        # prow에서 필요한 부분을 rhistory에 추가\n",
    "        row3['end_unix'] = present_time + 180\n",
    "        rhistory = pd.concat([rhistory, row3]).reset_index(drop=True)\n",
    "\n",
    "\n",
    "    # 2. 시작 유닉스 타임컬럼 생성 후 종류 유닉스 타임에서 현시별 현시기간 컬럼의 합을 뺀 값으로 입력\n",
    "    # - 현시시간의 합을 뺀 시간의 +- 10초 이내에 이전 주기정보가 존재하면 그 유닉스 시간을 시작 유닉스시간 값으로 하고, 존재하지 않으면 현시시간의 합을 뺀 유닉스 시간을 시작 유닉스 시간으로 지정\n",
    "    for i, row in rhistory.iterrows():\n",
    "        inter_no = row.inter_no\n",
    "        end_unix = row.end_unix\n",
    "        elapsed_time = row[[f'dura_{alph}{j}' for alph in ['A', 'B'] for j in range(1,9)]].sum() // 2 # 현시시간 합\n",
    "        # 이전 유닉스 존재하지 않음 : 현시시간 합의 차\n",
    "        start_unix = end_unix - elapsed_time\n",
    "        pre_rows = history[:i] # previous rows\n",
    "        if inter_no in pre_rows.inter_no.unique(): # 이전 유닉스 존재\n",
    "            pre_unix = pre_rows[pre_rows.inter_no == inter_no]['end_unix'].iloc[-1] # previous unix time\n",
    "            # 이전 유닉스 존재, abs < 10 : 이전 유닉스\n",
    "            if abs(pre_unix - start_unix) < 10:\n",
    "                start_unix = pre_unix\n",
    "            # 이전 유닉스 존재, abs >=10 : 현시시간 합의 차\n",
    "            else:\n",
    "                pass\n",
    "        rhistory.loc[i, 'start_unix'] = start_unix \n",
    "    rhistory[rhistory.isna()] = 0\n",
    "    rhistory['start_unix'] = rhistory['start_unix'].astype(int)\n",
    "    rhistory = rhistory[['inter_no', 'start_unix'] + [f'dura_{alph}{j}' for alph in ['A', 'B'] for j in range(1,9)] + ['cycle']]\n",
    "    return rhistory\n",
    "\n",
    "rhistory = make_rhistory(history, present_time)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "def processing(rhistory, hours):\n",
    "    rhists = []\n",
    "    for inter_no in sorted(rhistory.inter_no.unique()):\n",
    "        rhist = rhistory.copy()[rhistory.inter_no==inter_no]\n",
    "        rhist = rhist.drop_duplicates(subset=['start_unix']).reset_index(drop=True)\n",
    "\n",
    "        # D_n 및 S_n 값 정의\n",
    "        rhist['D_n'] = 0 # D_n : 시간차이\n",
    "        rhist['S_n'] = 0 # S_n : 현시시간합\n",
    "        for n in range(len(rhist)):\n",
    "            curr_unix = rhist.iloc[n].start_unix # current start_unix\n",
    "            rhist.loc[n, ['D_n', 'S_n']] = calculate_DS(rhist, curr_unix, hours, timetable)\n",
    "\n",
    "        # 이전시각, 현재시각\n",
    "        prev_unix = rhist.loc[0, 'start_unix'] # previous start_unix\n",
    "        curr_unix = rhist.loc[1, 'start_unix'] # current start_unix\n",
    "\n",
    "        # rhist의 마지막 행에 도달할 때까지 반복\n",
    "        while True:\n",
    "            n = rhist[rhist.start_unix==curr_unix].index[0]\n",
    "            cycle = rhist.loc[n, 'cycle']\n",
    "            D_n = rhist.loc[n, 'D_n']\n",
    "            S_n = rhist.loc[n, 'S_n']\n",
    "            # 참값인 경우\n",
    "            if (abs(D_n - S_n) <= 5):\n",
    "                pass\n",
    "            # 참값이 아닌 경우\n",
    "            else:\n",
    "                # 2-1-1. 결측치 처리 : 인접한 두 start_unix의 차이가 계획된 주기의 두 배보다 크면 결측이 일어났다고 판단, 신호계획의 현시시간으로 \"대체\"\n",
    "                if curr_unix - prev_unix >= 2 * cycle:\n",
    "                    # prev_unix를 계획된 주기만큼 늘려가면서 한 행씩 채워나간다.\n",
    "                    # (curr_unix와의 차이가 계획된 주기보다 작거나 같아질 때까지)\n",
    "                    while curr_unix - prev_unix > cycle:\n",
    "                        prev_unix += cycle\n",
    "                        # 신호 계획(prow) 불러오기\n",
    "                        start_seconds = np.array(timetable.start_seconds)\n",
    "                        idx = (start_seconds <= prev_unix).sum() - 1\n",
    "                        start_hour = timetable.iloc[idx].start_hour\n",
    "                        start_minute = timetable.iloc[idx].start_minute\n",
    "                        prow = plan.copy()[(plan.inter_no==inter_no) & (plan.start_hour==start_hour) & (plan.start_minute==start_minute)] # planned row\n",
    "                        # prow에서 필요한 부분을 rhist에 추가\n",
    "                        prow['start_unix'] = prev_unix\n",
    "                        prow = prow.drop(['start_hour', 'start_minute', 'offset'], axis=1)\n",
    "                        cycle = prow.iloc[0].cycle\n",
    "                        rhist = pd.concat([rhist, prow])\n",
    "                        rhist = rhist.sort_values(by='start_unix').reset_index(drop=True)\n",
    "                        n += 1\n",
    "\n",
    "                # 2-1-2. 이상치 처리 : 비율에 따라 해당 행을 \"삭제\"(R_n <= 0.5) 또는 \"조정\"(R_n > 0.5)한다\n",
    "                R_n = (curr_unix - prev_unix) / cycle # R_n : 비율\n",
