Files
scud_ai/services/scud_export.py
T

400 lines
17 KiB
Python
Raw Blame History

This file contains ambiguous Unicode characters
This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.
"""
===============================================================================
FILE: services/scud_export.py
ROLE: Прямой экспорт данных СКУД Орион (MS SQL) в SQLite и чистый Excel (XlsxWriter).
Корректная фильтрация транзитных проходов турникетов парковки и двора.
===============================================================================
"""
import argparse
import logging
import os
import sys
import warnings
from datetime import datetime, timedelta
CURRENT_DIR = os.path.dirname(os.path.abspath(__file__))
ROOT_DIR = os.path.abspath(os.path.join(CURRENT_DIR, ".."))
if ROOT_DIR not in sys.path:
sys.path.insert(0, ROOT_DIR)
import pandas as pd
import pyodbc
import xlsxwriter
from config import SCUD_DIR, clean_scud_fio_light, load_exceptions
from core.database import save_scud_to_db, has_scud_logs_for_date, has_yesterday_final_snapshot
from core.repositories.scud_repo import save_raw_events_to_db
warnings.filterwarnings("ignore", message="pandas only supports SQLAlchemy connectable")
SCRIPT_DIR = os.path.dirname(os.path.abspath(__file__))
LOG_DIR = os.path.join(SCRIPT_DIR, "..", "logs")
os.makedirs(LOG_DIR, exist_ok=True)
TODAY_DATE_STR = datetime.now().strftime("%d.%m.%Y")
LOG_FILE = os.path.join(LOG_DIR, f"export_{TODAY_DATE_STR}.log")
logger = logging.getLogger("scud_export")
logger.setLevel(logging.INFO)
logger.propagate = False
if not logger.handlers:
try:
_file_handler = logging.FileHandler(LOG_FILE, encoding="utf-8")
_formatter = logging.Formatter("[%(asctime)s] [%(levelname)s] %(message)s", datefmt="%Y-%m-%d %H:%M:%S")
_file_handler.setFormatter(_formatter)
logger.addHandler(_file_handler)
except (PermissionError, OSError) as e:
sys.stderr.write(f"Предупреждение: невозможно создать лог-файл {LOG_FILE}: {e}\n")
_console_handler = logging.StreamHandler(sys.stdout)
_formatter = logging.Formatter("[%(asctime)s] [%(levelname)s] %(message)s", datefmt="%Y-%m-%d %H:%M:%S")
_console_handler.setFormatter(_formatter)
logger.addHandler(_console_handler)
def log(message: str, level: str = "INFO"):
level_map = {
"INFO": logging.INFO,
"ERROR": logging.ERROR,
"WARNING": logging.WARNING,
"SUCCESS": logging.INFO,
}
if level == "SUCCESS":
logger.info(f"[SUCCESS] {message}")
else:
logger.log(level_map.get(level, logging.INFO), message)
SERVER_NAME = r"172.16.200.147\SQL"
DATABASE_NAME = "Orion-14.01.21-1"
SQL_USER = "sa"
SQL_PASSWORD = "123456"
ODBC_DRIVER = "ODBC Driver 18 for SQL Server"
SQL_QUERY_TEMPLATE = r"""
DECLARE @InputDate DATE = '{target_date}';
DECLARE @TargetDate DATE = @InputDate;
DECLARE @StartDate DATETIME = CAST(@TargetDate AS DATETIME);
DECLARE @EndDate DATETIME = DATEADD(SECOND, -1, DATEADD(DAY, 1, @StartDate));
WITH DailyLogs AS (
SELECT
log.HozOrgan AS EmployeeID,
log.TimeVal,
log.Event,
log.Mode,
-- Приоритет отдается физическому направлению контроллера (Mode):
CASE
WHEN log.Mode = 2 THEN 'OUT'
WHEN log.Mode = 1 THEN 'IN'
WHEN log.Event IN (2, 27, 29, 33, 55, 65) THEN 'OUT'
WHEN log.Event IN (1, 21, 26, 54, 64) THEN 'IN'
ELSE 'OTHER'
END AS Direction,
-- Вычисляем самое последнее событие сотрудника за день:
ROW_NUMBER() OVER (PARTITION BY log.HozOrgan ORDER BY log.TimeVal DESC, log.Event DESC) AS RnLast
