refactor: big, more patterns\n\nBREAKING CHANGES

main.py

@@ -5,6 +5,7 @@
 
 import json
 import os
+import random
 import time
 import traceback
 import uuid
@@ -26,50 +27,75 @@ FACULTETS = sorted([
 DIRNAME = "excels"
 DIFFABLE_DATES = "diffable_dates.txt"
 
+SKIP_DIFFABLE_DATES = True
+
 DEBUG_ONE_FAC = None #'fevt'
-result_groups = {}
+LOGGING = False
+
+unique_raws = set()
 result = {
     "version": 1,
     "notice": "ОТКАЗ ОТ ОТВЕТСТВЕННОСТИ: Данные, доступ к API и т.д. предоставляется КАК-ЕСТЬ (AS-IS) без каких либо, явно или не явно подразумеваемых гарантий.\n\nПарсер написал: Миронов Станислав\n\nИсточник данных: https://www.vstu.ru/student/raspisaniya/zanyatiy/index.php",
     "actual_at": round(time.time()),
-    "documentation": "https://fazziclay.com/api/v1/vstu_schedule_parser/scheme.json",
+    "documentation": "https://fazziclay.com/api/v1/vstu_schedule_parser/scheme.json (temporary outdated)",
-    "daypicture": "QwQ",
+    "daypicture": "0w0",
-    "daycite": "running on a rope",
+    "daycite": "KIlLSWITCH",
     "contact": "https://fazziclay.com/",
     "university": "VSTU",
     "university_site": "https://www.vstu.ru/",
     "source": "https://fazziclay.com/api/v1/vstu_schedule_parser/result.json",
     "stat": {
         "total_parsing_time": -1,
+        "excels": {
+            "fine": 0,
+            "bad": 0
+        },
+        "groups": 0,
+        "unique_raws": -1
     },
     "api_notices": {
-        "updated_at": 1759651871,
+        "updated_at": 1773523692,
-        "text": "Пожалуйста сохраняйте 'updated_at', это время изменения ЭТОГО текста. Тут возможно будут появлятся важные BREAKING CHANGES и дедлайны к ним.\nПо хорошему если updated_at другой по сравнению с вашем кэшем это сообщение должно отправляться вам в телеграм как уведомление о поедстоящих изменениях\nwarning=True значит 'text' содержит важное а не как щас hint.\n\n  ~fazziclay aka Stanislav;\n\n2025-10-05: добавлено data_source_hash в эксель и в группу. Это SHA1 of скачанный эксель файл.",
+        "text_pre1": "Пожалуйста сохраняйте 'updated_at', это время изменения ЭТОГО текста. Тут возможно будут появлятся важные BREAKING CHANGES и дедлайны к ним.\nПо хорошему если updated_at другой по сравнению с вашем кэшем это сообщение должно отправляться вам в телеграм как уведомление о поедстоящих изменениях\nwarning=True значит 'text' содержит важное а не как щас hint.\n\n  ~fazziclay aka Stanislav;\n\n2025-10-05: добавлено data_source_hash в эксель и в группу. Это SHA1 of скачанный эксель файл.",
-        "warning": False,
+        "text": "2026-03-15 BREAKING CHANGES! By Stanislav Mironov.\n\nИзменено многое в угоду унифкации и расширению спаршенных групп. Пока alpha",
+        "warning": True,
         "tut-plavayuschaya-struktura": "required only 'updated_at', 'text' and 'warning'"
     },
-    "doubled_groups": [],
     "debug": {
-        "bleu~~": 2
+        "bleu~~": 3
     },
     "excels": [],
     "facultets": FACULTETS,
-
+    "group_names_parsed": [],
-    "emptykey1": "",
+    "unique_raws": unique_raws,
-    "emptykey2": "",
-
-    "groups": result_groups,
-
-    "emptykey3": "",
-    "emptykey4": "",
     "see_header_at_top_of_this_file": "SEE TOP OF THIS FILE | ОБРАТИТЕ ВНИМАНИЕ НА ВЕРХ ЭТОГО ФАЙЛА"
 }
 