    "                # R_n이 0.5보다 작거나 같으면 해당 행을 삭제\n",
    "                if R_n <= 0.5:\n",
    "                    rhist = rhist.drop(index=n).reset_index(drop=True)\n",
    "                    if n >= rhist.index[-1]:\n",
    "                        break\n",
    "                    # 행삭제에 따른 curr_unix, R_n 재정의\n",
    "                    curr_unix = rhist.loc[n, 'start_unix']\n",
    "                    R_n = (curr_unix - prev_unix) / cycle # R_n : 비율\n",
    "\n",
    "                # R_n이 0.5보다 크면 해당 행 조정 (비율을 유지한 채로 현시시간 대체)\n",
    "                if R_n > 0.5:\n",
    "                    # 신호 계획(prow) 불러오기\n",
    "                    start_seconds = np.array(timetable.start_seconds)\n",
    "                    idx = (start_seconds <= curr_unix).sum() - 1\n",
    "                    start_hour = timetable.iloc[idx].start_hour\n",
    "                    start_minute = timetable.iloc[idx].start_minute\n",
    "                    prow = plan[(plan.inter_no==inter_no) & (plan.start_hour==start_hour) & (plan.start_minute==start_minute)] # planned row\n",
    "                    # 조정된 현시시간 (prow에 R_n을 곱하고 정수로 바꿈)\n",
    "                    adjusted_dur = prow.copy()[[f'dura_{alph}{j}' for alph in ['A', 'B'] for j in range(1,9)]] * R_n\n",
    "                    int_parts = adjusted_dur.iloc[0].apply(lambda x: int(x))\n",
    "                    frac_parts = adjusted_dur.iloc[0] - int_parts\n",
    "                    difference = round(adjusted_dur.iloc[0].sum()) - int_parts.sum()\n",
    "                    for _ in range(difference): # 소수 부분이 가장 큰 상위 'difference'개의 값에 대해 올림 처리\n",
    "                        max_frac_index = frac_parts.idxmax()\n",
    "                        int_parts[max_frac_index] += 1\n",
    "                        frac_parts[max_frac_index] = 0  # 이미 처리된 항목은 0으로 설정\n",
    "                    # rhist에 조정된 현시시간을 반영\n",
    "                    rhist.loc[n, [f'dura_{alph}{j}' for alph in ['A', 'B'] for j in range(1,9)]] = int_parts.values\n",
    "                    rhist.loc[n, 'cycle'] = int_parts.sum().sum() // 2\n",
    "\n",
    "            if n >= rhist.index[-1]:\n",
    "                break\n",
    "            prev_unix = curr_unix\n",
    "            curr_unix = rhist.loc[n+1, 'start_unix']\n",
    "\n",
    "        # 생략해도 무방할 코드\n",
    "        rhist = rhist.reset_index(drop=True)\n",
    "        rhist = rhist.sort_values(by=['start_unix'])\n",
    "\n",
    "        # D_n 및 S_n 값 재정의\n",
    "        for n in range(len(rhist)):\n",
    "            curr_unix = rhist.iloc[n].start_unix # current start_unix\n",
    "            rhist.loc[n, ['D_n', 'S_n']] = calculate_DS(rhist, curr_unix, hours, timetable)\n",
    "        rhists.append(rhist)\n",
    "    rhists = pd.concat(rhists).sort_values(by=['start_unix','inter_no'])\n",
    "    rhists = rhists[rhists.start_unix >= present_time - 3600]\n",
    "    rhists = rhists.drop(columns=['D_n', 'S_n'])\n",
    "    return rhists\n",
    "rhists = processing(rhistory, hours)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "def make_hrhists(rhists, isplits, timetable):\n",
    "    # 계층화된 형태로 변환\n",
    "    hrhists = [] # hierarchied recent history\n",
    "    for i, row in rhists.iterrows():\n",
    "        inter_no = row.inter_no\n",
    "        start_unix = row.start_unix\n",
    "\n",
    "        ind = (timetable['start_seconds'] <= row.start_unix).sum() - 1\n",
    "        start_hour = timetable.iloc[ind].start_hour\n",
    "        start_minute = timetable.iloc[ind].start_minute\n",
    "        isplit = isplits[(inter_no, start_hour, start_minute)]\n",
    "        phas_As = [isplit[j][0] for j in isplit.keys()]\n",
    "        phas_Bs = [isplit[j][1] for j in isplit.keys()]\n",
    "        durs_A = row[[f'dura_A{j}' for j in range(1,9)]]\n",
    "        durs_B = row[[f'dura_B{j}' for j in range(1,9)]]\n",
    "        durations = []\n",
    "        for j in range(1, len(isplit)+1):\n",
    "            ja = isplit[j][0]\n",
    "            jb = isplit[j][1]\n",
    "            if ja == jb:\n",
    "                durations.append(min(durs_A[ja-1], durs_B[jb-1]))\n",
    "            else:\n",
    "                durations.append(abs(durs_A[ja-1] - durs_B[ja-1]))\n",
    "        new_rows = pd.DataFrame({'inter_no':[inter_no] * len(durations), 'start_unix':[start_unix] * len(durations),\n",
    "                                'phas_A':phas_As, 'phas_B':phas_Bs, 'duration':durations})\n",
    "        hrhists.append(new_rows)\n",
    "    hrhists = pd.concat(hrhists)\n",