FROM pLogData log WITH (NOLOCK)
WHERE log.TimeVal BETWEEN @StartDate AND @EndDate
AND log.HozOrgan IS NOT NULL
AND log.HozOrgan > 0
AND log.Event IN (1, 2, 21, 26, 27, 28, 29, 32, 33, 54, 55, 64, 65)
),
Passages AS (
SELECT
EmployeeID,
MIN(TimeVal) AS FirstRawEvent,
MAX(TimeVal) AS LastRawEvent,
-- Первый вход за день:
MIN(CASE WHEN Direction = 'IN' THEN TimeVal END) AS FirstIn,
-- Фиксируем выход ТОЛЬКО если самое последнее событие за день было именно выходом (OUT):
MAX(CASE WHEN RnLast = 1 AND Direction = 'OUT' THEN TimeVal END) AS FinalOut
FROM DailyLogs
GROUP BY EmployeeID
),
EvaluatedPassages AS (
SELECT
p.*,
-- Время выхода проставляется только при окончательном уходе сотрудника из здания:
CASE
WHEN p.FinalOut IS NOT NULL
AND p.FirstIn IS NOT NULL
AND p.FinalOut > DATEADD(MINUTE, 5, p.FirstIn)
THEN p.FinalOut
ELSE NULL
END AS FilteredLastOut
FROM Passages p
)
SELECT
N'ЛЕНМОРНИИПРОЕКТ' AS [Фирма],
ISNULL(CAST(div.Name AS NVARCHAR(255)), N'Без подразделения') AS [Подразделение],
LTRIM(RTRIM(
ISNULL(CAST(p.Name AS NVARCHAR(255)), N'') +
CASE WHEN p.FirstName IS NOT NULL AND CAST(p.FirstName AS NVARCHAR(255)) <> ''
THEN N' ' + CAST(p.FirstName AS NVARCHAR(255)) ELSE N'' END +
CASE WHEN p.MidName IS NOT NULL AND CAST(p.MidName AS NVARCHAR(255)) <> ''
THEN N' ' + CAST(p.MidName AS NVARCHAR(255)) ELSE N'' END
)) AS [Сотрудник],
ISNULL(CAST(post.Name AS NVARCHAR(255)), N'—') AS [Должность],
ISNULL(CAST(p.TabNumber AS NVARCHAR(50)), N'—') AS [Таб_№],
CONVERT(VARCHAR(10), @TargetDate, 104) AS [Дата],
ISNULL(CAST(CONVERT(VARCHAR(8), pass.FirstIn, 108) AS NVARCHAR(20)), N'Нет входа') AS [Начало_дня],
CASE
WHEN pass.FirstIn IS NULL AND pass.FirstRawEvent IS NOT NULL
THEN CAST(CONVERT(VARCHAR(8), pass.FirstRawEvent, 108) AS NVARCHAR(20))
ELSE N'—'
END AS [Первая_активность],
CASE
WHEN pass.FilteredLastOut IS NOT NULL
THEN CAST(CONVERT(VARCHAR(8), pass.FilteredLastOut, 108) AS NVARCHAR(20))
ELSE N'Нет выхода'
END AS [Конец_дня],
CASE
WHEN pass.EmployeeID IS NOT NULL AND (pass.FirstIn IS NOT NULL OR pass.FirstRawEvent IS NOT NULL) THEN
RIGHT('0' + CAST(DATEDIFF(MINUTE,
ISNULL(pass.FirstIn, pass.FirstRawEvent),
CASE
WHEN pass.FilteredLastOut IS NOT NULL THEN pass.FilteredLastOut
WHEN @TargetDate = CAST(GETDATE() AS DATE) THEN GETDATE()
ELSE ISNULL(pass.FirstIn, pass.FirstRawEvent)
END) / 60 AS VARCHAR), 2) + ':' +
RIGHT('0' + CAST(DATEDIFF(MINUTE,
ISNULL(pass.FirstIn, pass.FirstRawEvent),
CASE
WHEN pass.FilteredLastOut IS NOT NULL THEN pass.FilteredLastOut
WHEN @TargetDate = CAST(GETDATE() AS DATE) THEN GETDATE()
ELSE ISNULL(pass.FirstIn, pass.FirstRawEvent)
END) % 60 AS VARCHAR), 2)
ELSE N'00:00'
END AS [Находился_в_здании],
CASE
WHEN pass.EmployeeID IS NOT NULL THEN N'Присутствовал'
ELSE N'Отсутствовал (Нет событий)'
END AS [Статус]
FROM pList p WITH (NOLOCK)
LEFT JOIN PDivision div WITH (NOLOCK) ON p.Section = div.ID
LEFT JOIN PPost post WITH (NOLOCK) ON p.Post = post.ID
LEFT JOIN EvaluatedPassages pass ON p.ID = pass.EmployeeID
WHERE
ISNULL(p.StatusRecord, 0) = 0
AND p.DateTimeInArchive IS NULL
AND ISNULL(CAST(div.Name AS NVARCHAR(255)), N'') NOT LIKE N'Аренд%'
AND ISNULL(CAST(div.Name AS NVARCHAR(255)), N'') NOT IN (N'Без подразделения', N'')
AND p.Name NOT LIKE N'бр.%'
AND p.Name NOT LIKE N'Гость%'
AND ISNULL(CAST(div.Name AS NVARCHAR(255)), N'') NOT IN (N'БГИ', N'КНР')
AND ISNULL(CAST(div.Name AS NVARCHAR(255)), N'') NOT LIKE N'Рабоч%'