+def process_obj(data):
+    try:
+        if isinstance(data, dict):
+            for key, value in data.items():
+                if key == "raw":
+                    unique_raws.update(value)
+                
+                process_obj(value)
+                
+        # Если это список, проходим по его элементам
+        elif isinstance(data, list):
+            for item in data:
+                process_obj(item)
+            
+    except Exception as e:
+        print("Failed process_obj")
+        print(e)
+
 def process_excel_file(facultet, excel_url, counter, latest_changed):
     is_xlsx = excel_url.endswith(".xlsx")
     download_place = f"{DIRNAME}/" + f"_[C{counter}]_" + facultet + ".xls" + ("x" if is_xlsx else "")
 
     excel_filename = excel_url.split("/")[-1]
+    if "ФЭУ" not in excel_filename:
+        print("SKIPPED")
+        return
 
     excel_info = {
         "filename": excel_filename,
@@ -80,9 +106,9 @@ def process_excel_file(facultet, excel_url, counter, latest_changed):
         "group_names_parsed": [],
         "facultet": facultet,
         "counter": counter,
-        "week_keys_metadata": {}
+        "sheets": []
     }
-    parser.LOGGING = False
+    parser.LOGGING = LOGGING
 
     try:
         aigenerated.download_file_from_url(excel_url, download_place)
@@ -94,44 +120,45 @@ def process_excel_file(facultet, excel_url, counter, latest_changed):
         
         while True:
             print(f"Parsing sheet №{reader.get_sheet_index()+1} (from 1)")
+            sheet_dict = {
+                "index": reader.get_sheet_index(),
+                "name": reader.get_sheet_name(),
+                "reader_info": reader.info(),
+                "group_names_parsed": [],
+                "groups": {}
+            }
+            excel_info['sheets'].append(sheet_dict)
             prs = parser.Parser(reader)
 
             print("Parser created; parser.parse();")
             prs.parse()
 
             print("parsed done!")
+            
+            if len(prs.raw_no_schedule) > 0:
+                sheet_dict["raw_no_schedule"] = prs.raw_no_schedule
+                
+            if len(prs.features) > 0:
+                sheet_dict["features"] = sorted(prs.features)
+            
             if prs.parser_error is not None:
-                excel_info["parser_error_cycle_" + str(reader.get_sheet_index()+1)] = prs.parser_error
+                sheet_dict["parser_error"] = prs.parser_error
                 
             if prs.parser_warnings is not None and len(prs.parser_warnings) > 0:
-                excel_info["parser_warnings_cycle_" + str(reader.get_sheet_index()+1)] = prs.parser_warnings
+                sheet_dict["parser_warnings"] = prs.parser_warnings
 
             for group_name in prs.groups.keys():
-                if group_name in result_groups.keys():
+                gr = prs.groups[group_name]
-                    print(f" -- WTF -- Doubled groups -- name: {group_name}")
+                gr["excel_url"] = excel_url
-                    if 'warning_doubled_groups_skip' not in excel_info.keys():
+                sheet_dict["group_names_parsed"].append(group_name)
-                        excel_info['warning_doubled_groups_skip'] = []
-                    
-                    excel_info['warning_doubled_groups_skip'].append(group_name)
-                    result['doubled_groups'].append(group_name)
-
-
-                    continue
-                        
-                gr = result_groups[group_name] = prs.groups[group_name]
-                gr['facultet'] = facultet
-                gr['data_source'] = excel_filename # same as 'filename' in excel_info's
-                gr['data_source_hash'] = sha1hash
-                gr['debug'] = {
-                    "excel_url": excel_url,
-                    "reader_info": reader.info(),
-                    "reader_sheet_index": reader.get_sheet_index(),
-                    "download_place": download_place
-                }
                 excel_info["group_names_parsed"].append(group_name)
-                excel_info['week_keys_metadata'] = prs.week_keys_metadata
+                result["group_names_parsed"].append(group_name)
+                result['stat']['groups'] += 1
+                sheet_dict['week_keys_metadata'] = prs.week_keys_metadata
+                sheet_dict['groups'][group_name] = gr
+                process_obj(gr['slots'])
 
-            print(f"Populates {len(prs.groups)} groups to result: " + " ".join(prs.groups.keys()))
+            print(f"Populates {len(prs.groups)} groups: " + " ".join(prs.groups.keys()))
 
             if not reader.has_next_sheet():
                 print("File ended")
@@ -159,10 +186,12 @@ def process_excel_file(facultet, excel_url, counter, latest_changed):
         })
     
     result['excels'].append(excel_info)
+    k = "fine" if len(excel_info['group_names_parsed']) > 0 else "bad"
+    result['stat']['excels'][k] += 1
 
 faileds = []
 def main():
-    global result_groups, result
+    global result
     t = utils.StepTimeCounter()
     try:
         try:
@@ -189,7 +218,9 @@ def main():
 
     if now_diffable_dates == prev_diffable_dates:
         print("No date changes in vstu.ru website. Stopping")
+        if not SKIP_DIFFABLE_DATES:
             return
+        print("SKIP_DIFFABLE_DATES is True, force resuming")
     
     counter = 10000
     for excel_link in EXCEL_LINKS:
@@ -200,14 +231,8 @@ def main():
         process_excel_file(facultet, excel_url, counter, latest_changed)
 