    "    hrhists = hrhists.sort_values(by = ['start_unix', 'inter_no', 'phas_A', 'phas_B']).reset_index(drop=True)\n",
    "    return hrhists\n",
    "hrhists = make_hrhists(rhists, isplits, timetable)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "def make_movements():\n",
    "    movements_path = '../Intermediates/movement/'\n",
    "    movements_list = [pd.read_csv(movements_path + file, index_col=0) for file in tqdm(os.listdir(movements_path))]\n",
    "    movements = pd.concat(movements_list)\n",
    "    movements = movements.drop(columns=['start_unix'])\n",
    "    movements = movements.drop_duplicates()\n",
    "    movements = movements.sort_values(by=['inter_no', 'phas_A', 'phas_B'])\n",
    "    movements = movements.reset_index(drop=True)\n",
    "    movements.to_csv('../Intermediates/movements.csv')\n",
    "    return movements\n",
    "movements = pd.read_csv('../Intermediates/movements.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "def update_movement(hrhists, movement, movements):\n",
    "    # 중복을 제거하고 (inter_no, start_unix) 쌍을 만듭니다.\n",
    "    hrhists_inter_unix = set(hrhists[['inter_no', 'start_unix']].drop_duplicates().itertuples(index=False, name=None))\n",
    "    movement_inter_unix = set(movement[['inter_no', 'start_unix']].drop_duplicates().itertuples(index=False, name=None))\n",
    "\n",
    "    # hrhists에는 있지만 movement에는 없는 (inter_no, start_unix) 쌍을 찾습니다.\n",
    "    missing_in_movement = hrhists_inter_unix - movement_inter_unix\n",
    "\n",
    "    # 새로운 행들을 생성합니다.\n",
    "    new_rows = []\n",
    "    if missing_in_movement:\n",
    "        for inter_no, start_unix in missing_in_movement:\n",
    "            # movements에서 해당 inter_no의 데이터를 찾습니다.\n",
    "            new_row = movements[movements['inter_no'] == inter_no].copy()\n",
    "            # start_unix 값을 설정합니다.\n",
    "            new_row['start_unix'] = start_unix\n",
    "            new_rows.append(new_row)\n",
    "\n",
    "        # 새로운 데이터프레임을 생성하고 기존 movement 데이터프레임과 합칩니다.\n",
    "        new_movement = pd.concat(new_rows, ignore_index=True)\n",
    "        movement_updated = pd.concat([movement, new_movement], ignore_index=True)\n",
    "    else:\n",
    "        movement_updated = movement\n",
    "    return movement_updated\n",
    "movement = pd.read_csv(f'../Intermediates/movement/movement_{present_time}.csv', index_col=0)\n",
    "movement_updated = update_movement(hrhists, movement, movements)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "def make_histid(present_time, movedur, inter2node):\n",
    "    # 이동류 매칭 테이블에서 진입id, 진출id를 가져와서 붙임.\n",
    "    for i, row in movedur.iterrows():\n",
    "        inter_no = row.inter_no\n",
    "        start_unix = row.start_unix\n",
    "        # incoming and outgoing edges A\n",
    "        move_A = row.move_A\n",
    "        if move_A in [17, 18]:\n",
    "            inc_edge_A = np.nan\n",
    "            out_edge_A = np.nan\n",
    "        else:\n",
    "            match_A = matching[(matching.inter_no == inter_no) & (matching.move_no == move_A)].iloc[0]\n",
    "            inc_edge_A = match_A.inc_edge\n",
    "            out_edge_A = match_A.out_edge\n",
    "        movedur.loc[i, ['inc_edge_A', 'out_edge_A']] = [inc_edge_A, out_edge_A]\n",
    "        # incoming and outgoing edges B\n",
    "        move_B = row.move_B\n",
    "        if move_B in [17, 18]:\n",
    "            inc_edge_B = np.nan\n",
    "            out_edge_B = np.nan\n",
    "        else:\n",
    "            match_B = matching[(matching.inter_no == inter_no) & (matching.move_no == move_B)].iloc[0]\n",
    "            inc_edge_B = match_B.inc_edge\n",
    "            out_edge_B = match_B.out_edge\n",
    "        movedur.loc[i, ['inc_edge_B', 'out_edge_B']] = [inc_edge_B, out_edge_B]\n",
    "\n",
    "    # 이동류 컬럼 제거\n",
    "    movedur = movedur.drop(['move_A', 'move_B'], axis=1)\n",
    "\n",
    "    histid = movedur.copy() # history with edge ids (incoming and outgoing edge ids)\n",
    "    histid['node_id'] = histid['inter_no'].map(inter2node)\n",
    "    histid = histid[['inter_no', 'node_id', 'start_unix', 'phas_A', 'phas_B', 'duration', 'inc_edge_A', 'out_edge_A', 'inc_edge_B', 'out_edge_B']]\n",
    "    histid_start = present_time - 1200\n",
    "    histid = histid[histid.start_unix > histid_start]\n",
    "    return histid\n",
    "# movedur\n",
    "movedur = pd.merge(movement_updated, hrhists, how='inner', on=['inter_no', 'start_unix', 'phas_A', 'phas_B']) # movements and durations\n",
    "movedur = movedur.sort_values(by=['start_unix', 'inter_no', 'phas_A','phas_B'])\n",
