This commit is contained in:
2026-03-04 01:21:06 +03:00
parent c7ad35d343
commit 9b004c3a86
7 changed files with 598 additions and 49 deletions
+251
View File
@@ -0,0 +1,251 @@
import argparse
import asyncio
import pandas as pd
import asyncpg
import re
import os
def clean_name(name: str) -> str:
if not isinstance(name, str):
return ""
name = re.sub(r'\s+[СC]$', '', name.strip())
name = re.sub(r'\s*\d+\s*$', '', name)
return name.strip()
def parse_excel_raw(file_path: str):
excel_file = pd.ExcelFile(file_path)
resident_hours = {}
room_hours = {}
for sheet_name in excel_file.sheet_names:
df = pd.read_excel(file_path, sheet_name=sheet_name)
# --- Парсинг жильцов ---
if 'ФИО' in df.columns and 'Часы' in df.columns:
for _, row in df.iterrows():
fio = clean_name(str(row['ФИО']))
if not fio or str(fio).lower() == 'nan':
continue
try:
hours_raw = str(row['Часы']).split('(')[0].strip()
amount = int(float(hours_raw))
except ValueError:
continue
if fio in resident_hours:
resident_hours[fio] += amount
else:
resident_hours[fio] = amount
# --- Парсинг комнат ---
elif df.iloc[0].astype(str).str.contains('Комната', case=False, na=False).any():
header_idx = df[df.apply(lambda r: r.astype(str).str.contains('Комната', case=False).any(), axis=1)].index[0]
df.columns = df.iloc[header_idx]
df = df.iloc[header_idx + 1:].reset_index(drop=True)
for _, row in df.iterrows():
room_str = str(row['Комната'])
if not room_str or room_str.lower() == 'nan':
continue
room_match = re.search(r'\b(\d{3,4})\b', room_str)
if not room_match:
continue
room_number = int(room_match.group(1))
try:
punishment_raw = str(row['Наказание']).split('(')[0].replace('часов', '').replace(',', ';').strip()
amounts = [int(x.strip()) for x in re.findall(r'\d+', punishment_raw)]
amount = sum(amounts) if amounts else 0
except (ValueError, AttributeError):
continue
if amount <= 0:
continue
if room_number in room_hours:
room_hours[room_number] += amount
else:
room_hours[room_number] = amount
return resident_hours, room_hours
async def run_parse(database_url: str, file_path: str):
raw_resident_hours, raw_room_hours = parse_excel_raw(file_path)
async with asyncpg.create_pool(database_url) as pool:
async with pool.acquire() as conn:
resident_rows = await conn.fetch("SELECT real_name FROM residents WHERE real_name IS NOT NULL;")
db_names = [row['real_name'] for row in resident_rows]
room_rows = await conn.fetch("SELECT number FROM rooms;")
db_rooms = {row['number'] for row in room_rows}
# --- Сопоставление жильцов ---
db_names_lower = {name.lower(): name for name in db_names}
matched_residents = {}
unmatched_residents = {}
for raw_fio, amount in raw_resident_hours.items():
fio_lower = raw_fio.lower()
matched_name = None
if fio_lower in db_names_lower:
matched_name = db_names_lower[fio_lower]
else:
parts = fio_lower.split()
if len(parts) >= 2:
search_key = f"{parts[0]} {parts[1]}"
if search_key in db_names_lower:
matched_name = db_names_lower[search_key]
if not matched_name:
for db_lower, db_orig in db_names_lower.items():
if fio_lower in db_lower or db_lower in fio_lower:
matched_name = db_orig
break
if matched_name:
matched_residents[matched_name] = matched_residents.get(matched_name, 0) + amount
else:
unmatched_residents[raw_fio] = unmatched_residents.get(raw_fio, 0) + amount
# --- Сопоставление комнат ---
matched_rooms = {}
unmatched_rooms = {}
for room_num, amount in raw_room_hours.items():
if room_num in db_rooms:
matched_rooms[room_num] = matched_rooms.get(room_num, 0) + amount
else:
unmatched_rooms[room_num] = unmatched_rooms.get(room_num, 0) + amount
# --- Запись в файлы ---
'''with open("parsed_residents.txt", "w", encoding="utf-8") as f:
for name, amount in sorted(matched_residents.items()):
f.write(f"{name} - {amount}\n")'''
with open("parsed_rooms.txt", "w", encoding="utf-8") as f:
for room, amount in sorted(matched_rooms.items()):
f.write(f"{room} - {amount}\n")
'''if unmatched_residents:
with open("unmatched_residents.txt", "w", encoding="utf-8") as f:
for name, amount in sorted(unmatched_residents.items()):
f.write(f"{name} - {amount}\n")'''
if unmatched_rooms:
with open("unmatched_rooms.txt", "w", encoding="utf-8") as f:
for room, amount in sorted(unmatched_rooms.items()):
f.write(f"{room} - {amount}\n")
async def run_load(database_url: str, admin_id: int):
if not os.path.exists("parsed_residents.txt") or not os.path.exists("parsed_rooms.txt"):
raise FileNotFoundError("Файлы parsed_residents.txt или parsed_rooms.txt не найдены. Сначала запустите parse.")