     print("Saving result.json")
-    group_names_alphabeticaly = sorted(result_groups.keys())
-    sorted_groups = {}
-    for group_name in group_names_alphabeticaly:
-        sorted_groups[group_name] = result_groups[group_name]
-
-    result['groups'] = sorted_groups
-
     result['stat']['total_parsing_time'] = t.step()
+    result['unique_raws'] = sorted(unique_raws)
 
     json.dump(result, open('result.json', 'w'), indent=2, ensure_ascii=False)
     print("Saved to result.json indent=2")

parser.py

@@ -3,10 +3,21 @@
 PAIR_NUMS = [
     "1-2", "3-4", "5-6", "7-8", "9-10", "11-12", "13-14", "15-16"
 ]
+WEEKDAYS_STARTSWITH = [
+    "понед",
+    "вторник",
+    "среда",
+    "четверг",
+    "пятница",
+    "суббота"
+]
 
+bad_group_names = [
+    "янв", "февр", "март", "апр", "май", "сент", "окт", "ноя", "дек", "июнь", "июль", "авг"
+]
+
+from datetime import time
 import json
-import uuid
-import aigenerated
 from coord import Coord, Merged
 from translations import ExcelSheetReader
 import utils
@@ -18,49 +29,126 @@ def pprint(*args, **kwargs):
     if LOGGING:
         print(*args, **kwargs)
         
+def is_weeknum(text):
+    for wd in WEEKDAYS_STARTSWITH:
+        if text.strip().replace(" ", "").lower().startswith(wd):
+            return True
+    return False
+
+def is_pair(text):
+    for p in PAIR_NUMS:
+        if text.strip().replace(" ", "").lower().startswith(p):
+            return True
+    return False
+
 class Parser:
     def __init__(self, reader: ExcelSheetReader):
         self.reader =  reader
-        self.groups = {}
+        self.groups = {} # Группы которые удалось распарсить
-        self.week_keys_metadata = {}
+        self.features = set() # фичи данной страницы
+        self.week_keys_metadata = {} # календарик
+        self.schedule_range_row = None # [min, max] диапазон col включительно где расписание
+        self.raw_no_schedule = [] # всё что не schedule_range_row отправляется сюда ('СОГЛАСОВАНО:', etc..)
         
-        self.weeknums: defaultdict = defaultdict(set) # no support json!
+        self.weeknums: defaultdict = defaultdict(set) # no support json! (для week_keys_metadata)
-        self.parser_error = None
+        self.parser_error = None # ошибка парсера перед выходом
-        self.parser_warnings = []
+        self.parser_warnings = [] # предупреждения парсера
         pprint("Parser created for '{0}'".format(reader.info()))
 
     def parse(self):
-        monday = self.reader.find("ПОНЕДЕЛЬНИК")
+        # Характерные признаки разных сеток
-        if monday is None:
+        no_pair_numeration = False
+        col_distance_pair_weekday = None
+        weekday_firstly_calendar = False
+        
+        first_weekday = self.reader.find_any(WEEKDAYS_STARTSWITH, startswith=True, nospace=True)
+        
+        if first_weekday is None:
+            self.features.add("no_weekdays")
             print(" -- Failed parse! -- ")
-            print("ПОНЕДЕЛЬНИК НЕ НАЙДЕН!")
+            print("дни недели не найдены!")
-            self.parser_error = "'ПОНЕДЕЛЬНИК' не найден в таблице."
+            self.parser_error = f"{WEEKDAYS_STARTSWITH} ни один найден в таблице. Дня недели нет."
             return
         
-        if monday.col != 4:
+        pair_num_any = self.reader.find_any(PAIR_NUMS, nospace=True)
-            print("--- warning parse! ---")
+        if pair_num_any is None:
-            print(f"Monday col != 4 (actual: {monday})") 
+            no_pair_numeration = True
-            self.parser_warnings.append(f"Monday col != 4 (actual: {monday}); Это, наверное, может работать не стабильно!")
+            self.features.add("no_pair_numeration")
+            self.parser_warnings.append(f"Нет нумерации академических часов {PAIR_NUMS}")
+
+        else:
+            self.features.add("pair_numeration")
+            col_distance_pair_weekday = pair_num_any.col - first_weekday.col        
+        
+        head_rx = first_weekday.row - 1 # выше первого понидельника
+        group_col_start = first_weekday.col + 2
+        if col_distance_pair_weekday is not None:
+            if col_distance_pair_weekday > 1:
+                weekday_firstly_calendar = True
+                self.features.add("weekdays_before_calendar")
+                group_col_start = pair_num_any.col + 1
             
-        head_rx = monday.row - 1 # выше первого понидельника
         if head_rx < 0:
-            raise Exception("head_rx < 0: Программа пыталась найти 'ПОНЕДЕЛЬНИК', но по всей видимости не нашла.")
+            raise Exception("head_rx < 0: Программа пыталась найти день недели, но по всей видимости не нашла.")
 