    "movedur = movedur[['inter_no', 'start_unix', 'phas_A', 'phas_B', 'move_A', 'move_B', 'duration']]\n",
    "histid = make_histid(present_time, movedur, inter2node)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "# def preprocess(m):\n",
    "#     '''\n",
    "#     통합테이블(histid)를 만드는 함수\n",
    " \n",
    "#     input : m\n",
    "#      - m ranges from 0 to 287, but 0 makes an error where 288 = 86400//300\n",
    "#      - present_time = fmins[m] : 현재시점\n",
    " \n",
    "#     output : histid (통합테이블, HISTory with edge_IDs)\n",
    "#      - 컬럼 : inter_no, node_id, start_unix, phas_A, phas_B, duration, inc_edge_A, out_edge_A, inc_edge_B, out_edge_B\n",
    " \n",
    "#     주요 데이터, 중간산출물 및 결과물 :\n",
    "#     # 데이터\n",
    "#      - history : 신호이력 (inter_no, end_unix, dura_Aj, dura_Bj, cycle, offset)\n",
    "#      - plan : 신호계획 (inter_no, start_hour, start_minute, dura_Aj, dura_Bj cycle, offset)\n",
    "#     # 중간산출물\n",
    "#      - rhists (recent history)\n",
    "#         - history에서 현재 시각 이전의 데이터를 가져옴.\n",
    "#         - end_unix를 start_unix로 변환\n",
    "#         - 참값판단 프로세스(결측·이상치 처리)\n",
    "#         - 컬럼 : inter_no, start_unix, dura_Aj, dura_Bj, cycle\n",
    "#      - hrhists (hierarchized recent history)\n",
    "#         - rhists를 계층화\n",
    "#         - 컬럼 : inter_no, start_unix, phas_A, phas_B, duration\n",
    "#      - movements\n",
    "#         - 각 교차로에 대하여 현시별로 이동류를 정해놓음.\n",
    "#         - join시 사용하기 위함.\n",
    "#         - 한 번 만들어놓고 두고두고 사용함.\n",
    "#         - 컬럼 : inter_no, phas_A, phas_B, move_A, move_B\n",
    "#      - movement\n",
    "#         - 현재 시점에서의 이동류정보\n",
    "#         - 컬럼 : inter_no, phas_A, phas_B, move_A, move_B, start_unix\n",
    "#      - movement_updated\n",
    "#         - movement와 hrhists를 join하기 전에, movement에는 없지만 hrhists에는 있는 start_unix에 대한 이동류 정보를 가져와 movement에 붙임\n",
    "#         - 이동류정보는 앞서 정의한 movements에서 가져옴.\n",
    "#         - 컬럼 : inter_no, phas_A, phas_B, move_A, move_B, start_unix\n",
    "#      - movedur\n",
    "#         - hrhists와 movement_updated를 join\n",
    "#         - 컬럼 : inter_no, phas_A, phas_B, move_A, move_B, start_unix, duration\n",
    "#     # 결과 : histid\n",
    "#         - 신호생성에 직접적으로 사용되는 데이터프레임\n",
    "#         - 컬럼 : inter_no, node_id, start_unix, phas_A, phas_B, duration, inc_edge_A, out_edge_A, inc_edge_B, out_edge_B\n",
    "#         - 한글컬럼 : 교차로번호, 노드id, 시작유닉스, A현시번호, B현시번호, 현시시간, 진입엣지(A), 진출엣지(A), 진입엣지(B), 진출엣지(B)\n",
    "#     '''\n",
    "#     midnight = int(datetime(2024, 1, 5, 0, 0, 0).timestamp())\n",
    "#     next_day = int(datetime(2024, 1, 6, 0, 0, 0).timestamp())\n",
    "#     fmins = range(midnight, next_day, 300) # fmins : unix time by Five MINuteS\n",
    "\n",
    "#     # 현재시각\n",
    "#     present_time = fmins[m]\n",
    "#     print(datetime.fromtimestamp(present_time))\n",
    "\n",
    "#     # 사용할 표준 테이블 목록\n",
    "#     plan = pd.read_csv('../Data/tables/plan.csv', index_col=0)\n",
    "#     history = pd.read_csv('../Data/tables/history.csv', index_col=0)\n",
    "\n",
    "#     # 참고할 딕셔너리, 데이터프레임, 리스트 등 목록\n",
    "#     splits, isplits = make_splits(plan)\n",
    "#     timetable = make_timetable(plan)\n",
    "#     inter_node = pd.read_csv('../Data/tables/inter_node.csv', index_col=0)\n",
    "#     inter_node = inter_node[inter_node.inter_type=='parent']\n",
    "#     inter2node = dict(zip(inter_node['inter_no'], inter_node['node_id']))\n",
    "#     hours = np.array(range(midnight - 7200, next_day + 1, 3600)) # 정각에 해당하는 시각들 목록\n",
    "\n",
    "#     # rhistory, rhists, hrhists\n",
    "#     rhistory = make_rhistory(history, present_time)\n",
    "#     rhists = processing(rhistory, hours)\n",
    "#     hrhists = make_hrhists(rhists, isplits, timetable)\n",
    "\n",
    "#     # movements, movement, movement_updated\n",
    "#     movements = pd.read_csv('../Intermediates/movements.csv')\n",
    "#     movement = pd.read_csv(f'../Intermediates/movement/movement_{present_time}.csv', index_col=0)\n",
    "#     movement_updated = update_movement(hrhists, movement, movements)\n",
    "\n",
    "#     # movedur\n",
    "#     movedur = pd.merge(movement_updated, hrhists, how='inner', on=['inter_no', 'start_unix', 'phas_A', 'phas_B']) # movements and durations\n",
    "#     movedur = movedur.sort_values(by=['start_unix', 'inter_no', 'phas_A','phas_B'])\n",
    "#     movedur = movedur[['inter_no', 'start_unix', 'phas_A', 'phas_B', 'move_A', 'move_B', 'duration']]\n",
    "\n",
    "#     # histid\n",
    "#     histid = make_histid(present_time, movedur, inter2node)\n",