resident_transactions = []
with open("parsed_residents.txt", "r", encoding="utf-8") as f:
for line in f:
line = line.strip()
if not line:
continue
parts = line.split(" - ")
if len(parts) == 2:
resident_transactions.append((parts[0].strip(), int(parts[1].strip())))
room_transactions = []
with open("parsed_rooms.txt", "r", encoding="utf-8") as f:
for line in f:
line = line.strip()
if not line:
continue
parts = line.split(" - ")
if len(parts) == 2:
room_transactions.append((int(parts[0].strip()), int(parts[1].strip())))
async with asyncpg.create_pool(database_url) as pool:
async with pool.acquire() as conn:
# Маппинги
res_rows = await conn.fetch("SELECT id, real_name FROM residents WHERE real_name IS NOT NULL;")
name_to_id = {row['real_name']: row['id'] for row in res_rows}
rm_rows = await conn.fetch("SELECT id, number FROM rooms;")
num_to_id = {row['number']: row['id'] for row in rm_rows}
res_tx_params = []
res_update_params = []
for name, amount in resident_transactions:
res_id = name_to_id.get(name)
if res_id:
res_tx_params.append((res_id, "increase", amount, admin_id, "Инициализация существующих часов"))
res_update_params.append((amount, res_id))
rm_tx_params = []
rm_update_params = []
for room_num, amount in room_transactions:
rm_id = num_to_id.get(room_num)
if rm_id:
rm_tx_params.append((rm_id, "increase", amount, admin_id, "Инициализация существующих часов"))
rm_update_params.append((amount, rm_id))
async with conn.transaction():
# Транзакции резидентов
if res_tx_params:
await conn.executemany(
"""
INSERT INTO hours_transactions
(resident_id, transaction_type, amount, admin_id, remark, created_at)
VALUES ($1, $2, $3, $4, $5, CURRENT_TIMESTAMP);
""", res_tx_params
)
# Обновление баланса резидентов
await conn.executemany(
"""
UPDATE residents
SET active_hours = active_hours + $1, updated_at = CURRENT_TIMESTAMP
WHERE id = $2;
""", res_update_params
)
# Транзакции комнат
if rm_tx_params:
await conn.executemany(
"""
INSERT INTO room_hours_transactions
(room_id, transaction_type, amount, admin_id, remark, created_at)
VALUES ($1, $2, $3, $4, $5, CURRENT_TIMESTAMP);
""", rm_tx_params
)
# Обновление баланса комнат
await conn.executemany(
"""
UPDATE rooms
SET active_hours = active_hours + $1
WHERE id = $2;
""", rm_update_params
)
async def main():
parser = argparse.ArgumentParser()
parser.add_argument("command", choices=["parse", "load"])
parser.add_argument("--db-url", required=True)
parser.add_argument("--file", required=False)
parser.add_argument("--admin-id", type=int, default=2047958833)
args = parser.parse_args()
if args.command == "parse":
if not args.file:
parser.error("Для команды parse требуется указать --file")
await run_parse(args.db_url, args.file)
elif args.command == "load":
await run_load(args.db_url, args.admin_id)
if __name__ == "__main__":
asyncio.run(main())