         head = self.reader.get_row_values(head_rx) # get all ROW (months, groups)
         pprint(f"head={head}")
-        self.groups = parse_groups(self.reader, head, monday, head_rx) # parse groups to self.groups
+        
+        head_joined = " ||| ".join([v for v in head if isinstance(v, str) and v.strip()])
+        print(head_joined)
+        if "1 неделя" in head_joined or "1 НЕДЕЛЯ" in head_joined or "2 неделя" in head_joined or "2 НЕДЕЛЯ" in head_joined or "ИЗМЕНЕНИЯ" in head_joined or "изменения" in head_joined or "vtf-vstu.ru" in head_joined:
+            head_rx -= 1
+            head = self.reader.get_row_values(head_rx) # get all ROW (months, groups)
+            pprint(f"head (upper)={head}")
+            self.features.add("post_groups_info_row")
+        
+        self.groups = parse_groups(self.reader, head, group_col_start, head_rx) # parse groups to self.groups
         pprint(f'self.groups={json.dumps(self.groups, indent=2, ensure_ascii=False)}')
         
         pprint("\n\n\n")
 
         for group in self.groups.values():
             pprint("\nSTART OF PROCESS GROUP\n")
-            self.process_group(group, monday)
+            self.process_group(group, first_weekday, pair_num_any.col if pair_num_any else None)
             pprint("\nEND OF PROCESS GROUP\n")
 
-        self.process_weekmetadatas(monday)
+        # week metadatas parse
+        S = 9999999
+        group_min_col = S
+        group_min_row = S
         
-    def process_weekmetadatas(self, first_monday: "Coord"):
+        for x in self.groups.values():
+            p = x['position']
+            group_min_row = min(p[0], group_min_row)
+            group_min_col = min(p[1], group_min_col)
+        
+        if group_min_row != S and group_min_col != S:        
+            pprint("Process weekmetadatas!")
+            self.process_weekmetadatas(Coord(row=group_min_row, col=group_min_col))
+            
+        # parse no-schedule raws (согласовано, и т.д.)
+        self.parse_raw_no_schedule()
+
+    
+    def parse_raw_no_schedule(self):
+        """Распарсить всё за пределами self.schedule_range_row в self.raw_no_schedule"""
+        if self.schedule_range_row is None:
+            return
+        
+        row = 0
+        while row < self.reader.get_row_count():
+            if row >= self.schedule_range_row[0] and row <= self.schedule_range_row[1]:
+                row = self.schedule_range_row[1] + 1
+            
+            row_values = self.reader.get_row_values(row)
+            row_values = [v for v in row_values if isinstance(v, str) and v.strip()]
+            if len(row_values) > 0:
+                self.raw_no_schedule.append(row_values)
+            
+            row += 1
+
+    def process_weekmetadatas(self, first_group: "Coord"):
+        """Обработать календарик"""
         for x in self.weeknums.keys():
             pprint(x)
             set_of_merged: set = self.weeknums[x]
@@ -82,14 +170,16 @@ class Parser:
                 self.parser_warnings.append(f"Processing weekmetadata for '{x}' failed because weekday excel block width != 1 (actual {weekday_merged.width()})")
                 continue
             
-            month_row = first_monday.row - 1
+            month_row = first_group.row
-            curr_col = weekday_merged.low.col - 1
+            curr_col = first_group.col - 1
             while curr_col >= 0:
                 month_pos = Coord(month_row, curr_col)
                 month_cell = month_pos.cell(self.reader)
                 if month_cell.is_empty():
                     pprint("month cell is empty")
-                    break
+                    curr_col -= 1
+                    continue
+                
                 month_name = str(month_cell.value).strip()
                 pprint(month_cell)
                 all_nums_of_month = utils.parse_all_dirt(self.reader, month_pos.replace(row=weekday_merged.low.row), right=1, down=weekday_merged.height())
@@ -118,6 +208,16 @@ class Parser:
     def push_weekday_meta(self, weekday: str, weeknum: int, week_key_name: str, merged: "Merged"):
         self.weeknums[week_key_name].add(merged)
         
+    def row_with_schedule_notify(self, row_coord):
+        if self.schedule_range_row is None:
+            self.schedule_range_row = [row_coord, row_coord]
+            
+        if self.schedule_range_row[1] < row_coord:
+            self.schedule_range_row[1] = row_coord
+            
+        if self.schedule_range_row[0] > row_coord:
+            self.schedule_range_row[0] = row_coord
+
     def parse_potokoviy(self, merged: Merged):
         speaker = None
         location = None
@@ -132,163 +232,157 @@ class Parser:
 
         return {"loc": str(location).strip(), "leader": str(speaker).strip(), "name": str(merged.cell(self.reader).value).strip()}
 