    "#     return histid"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2024-01-05 08:45:00\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>inter_no</th>\n",
       "      <th>node_id</th>\n",
       "      <th>start_unix</th>\n",
       "      <th>phas_A</th>\n",
       "      <th>phas_B</th>\n",
       "      <th>duration</th>\n",
       "      <th>inc_edge_A</th>\n",
       "      <th>out_edge_A</th>\n",
       "      <th>inc_edge_B</th>\n",
       "      <th>out_edge_B</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>165</th>\n",
       "      <td>177</td>\n",
       "      <td>i2</td>\n",
       "      <td>1704410710</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>40</td>\n",
       "      <td>-571542809_01</td>\n",
       "      <td>571542811_01</td>\n",
       "      <td>571542811_02</td>\n",
       "      <td>571542809_01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>166</th>\n",
       "      <td>177</td>\n",
       "      <td>i2</td>\n",
       "      <td>1704410710</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>25</td>\n",
       "      <td>571542811_02</td>\n",
       "      <td>571542107_01</td>\n",
       "      <td>-571542809_01</td>\n",
       "      <td>571542809_01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>167</th>\n",
       "      <td>177</td>\n",
       "      <td>i2</td>\n",
       "      <td>1704410710</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>71</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>168</th>\n",
       "      <td>177</td>\n",
       "      <td>i2</td>\n",
       "      <td>1704410710</td>\n",
       "      <td>4</td>\n",
       "      <td>4</td>\n",
       "      <td>34</td>\n",
       "      <td>-571542809_01</td>\n",
       "      <td>571542811_01</td>\n",
       "      <td>571542107_02</td>\n",
       "      <td>571542809_01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>578</th>\n",
       "      <td>176</td>\n",
       "      <td>i1</td>\n",
       "      <td>1704410720</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>37</td>\n",
       "      <td>-571542810_01</td>\n",
       "      <td>-571542797_02.99</td>\n",
       "      <td>571542797_02.99</td>\n",
       "      <td>571542810_01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>529</th>\n",
       "      <td>202</td>\n",
       "      <td>i9</td>\n",
       "      <td>1704411920</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>221</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>614</th>\n",
       "      <td>206</td>\n",
       "      <td>i7</td>\n",
       "      <td>1704411940</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>63</td>\n",
       "      <td>-571511538_02</td>\n",
       "      <td>571542073_02</td>\n",
       "      <td>571542073_01</td>\n",
       "      <td>571511538_02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>615</th>\n",
       "      <td>206</td>\n",
       "      <td>i7</td>\n",
       "      <td>1704411940</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>63</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>616</th>\n",
       "      <td>206</td>\n",
       "      <td>i7</td>\n",
       "      <td>1704411940</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>37</td>\n",
       "      <td>-571511538_02</td>\n",
       "      <td>571542073_02</td>\n",
       "      <td>571542073_01</td>\n",
       "      <td>571511538_02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>617</th>\n",
       "      <td>206</td>\n",
       "      <td>i7</td>\n",
       "      <td>1704411940</td>\n",
       "      <td>4</td>\n",
       "      <td>4</td>\n",
       "      <td>37</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>248 rows × 10 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "     inter_no node_id  start_unix  phas_A  phas_B  duration     inc_edge_A  \\\n",
       "165       177      i2  1704410710       1       1        40  -571542809_01   \n",
       "166       177      i2  1704410710       2       2        25   571542811_02   \n",
       "167       177      i2  1704410710       3       3        71            NaN   \n",
       "168       177      i2  1704410710       4       4        34  -571542809_01   \n",
       "578       176      i1  1704410720       1       1        37  -571542810_01   \n",
       "..        ...     ...         ...     ...     ...       ...            ...   \n",
       "529       202      i9  1704411920       2       2       221            NaN   \n",
       "614       206      i7  1704411940       1       1        63  -571511538_02   \n",
       "615       206      i7  1704411940       2       2        63            NaN   \n",