AND ISNULL(CAST(div.Name AS NVARCHAR(255)), N'') NOT LIKE N'Врем%'
AND ISNULL(CAST(div.Name AS NVARCHAR(255)), N'') NOT LIKE N'Практика%'
AND ISNULL(CAST(div.Name AS NVARCHAR(255)), N'') NOT LIKE N'тест%'
AND ISNULL(CAST(post.Name AS NVARCHAR(255)), N'') NOT LIKE N'Практикант%'
ORDER BY p.Name ASC;
"""
SQL_RAW_EVENTS_QUERY = r"""
DECLARE @InputDate DATE = '{target_date}';
DECLARE @StartDate DATETIME = CAST(@InputDate AS DATETIME);
DECLARE @EndDate DATETIME = DATEADD(SECOND, -1, DATEADD(DAY, 1, @StartDate));
SELECT
log.TimeVal,
log.HozOrgan,
LTRIM(RTRIM(
ISNULL(CAST(p.Name AS NVARCHAR(255)), N'') +
CASE WHEN p.FirstName IS NOT NULL AND CAST(p.FirstName AS NVARCHAR(255)) <> ''
THEN N' ' + CAST(p.FirstName AS NVARCHAR(255)) ELSE N'' END +
CASE WHEN p.MidName IS NOT NULL AND CAST(p.MidName AS NVARCHAR(255)) <> ''
THEN N' ' + CAST(p.MidName AS NVARCHAR(255)) ELSE N'' END
)) AS [Сотрудник],
ISNULL(CAST(div.Name AS NVARCHAR(255)), N'Без подразделения') AS [Подразделение],
log.Event,
log.Mode,
CASE
WHEN log.Mode = 2 THEN 'OUT'
WHEN log.Mode = 1 THEN 'IN'
WHEN log.Event IN (2, 27, 29, 33, 55, 65) THEN 'OUT'
WHEN log.Event IN (1, 21, 26, 54, 64) THEN 'IN'
ELSE 'OTHER'
END AS Direction
FROM pLogData log WITH (NOLOCK)
INNER JOIN pList p WITH (NOLOCK) ON log.HozOrgan = p.ID
LEFT JOIN PDivision div WITH (NOLOCK) ON p.Section = div.ID
WHERE log.TimeVal BETWEEN @StartDate AND @EndDate
AND log.HozOrgan IS NOT NULL
AND log.HozOrgan > 0
AND log.Event IN (28, 32)
AND ISNULL(p.StatusRecord, 0) = 0
ORDER BY log.TimeVal ASC;
"""
def save_df_to_clean_excel(df: pd.DataFrame, file_path: str, sheet_name: str = "Отчет"):
workbook = xlsxwriter.Workbook(file_path, {'constant_memory': False})
worksheet = workbook.add_worksheet(sheet_name)
fmt_header = workbook.add_format({
'bold': True,
'bg_color': '#D9E1F2',
'border': 1,
'border_color': '#D3D3D3',
'align': 'center',
'valign': 'vcenter',
'font_name': 'Calibri',
'font_size': 11
})
fmt_cell = workbook.add_format({
'border': 1,
'border_color': '#D3D3D3',
'valign': 'vcenter',
'align': 'left',
'font_name': 'Calibri',
'font_size': 11
})
headers = list(df.columns)
col_widths = [len(str(h)) for h in headers]
for col_idx, header in enumerate(headers):
worksheet.write(0, col_idx, str(header), fmt_header)
for row_idx, row_values in enumerate(df.values, start=1):
for col_idx, val in enumerate(row_values):
if pd.isna(val) or val is None:
val_str = ""
elif isinstance(val, bool):
val_str = "Да" if val else "Нет"
else:
val_str = str(val)
worksheet.write(row_idx, col_idx, val_str, fmt_cell)
if len(val_str) > col_widths[col_idx]:
col_widths[col_idx] = len(val_str)
for col_idx, width in enumerate(col_widths):
worksheet.set_column(col_idx, col_idx, min(max(width + 3, 10), 45))
workbook.close()
def get_targets(input_date: str | None):
targets = []
if input_date:
try:
parsed = datetime.strptime(input_date, "%d.%m.%Y").date()
targets.append({"name": "Указанная дата", "date": parsed})
except ValueError:
log(f"ОШИБКА: Неверный формат даты '{input_date}'. Используйте ДД.ММ.ГГГГ", "ERROR")
sys.exit(1)
else:
now = datetime.now()
yesterday = (now - timedelta(days=3 if now.weekday() == 0 else 1)).date()
today = now.date()
targets.append({"name": "Вчера", "date": yesterday})
targets.append({"name": "Сегодня", "date": today})
return targets
def run_export(input_date: str | None = None, save_xlsx: bool = True, debug: bool = False):