-    def process_group(self, group: dict, monday: Coord):
+    def process_group(self, group: dict, first_weekday: Coord, pair_pos_col):
         """
         Обработать группы, выполняется для каждой группы, после того как они распарены (parse_groups)
             group = {'name': 'ИВТ-260', 'position': [5, 6], 'position_human': 'G6:J6'}
         """
         pprint(f"process_group group={group}")
         group_name = group['name']
-        pprint(group_name)
+        pprint(F"Имя группы: {group_name}")
-        row = group['position'][0] + 1 # counter for while, +1 for shift down; также номер строки в таблице (вроде с нуля)
+        row_c1 = group['position'][0] + 1 # counter for while, +1 for shift down; также номер строки в таблице (вроде с нуля)
+        self.row_with_schedule_notify(group['position'][0])
+        group_header_pos = Coord(group['position'][0], group['position'][1])
+        width = group['width']
         weeknum = 1 # номер недели, щёлкнет +1 при каком-то условии.
         previous_pair = None
-        while row < self.reader.get_row_count(): # maybe условие чтобы не уйти ниже чем есть строк
+        
-            pos = Coord(row, group['position'][1]) # текущая позиция, верхний левый угол (=low)
+        weekcycles = 0
-            pprint(f"while pos={pos}")
+        while row_c1 < self.reader.get_row_count():
-            pos_right = pos.shift(right=3)
+            pos_c1 = Coord(row_c1, group['position'][1]) # текущая позиция, верхний левый угол (=low)
-            pair_pos = pos.replace(col=monday.col + 1)
+            self.row_with_schedule_notify(pos_c1.row)
-            weekday_pos = pos.replace(col=monday.col)
+
-            merged = self.reader.get_merged_coord(pos)
+            
-            merged_cell = merged.cell(self.reader)
+            if pos_c1.cell(self.reader).is_nospace_nocase_same(group_name):
-            cv = merged_cell.value
+                pprint("Ended with grpup name; stop moving down, break")
-            # В конце (12 пара:>) название группы, можно использовать как якорь
-            if utils.unspace(cv) == group_name:
-                pprint("Lesson == group name; ending group loop.")
                 break
             
+            weekday_pos = pos_c1.replace(col=first_weekday.col)
+            weekday_cell = weekday_pos.cell(self.reader)
             weekday_mr = self.reader.get_merged_coord(weekday_pos)
-            weekday = utils.unspace(weekday_mr.cell(self.reader).value)
+            weekday = weekday_cell.value
-            pair_mr = self.reader.get_merged_coord(pair_pos)
-            pair = utils.unspace(pair_mr.cell(self.reader).value)
 
-            skip = 0
+            if not is_weeknum(weekday):
-            if weekday == "":
+                row_c1 += 1
-                if weeknum == 1:
+                pprint("Not weeknum!")
-                    weeknum += 1
+                if weekcycles > 0:
-                    pprint("------")
+                    if (weeknum != 2):
-                    skip = 1
+                        pprint("Weeknum now 2")
-                    row += 1
+                        weekday = 0
-                else:
+                    weeknum = 2
-                    break
+                continue
-            
-            if not skip:
-                next = 3 # на сколько пыгнуть для следующего шага?
             
+            pprint(weekday)            
             weekday_key_name = weekday + ("_1" if weeknum == 1 else "_2")
             self.push_weekday_meta(weekday, weeknum, weekday_key_name, weekday_mr)
             
-                is_empty_lesson = len(utils.parse_all_dirt(self.reader, pos, 4, 3)) == 0 # если в поле не найдено ничего..
+            # state
-                parsed_discipline_name = None
+            event_no = 1
-                parsed_location = None
+            is_widely = False
-                parsed_leader = None
+            override_col_range = None
-                pairs = 1
+            all_raw = set()
-                wtf_tomanypairs = False
+            pairs = set()
-                is_solid = pos_right in merged
+            times = []
-                parsed_uncotigorized = []
+            first_coord = None
-                is_wide_maybe_potokoviy = merged.width() > 4 # потоковая ли лекция (занимает несколько групп.)
             