       "616       206      i7  1704411940       3       3        37  -571511538_02   \n",
       "617       206      i7  1704411940       4       4        37            NaN   \n",
       "\n",
       "           out_edge_A       inc_edge_B    out_edge_B  \n",
       "165      571542811_01     571542811_02  571542809_01  \n",
       "166      571542107_01    -571542809_01  571542809_01  \n",
       "167               NaN              NaN           NaN  \n",
       "168      571542811_01     571542107_02  571542809_01  \n",
       "578  -571542797_02.99  571542797_02.99  571542810_01  \n",
       "..                ...              ...           ...  \n",
       "529               NaN              NaN           NaN  \n",
       "614      571542073_02     571542073_01  571511538_02  \n",
       "615               NaN              NaN           NaN  \n",
       "616      571542073_02     571542073_01  571511538_02  \n",
       "617               NaN              NaN           NaN  \n",
       "\n",
       "[248 rows x 10 columns]"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "def preprocess(m):\n",
    "    '''\n",
    "    통합테이블(histid)를 만드는 함수\n",
    " \n",
    "    input : m\n",
    "     - m ranges from 0 to 287, but 0 makes an error where 288 = 86400//300\n",
    "     - present_time = fmins[m] : 현재시점\n",
    " \n",
    "    output : histid (통합테이블, HISTory with edge_IDs)\n",
    "     - 컬럼 : inter_no, node_id, start_unix, phas_A, phas_B, duration, inc_edge_A, out_edge_A, inc_edge_B, out_edge_B\n",
    " \n",
    "    주요 데이터, 중간산출물 및 결과물 :\n",
    "    # 데이터\n",
    "     - history : 신호이력 (inter_no, end_unix, dura_Aj, dura_Bj, cycle, offset)\n",
    "     - plan : 신호계획 (inter_no, start_hour, start_minute, dura_Aj, dura_Bj cycle, offset)\n",
    "    # 중간산출물\n",
    "     - rhists (recent history)\n",
    "        - history에서 현재 시각 이전의 데이터를 가져옴.\n",
    "        - end_unix를 start_unix로 변환\n",
    "        - 참값판단 프로세스(결측·이상치 처리)\n",
    "        - 컬럼 : inter_no, start_unix, dura_Aj, dura_Bj, cycle\n",
    "     - hrhists (hierarchized recent history)\n",
    "        - rhists를 계층화\n",
    "        - 컬럼 : inter_no, start_unix, phas_A, phas_B, duration\n",
    "     - movements\n",
    "        - 각 교차로에 대하여 현시별로 이동류를 정해놓음.\n",
    "        - join시 사용하기 위함.\n",
    "        - 한 번 만들어놓고 두고두고 사용함.\n",
    "        - 컬럼 : inter_no, phas_A, phas_B, move_A, move_B\n",
    "     - movement\n",
    "        - 현재 시점에서의 이동류정보\n",
    "        - 컬럼 : inter_no, phas_A, phas_B, move_A, move_B, start_unix\n",
    "     - movement_updated\n",
    "        - movement와 hrhists를 join하기 전에, movement에는 없지만 hrhists에는 있는 start_unix에 대한 이동류 정보를 가져와 movement에 붙임\n",
    "        - 이동류정보는 앞서 정의한 movements에서 가져옴.\n",
    "        - 컬럼 : inter_no, phas_A, phas_B, move_A, move_B, start_unix\n",
    "     - movedur\n",
    "        - hrhists와 movement_updated를 join\n",
    "        - 컬럼 : inter_no, phas_A, phas_B, move_A, move_B, start_unix, duration\n",
    "    # 결과 : histid\n",
    "        - 신호생성에 직접적으로 사용되는 데이터프레임\n",
    "        - 컬럼 : inter_no, node_id, start_unix, phas_A, phas_B, duration, inc_edge_A, out_edge_A, inc_edge_B, out_edge_B\n",
    "        - 한글컬럼 : 교차로번호, 노드id, 시작유닉스, A현시번호, B현시번호, 현시시간, 진입엣지(A), 진출엣지(A), 진입엣지(B), 진출엣지(B)\n",
    "    '''\n",
    "    midnight = int(datetime(2024, 1, 5, 0, 0, 0).timestamp())\n",
    "    next_day = int(datetime(2024, 1, 6, 0, 0, 0).timestamp())\n",
    "    fmins = range(midnight, next_day, 300) # fmins : unix time by Five MINuteS\n",
    "\n",
    "    # 현재시각\n",
    "    present_time = fmins[m]\n",
    "    print(datetime.fromtimestamp(present_time))\n",
    "\n",
    "    # 사용할 표준 테이블 목록\n",
    "    plan = pd.read_csv('../Data/tables/plan.csv', index_col=0)\n",
    "    history = pd.read_csv('../Data/tables/history.csv', index_col=0)\n",
    "\n",
    "    # 참고할 딕셔너리, 데이터프레임, 리스트 등 목록\n",
    "    splits, isplits = make_splits(plan)\n",
    "    timetable = make_timetable(plan)\n",
    "    inter_node = pd.read_csv('../Data/tables/inter_node.csv', index_col=0)\n",
    "    inter_node = inter_node[inter_node.inter_type=='parent']\n",
    "    inter2node = dict(zip(inter_node['inter_no'], inter_node['node_id']))\n",
    "    hours = np.array(range(midnight - 7200, next_day + 1, 3600)) # 정각에 해당하는 시각들 목록\n",
    "\n",
    "    # rhistory, rhists, hrhists\n",
    "    rhistory = make_rhistory(history, present_time)\n",
    "    rhists = processing(rhistory, hours)\n",
    "    hrhists = make_hrhists(rhists, isplits, timetable)\n",
    "\n",
    "    # movements, movement, movement_updated\n",
    "    movements = pd.read_csv('../Intermediates/movements.csv')\n",