if debug:
logger.setLevel(logging.DEBUG)
log("=== ВКЛЮЧЕН РЕЖИМ ОТЛАДКИ (DEBUG MODE) ===", "WARNING")
log("=== [ЭТАП 0] Выгрузка свежих данных СКУД напрямую из БД Орион ===")
os.makedirs(SCUD_DIR, exist_ok=True)
targets = get_targets(input_date)
conn_str = (
f"DRIVER={{{ODBC_DRIVER}}};"
f"SERVER={SERVER_NAME};"
f"DATABASE={DATABASE_NAME};"
f"UID={SQL_USER};"
f"PWD={SQL_PASSWORD};"
f"TrustServerCertificate=yes;"
f"Encrypt=no;"
)
success = True
for target in targets:
processing_date = target["date"]
processing_date_str = processing_date.strftime("%d.%m.%Y")
period_label = target["name"]
is_yesterday = (period_label == "Вчера")
if is_yesterday and has_yesterday_final_snapshot(processing_date_str):
log(f"[ℹ️] Вчерашний день ({processing_date_str}) уже зафиксирован финишным снапшотом Y. Пропускаем запрос к MS SQL.")
continue
if is_yesterday:
snapshot_time = f"{processing_date.strftime('%Y-%m-%d')} 23:59:59"
else:
snapshot_time = datetime.now().strftime("%Y-%m-%d %H:%M:%S")
log(f"--- Обработка периода: {period_label} ({processing_date_str}) --- [Снапшот: {snapshot_time}]")
sql_query = SQL_QUERY_TEMPLATE.format(target_date=processing_date.strftime("%Y-%m-%d"))
connection = None
try:
connection = pyodbc.connect(conn_str, timeout=120)
df = pd.read_sql(sql_query, connection)
log(f"База вернула {len(df)} строк за {processing_date_str}")
if len(df) > 0:
df['fio_clean'] = df['Сотрудник'].apply(clean_scud_fio_light)
df['anomaly_flag'] = 'NONE'
mask_anomaly = (df['Начало_дня'] == 'Нет входа') & (df['Первая_активность'] != '—')
df.loc[mask_anomaly, 'anomaly_flag'] = 'ANOMALY_NO_IN_HAS_ACTIVITY'
df['Пришел'] = df['Статус'].str.contains('Присутствовал', case=False, na=False) & (~mask_anomaly)
save_scud_to_db(df, processing_date_str, snapshot_time=snapshot_time, is_yesterday=is_yesterday)
raw_sql = SQL_RAW_EVENTS_QUERY.format(target_date=processing_date.strftime("%Y-%m-%d"))
df_raw = pd.read_sql(raw_sql, connection)
if len(df_raw) > 0:
df_raw['fio_clean'] = df_raw['Сотрудник'].apply(clean_scud_fio_light)
inserted_count = save_raw_events_to_db(df_raw, processing_date_str)
log(f"[✓] В scud_events_raw сохранено {inserted_count} сырых событий проходов за {processing_date_str}!", "SUCCESS")
log(f"[✓] Записи за {processing_date_str} успешно сохранены в SQLite!", "SUCCESS")
if save_xlsx:
file_name = f"Сотрудники_{processing_date_str}.xlsx"
file_path = os.path.join(SCUD_DIR, file_name)
if os.path.exists(file_path):
try:
os.remove(file_path)
except OSError as e:
log(f"ОШИБКА при удалении старого файла {file_name}: {e}", "ERROR")
save_df_to_clean_excel(df, file_path, sheet_name="Отчет")
log(f"[✓] Успешно экспортирован файл: data/scud/{file_name}", "SUCCESS")
else:
log(f"Запрос за {processing_date_str} вернул 0 строк.", "WARNING")
except Exception as e:
log(f"🛑 ОШИБКА выгрузки СКУД за {processing_date_str}: {e}", "ERROR")
success = False
finally:
if connection is not None:
connection.close()
log("=== Выгрузка СКУД завершена ===\n")
return success
if __name__ == "__main__":
parser = argparse.ArgumentParser()
parser.add_argument("--date", dest="input_date", default=None)
parser.add_argument("-d", "--debug", action="store_true")
parser.add_argument("--no-xlsx", dest="save_xlsx", action="store_false", default=True)
args = parser.parse_args()
run_export(args.input_date, save_xlsx=args.save_xlsx, debug=args.debug)