-                if not is_empty_lesson:
+            row_c2 = row_c1
-                    cur = pos.shift(down=2)
+            while row_c2 <= weekday_mr.high.row:
-                    while utils.has_no_bottom_border(self.reader, cur):
+                pos_c2 = Coord(row_c2, group['position'][1]) # текущая позиция (внутри группы, внутри дня недели), верхний левый угол (=low)
-                        next += 3
+                cell_c2 = pos_c2.cell(self.reader)
-                        pairs += 1
+                mr_c2 = self.reader.get_merged_coord(pos_c2)
-                        pprint(f"next = {next} cur={cur}")
-                        if pairs >= 7:
-                            wtf_tomanypairs = True
-                            break
-                        cur = cur.shift(down=3)
                 
-                    if is_wide_maybe_potokoviy:
+                if first_coord is None:
-                        ret = self.parse_potokoviy(merged)
+                    first_coord = pos_c2.row
-                        parsed_location = ret['loc']
-                        parsed_leader = ret['leader']
-                        parsed_discipline_name = ret['name']
-                        parsed_uncotigorized = list(utils.parse_all_dirt(self.reader, merged.low, merged.width(), next))
                 
+                pair_num = None
+                pair_num_mr = None
+                if pair_pos_col is not None:
+                    pair_num = pos_c2.replace(col=pair_pos_col)
+                    pair_num_mr = self.reader.get_merged_coord(pair_num)
+                
+                if (not is_widely) and (mr_c2.low.col < group_header_pos.col or mr_c2.high.col > group_header_pos.col + width - 1):
+                    is_widely = True
+                    override_col_range = (mr_c2.low.col, mr_c2.high.col)
+                
+                col_low = group_header_pos.col
+                col_high = group_header_pos.col + width - 1
+                if override_col_range is not None:
+                    col_low = min(col_low, override_col_range[0])
+                    col_high = max(col_high, override_col_range[1])
+
+                dirty_line = utils.parse_all_dirt(self.reader, Coord(row_c2, col_low), (col_high - col_low + 1), 1, with_cells=True)
+                if len(dirty_line) > 0:
+                    if pair_num_mr is not None:
+                        pair_num_to_add = pair_num_mr.cell(self.reader).value.replace(" ", "").strip()
+                        if len(pair_num_to_add) == 0:
+                            pair_num_to_add = "???"
+                        pairs.add(pair_num_to_add)
+                    
+                    for cell in dirty_line:
+                        if not cell.is_time:
+                            all_raw.add(str(cell.value))
                         else:
-                        if (is_solid):
+                            dt: time = cell.value
-                            parsed_discipline_name = cv
+                            times.append(str(dt))
                 
-                        parsed_uncotigorized = list(utils.parse_all_dirt(self.reader, merged.low, 4, next))
+                def clean_state():
+                    nonlocal is_widely, override_col_range, event_no, all_raw, pairs, times, first_coord
+                    is_widely = False
+                    override_col_range = None
+                    event_no += 1
+                    all_raw = set()
+                    pairs = set()
+                    first_coord = None
+                    times = []
                 
-                # попытка исправить пару (1-2) если пустая.
-                fuck_empty_pair_in_excel = pair == ""
-                previous_dump = previous_pair
-                if fuck_empty_pair_in_excel:
-                    if previous_pair is None or previous_pair == "":
-                        pair = f"EMPTY_IN_EXCEL"
-                    else:
-                        pair = utils.next_element(PAIR_NUMS, previous_pair)
                 
-                if pair != "":
+                if not utils.has_no_bottom_border(self.reader, pos_c2) and not(mr_c2.high.row - row_c2 > 0):
-                    previous_pair = pair if next == 3 else None # костыль чтобы избежать гипотетически не верной даты.
+                    if not (len(all_raw) == 0):
+                        # this code last for current state event
+                        pprint(f"№{event_no} {pairs}: {'[wide] ' if is_widely else ''} raw={all_raw}")
                         
-                # пытаемся из некотегорезированных данных выцепить место и лидера (препода)
-                prepods = set()
-                if parsed_leader is not None: prepods.add(parsed_leader.strip())
-
-                locations = set()
-                if parsed_location is not None: locations.add(parsed_location.strip().replace("  ", ""))
-
-                for x in list(parsed_uncotigorized):
-                    if aigenerated.is_surname_string(x):
-                        prepods.add(x.strip())
-
-                    if aigenerated.is_room_number(x):
-                        locations.add(x.strip().replace("  ", "") if x is not None else None)
-
-                # попытка починить пустую дисциплину
-                if parsed_discipline_name is None:
-                    l = sorted(utils.remove_from_list(list(parsed_uncotigorized), list(locations | prepods | set([parsed_location, parsed_leader]))))
-                    parsed_discipline_name = " ".join(l)
-
-                # чистим сеты от мусора
-                utils.discards_list(prepods, nones=True, emptystrings=True)
-                utils.discards_list(locations, nones=True, emptystrings=True)
-                utils.discards_list(parsed_uncotigorized, nones=True, emptystrings=True)
-
-                # если не пустой предмет то записываем его
-                if not is_empty_lesson:
                         slots = group['slots']
                         w = weekday_key_name
                         if w not in slots.keys():
                             slots[w] = {}
                         