    "    movement = pd.read_csv(f'../Intermediates/movement/movement_{present_time}.csv', index_col=0)\n",
    "    movement_updated = update_movement(hrhists, movement, movements)\n",
    "\n",
    "    # movedur\n",
    "    movedur = pd.merge(movement_updated, hrhists, how='inner', on=['inter_no', 'start_unix', 'phas_A', 'phas_B']) # movements and durations\n",
    "    movedur = movedur.sort_values(by=['start_unix', 'inter_no', 'phas_A','phas_B'])\n",
    "    movedur = movedur[['inter_no', 'start_unix', 'phas_A', 'phas_B', 'move_A', 'move_B', 'duration']]\n",
    "\n",
    "    # histid\n",
    "    histid = make_histid(present_time, movedur, inter2node)\n",
    "    return histid\n",
    "preprocess(105)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1704411900\n",
      "2024-01-05 08:45:00\n",
      "175\n",
      "1130\n",
      "True True\n",
      "176\n",
      "1190\n",
      "True True\n",
      "177\n",
      "1200\n",
      "True True\n",
      "178\n",
      "1190\n",
      "True True\n",
      "201\n",
      "1190\n",
      "True True\n",
      "202\n",
      "1110\n",
      "True True\n",
      "206\n",
      "1180\n",
      "True True\n",
      "210\n",
      "1130\n",
      "True True\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "print(present_time)\n",
    "histid = preprocess(105)\n",
    "k = 0\n",
    "for inter_no in sorted(histid.inter_no.unique()):\n",
    "    k += 1\n",
    "    df = histid[histid.inter_no==inter_no]\n",
    "    M = df.start_unix.max()\n",
    "    m = df.start_unix.min()\n",
    "    print(inter_no)\n",
    "    print(M - m)\n",
    "    plt.scatter([M, m], [k] * 2, c='y')\n",
    "    plt.scatter([present_time], [k], c='b')\n",
    "    plt.axvline(x=histid.start_unix.max(), c='r')\n",
    "    plt.axvline(x=histid.start_unix.min(), c='r')\n",
    "    print(m < present_time, present_time < M)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1704411900\n",
      "2024-01-05 08:45:00\n",
      "175\n",
      "1130\n",
      "True True\n",
      "176\n",
      "1190\n",
      "True True\n",
      "177\n",
      "1200\n",
      "True True\n",
      "178\n",
      "1190\n",
      "True True\n",
      "201\n",
      "1190\n",
      "True True\n",
      "202\n",
      "1110\n",
      "True True\n",
      "206\n",
      "1180\n",
      "True True\n",
      "210\n",
      "1130\n",
      "True True\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "print(present_time)\n",
    "histid = preprocess(105)\n",
    "k = 0\n",
    "for inter_no in sorted(histid.inter_no.unique()):\n",
    "    k += 1\n",
    "    df = histid[histid.inter_no==inter_no]\n",
    "    M = df.start_unix.max()\n",
    "    m = df.start_unix.min()\n",
    "    print(inter_no)\n",
    "    print(M - m)\n",
    "    plt.scatter([M, m], [k] * 2, c='y')\n",
    "    plt.scatter([present_time], [k], c='b')\n",
    "    plt.axvline(x=histid.start_unix.max(), c='r')\n",
    "    plt.axvline(x=histid.start_unix.min(), c='r')\n",
    "    print(m < present_time, present_time < M)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2024-01-05 08:45:00\n",
      "2024-01-05 08:50:00\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>inter_no</th>\n",
       "      <th>node_id</th>\n",
       "      <th>start_unix</th>\n",
       "      <th>phas_A</th>\n",
       "      <th>phas_B</th>\n",
       "      <th>duration</th>\n",
       "      <th>inc_edge_A</th>\n",
       "      <th>out_edge_A</th>\n",
       "      <th>inc_edge_B</th>\n",
       "      <th>out_edge_B</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>515</th>\n",
       "      <td>206</td>\n",
       "      <td>i7</td>\n",
       "      <td>1704411040</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>44</td>\n",
       "      <td>-571511538_02</td>\n",
       "      <td>571542073_02</td>\n",
       "      <td>571542073_01</td>\n",
       "      <td>571511538_02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>516</th>\n",
       "      <td>206</td>\n",
       "      <td>i7</td>\n",
       "      <td>1704411040</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>44</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>517</th>\n",
       "      <td>206</td>\n",
       "      <td>i7</td>\n",
       "      <td>1704411040</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>26</td>\n",
       "      <td>-571511538_02</td>\n",
       "      <td>571542073_02</td>\n",
       "      <td>571542073_01</td>\n",
       "      <td>571511538_02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>518</th>\n",
       "      <td>206</td>\n",
       "      <td>i7</td>\n",