+                        pair_name = "????"
+                        try:
+                            pair_name = sorted(pairs)[0]
+                        except: pass
+                        
                         today = slots[w]
-                    today[pair] = {
+                        obj = {
-                        "excel_pos": str(pos),
+                            "object": "event",
-                        "discipline_name": parsed_discipline_name.strip(),
+                            "pairs": sorted(pairs),
-                        "locations": sorted(locations),
+                            "is_flow": is_widely,
-                        "leads": sorted(prepods),
+                            "excel_range": utils.merged_humanize((first_coord, col_low, row_c2, col_high)),
-                        "is_solid": is_solid,
+                            "raw": sorted(all_raw),
-                        "time_coeff": pairs,
-                        "is_flow": is_wide_maybe_potokoviy,
-                        "lefttopmerged": {
-                            "width": merged.width(),
-                            "height": merged.height(),
-                            "excel_range": utils.merged_humanize(merged.as_numbers())
-                        },
-                        "raw": sorted(parsed_uncotigorized),
                             "weekday": utils.weekday_to_num(weekday),
                             "weeknum": weeknum
                         }
-                    if fuck_empty_pair_in_excel:
+                        if len(times) > 0:
-                        today[pair]['pair_num_empty'] = {
+                            obj['times'] = times
-                            "prev": previous_dump,
+                        
-                            "restored": pair != "",
+                        if pair_pos_col is None:
-                            "pair": pair
+                            slots[w] = obj
-                        }
+                        else:
-                    if wtf_tomanypairs:
+                            today[pair_name] = obj
-                        today[pair]['to_many_parsing_time_coeff'] = True
+                    # here may be a empty all_raw
+                    clean_state()
+                    first_coord = None
                     
                 
-                # INCREMENT на next и конец цикла.
+                if row_c2 >= weekday_mr.high.row:
-                row += next
+                    clean_state()
+                    pprint("Last for weekday")
 
+                row_c2 += 1
             
+            row_c1 += weekday_mr.height()
+            weekcycles += 1
 
-def parse_groups(reader: "ExcelSheetReader", head, monday: Coord, head_rx):
+def parse_groups(reader: "ExcelSheetReader", head, col_start, head_rx):
     """Распознать список групп и метаданные к ним, по сути получить список названий группы и координат её верхнего header-а (AQ6:AT6)"""
     groups = {}
     i = 0
@@ -296,18 +390,23 @@ def parse_groups(reader: "ExcelSheetReader", head, monday: Coord, head_rx):
         x = head[i]
         pprint(f"while i={i} head[i]={x}")
         merged = reader.get_merged_coord(Coord(head_rx, i))
-        if i > monday.col + 1:
+        if i >= col_start:
-            if merged is None or x == "":
+            if merged is None or x == "" or x is None:
-                break
-
-            if merged.width() != 4:
-                pprint(f"WARNING: group header witdh !=4 (found: {merged.width()}); blocks !=4 not supported by parser.")
                 break
 
             name = utils.unspace(x)
+            skip = False
+            if "-" not in name:
+                for x in bad_group_names:
+                    if x in name.lower():
+                        skip = True
+                        pprint(f"Skip groupname {name} because not dash in name and in blacklist")
+                    
+            if not skip:
                 groups[name] = {
                     "name": name,
                     "position": [head_rx, i],
+                    "width": merged.width(),
                     "position_human": utils.merged_humanize(merged.as_numbers()),
                     "slots": {}
                 }

translations.py

@@ -1,6 +1,7 @@
 # --- Абстрактный базовый класс (Контракт) ---
 
 from abc import ABC, abstractmethod
+from datetime import datetime, time
 
 import openpyxl
 import xlrd
@@ -10,10 +11,19 @@ from coord import Coord, Merged
 EMPTY_CTYPES = [xlrd.XL_CELL_EMPTY, xlrd.XL_CELL_BLANK]
 
 class TranschendentnostCell:
-    def __init__(self, value, is_empty):
+    def __init__(self, value, is_empty, is_time=False):
         self.value = value
+        self.is_time = isinstance(value, time) or is_time
         self._is_empty = is_empty
         
+    def is_nospace_nocase_same(self, query):
+        try:
+            if self.value.lower().replace(" ", "").strip() == query.lower().replace(" ", "").strip():
+                return True
+        except: pass
+        
+        return False
+
     def is_empty(self):
         return self._is_empty
 
@@ -29,6 +39,10 @@ class ExcelSheetReader(ABC):
     def get_sheet_index(self):
         pass
     
+    @abstractmethod
+    def get_sheet_name(self):
+        pass
+
     @abstractmethod
     def has_next_sheet(self):
         pass
@@ -71,16 +85,28 @@ class ExcelSheetReader(ABC):
         return "TODO: info"
     
     @abstractmethod
-    def cell(self, row, col):
+    def cell(self, row, col) -> TranschendentnostCell:
         """Возвращает абстрактную клетку"""
         pass
     
-    def find(self, query = None):
+    def find(self, query = None, startswith=False, nospace=False):
+        return self.find_any([query], startswith=startswith, nospace=nospace)
+    
+    def find_any(self, query = None, startswith=False, nospace=False):
         for rx in range(self.get_row_count()):
             i = 0
             for x in self.get_row_values(rx):
-                if x == query:
+                if nospace:
+                    x = str(x).replace(" ", "").strip()
+                
+                for query_selected in query:
+                    if x == query_selected:
                         return Coord(rx, i)
+                    elif startswith:
+                        try:
+                            if str(x).lower().startswith(query_selected.lower()):
+                                return Coord(rx, i)
+                        except: pass
                 i += 1
 
         return None
@@ -118,6 +144,9 @@ class XlrdSheetReader(ExcelSheetReader):
     def init_sheet(self):
         self.sheet = self.book.sheet_by_index(self.sheet_index)
         
+    def get_sheet_name(self):
+        return self.sheet.name
+
     def has_next_sheet(self):
         return self.sheet_index < len(self.book.sheet_names())-1
 
@@ -140,7 +169,24 @@ class XlrdSheetReader(ExcelSheetReader):
     def cell(self, row, col):
         """Возвращает абстрактную клетку"""
         c = self.sheet.cell(row, col)
-        return TranschendentnostCell(c.value, c.ctype in EMPTY_CTYPES)
+        is_empty = c.ctype in EMPTY_CTYPES
+        is_time = c.ctype == xlrd.XL_CELL_DATE
+        value = c.value
+        if is_empty:
+            value = ""
+        elif is_time:
+            if isinstance(value, float):
+                if value <= 1:
+                    seconds = round(value * 86400)
+                    minutes, seconds = divmod(seconds, 60)
+                    hours, minutes = divmod(minutes, 60)
+                    value = time(hour=hours, second=seconds, minute=minutes)
+                else:
+                    print(f"TODO: value is {value} its unix? not 0.xxxxxxxx")
+            else:
+                is_time = False
+                print("IsTime but not float!")
+        return TranschendentnostCell(value, is_empty, is_time=is_time)
 
     def get_border_style(self, coord: Coord, side):
         row = coord.row
@@ -192,6 +238,9 @@ class OpenpyxlSheetReader(ExcelSheetReader):
     def get_sheet_index(self):
         return self.sheet_index
     
+    def get_sheet_name(self):
+        return self.workbook.sheetnames[self.sheet_index]
+    
     def has_next_sheet(self):
         return self.sheet_index < len(self.workbook.sheetnames)-1
 
@@ -221,7 +270,7 @@ class OpenpyxlSheetReader(ExcelSheetReader):
         c = self._get_cell(row, col)
         
         is_empty = (c.value is None)
-        return TranschendentnostCell("" if is_empty else c.value, is_empty)
+        return TranschendentnostCell("" if is_empty else c.value, is_empty, is_time=isinstance(c.value, time))
 
     def get_cell_value(self, row, col):
         cell = self._get_cell(row, col)

utils.py

@@ -53,7 +53,7 @@ def remove_from_list(l: list, todel: list):
 
     return l
 
-def parse_all_dirt(reader: "ExcelSheetReader", min_pos: Coord, right, down):
+def parse_all_dirt(reader: "ExcelSheetReader", min_pos: Coord, right, down, with_cells=False):
     RET = set()
 
     row = min_pos.row
@@ -61,10 +61,9 @@ def parse_all_dirt(reader: "ExcelSheetReader", min_pos: Coord, right, down):
         col = min_pos.col
         while col < min_pos.col + right:
             #print(excel_coordinate(row, col))
-            cv = reader.get_cell_value(row, col)
+            cv = reader.cell(row, col)
-            value = str(cv).strip()
+            if cv is not None and not cv.is_empty():
-            if cv is not None and len(value) > 0:
+                RET.add(cv if with_cells else str(cv.value))
-                RET.add(value)
             col += 1
         row += 1
 
@@ -165,7 +164,7 @@ def find(sh, query = None):
     return None
 
 def weekday_to_num(st: str):
-    if st.upper().strip() == "ПОНЕДЕЛЬНИК":
+    if st.upper().strip().startswith("ПОНЕД"):
         return 1
     if st.upper().strip() == "ВТОРНИК":
         return 2
@@ -177,7 +176,7 @@ def weekday_to_num(st: str):
         return 5
     if st.upper().strip() == "СУББОТА":
         return 6
-    if st.upper().strip() == "ВОСКРЕСЕНЬЕ":
+    if st.upper().strip().startswith("ВОСКР"):
         return 7
     
     return -1