       "      <td>1704411040</td>\n",
       "      <td>4</td>\n",
       "      <td>4</td>\n",
       "      <td>26</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>702</th>\n",
       "      <td>177</td>\n",
       "      <td>i2</td>\n",
       "      <td>1704411049</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>40</td>\n",
       "      <td>-571542809_01</td>\n",
       "      <td>571542811_01</td>\n",
       "      <td>571542811_02</td>\n",
       "      <td>571542809_01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>520</th>\n",
       "      <td>202</td>\n",
       "      <td>i9</td>\n",
       "      <td>1704412220</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>207</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>477</th>\n",
       "      <td>206</td>\n",
       "      <td>i7</td>\n",
       "      <td>1704412240</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>69</td>\n",
       "      <td>-571511538_02</td>\n",
       "      <td>571542073_02</td>\n",
       "      <td>571542073_01</td>\n",
       "      <td>571511538_02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>478</th>\n",
       "      <td>206</td>\n",
       "      <td>i7</td>\n",
       "      <td>1704412240</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>69</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>479</th>\n",
       "      <td>206</td>\n",
       "      <td>i7</td>\n",
       "      <td>1704412240</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>41</td>\n",
       "      <td>-571511538_02</td>\n",
       "      <td>571542073_02</td>\n",
       "      <td>571542073_01</td>\n",
       "      <td>571511538_02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>480</th>\n",
       "      <td>206</td>\n",
       "      <td>i7</td>\n",
       "      <td>1704412240</td>\n",
       "      <td>4</td>\n",
       "      <td>4</td>\n",
       "      <td>41</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>241 rows × 10 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "     inter_no node_id  start_unix  phas_A  phas_B  duration     inc_edge_A  \\\n",
       "515       206      i7  1704411040       1       1        44  -571511538_02   \n",
       "516       206      i7  1704411040       2       2        44            NaN   \n",
       "517       206      i7  1704411040       3       3        26  -571511538_02   \n",
       "518       206      i7  1704411040       4       4        26            NaN   \n",
       "702       177      i2  1704411049       1       1        40  -571542809_01   \n",
       "..        ...     ...         ...     ...     ...       ...            ...   \n",
       "520       202      i9  1704412220       2       2       207            NaN   \n",
       "477       206      i7  1704412240       1       1        69  -571511538_02   \n",
       "478       206      i7  1704412240       2       2        69            NaN   \n",
       "479       206      i7  1704412240       3       3        41  -571511538_02   \n",
       "480       206      i7  1704412240       4       4        41            NaN   \n",
       "\n",
       "       out_edge_A    inc_edge_B    out_edge_B  \n",
       "515  571542073_02  571542073_01  571511538_02  \n",
       "516           NaN           NaN           NaN  \n",
       "517  571542073_02  571542073_01  571511538_02  \n",
       "518           NaN           NaN           NaN  \n",
       "702  571542811_01  571542811_02  571542809_01  \n",
       "..            ...           ...           ...  \n",
       "520           NaN           NaN           NaN  \n",
       "477  571542073_02  571542073_01  571511538_02  \n",
       "478           NaN           NaN           NaN  \n",
       "479  571542073_02  571542073_01  571511538_02  \n",
       "480           NaN           NaN           NaN  \n",
       "\n",
       "[241 rows x 10 columns]"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "preprocess(105)\n",
    "preprocess(106)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2024-01-05 13:40:00\n",
      "[175, 176, 177, 178, 201, 202, 206, 210]\n",
      "[175, 176, 177, 178, 201, 202, 206, 210]\n"
     ]
    }
   ],
   "source": [
    "histid = preprocess(164)\n",
    "print(sorted(history.inter_no.unique()))\n",
    "print(sorted(histid.inter_no.unique()))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [],
   "source": [
    "# for m in range(100, 200):\n",
    "#     histid = preprocess(m)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "rts",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.10"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}