gctools/database/db_create.py

import json
import os
from dataclasses import dataclass, field
from pathlib import Path

import pandas as pd
import pyodbc


@dataclass
class DsnConfig:
    user: str = "sa"
    password: str = "Mffu3011#"
    server: str = "LOCALHOST\\GLOBALCUBE"
    database: str = "CARLO"
    driver: str = "mssql"
    schema: str = "import"


@dataclass
class DbCreateConfig:
    name: str = "CARLO"
    csv_file: str = "CARLO.csv"
    clients: dict[str, str] = field(default_factory=lambda: {"1": "M und S Fahrzeughandel GmbH"})
    filter: list[str] = (["2018-01-01T00:00:00", "2022-01-01T00:00:00"],)
    source_dsn: DsnConfig = None
    target_dsn: DsnConfig = None
    stage_dir: str = "..\\temp"
    batch_dir: str = "..\\batch"
    logs_dir: str = "..\\logs"


class database_inspect:
    tables = []

    def __init__(self, dsn: DsnConfig, source=False):
        self.dsn = dsn
        self.type = "SOURCE" if source else "DEST"
        self.cursor = self.connect()

    @property
    def conn_string(self):
        if self.dsn.driver == "mssql":
            return ";".join(
                [
                    "Driver={SQL Server Native Client 11.0}",
                    f"Server={self.dsn.server}",
                    f"Database={self.dsn.database}",
                    f"Uid={self.dsn.user}",
                    f"Pwd={self.dsn.password}",
                ]
            )
        if self.dsn.driver == "mysql":
            return f"mysql+pymysql://{self.dsn.user}:{self.dsn.password}@{self.dsn.server}/{self.dsn.database}?charset=utf8mb4"
        return ";".join(
            [
                "Driver={PostgreSQL Unicode}",
                f"Server={self.dsn.server}",
                "Port=5432",
                f"Database={self.dsn.database}",
                f"Uid={self.dsn.user}",
                f"Pwd={self.dsn.password}",
            ]
        )
        # f"DSN={self.dsn.server};UID={self.dsn.user};PWD={self.dsn.password}"

    @property
    def conn_ini(self):
        return "\r\n".join(
            [
                f'{self.type}_SERVER="{self.dsn.server}"',
                f'{self.type}_USER="{self.dsn.user}"',
                f'{self.type}_PASSWORD="{self.dsn.password}"',
                f'{self.type}_DATABASE="{self.dsn.database}"',
            ]
        )

    @property
    def bcp_conn_params(self):
        return f"-S {self.dsn.server} -d {self.dsn.database} -U {self.dsn.user} -P {self.dsn.password}"

    def connect(self):
        c = pyodbc.connect(self.conn_string)
        return c.cursor()

    def get_tables(self):
        tables = [x[2] for x in self.cursor.tables(tableType="TABLE")]
        views = [x[2] for x in self.cursor.tables(tableType="VIEW")]
        self.tables = tables + views
        return self.tables

    def get_prefix(self):
        if (len(self.tables)) == 0:
            self.get_tables()
        source_tables_prefix = dict(enumerate(sorted(list(set([t.split("$")[0] for t in self.tables if "$" in t]))), 1))
        if len(source_tables_prefix) == 0:
            q = self.cursor.execute("select name FROM sys.databases")
            source_tables_prefix = [x[0] for x in q.fetchall()]
        return source_tables_prefix

    def get_columns(self, table):
        source_insp_cols = [col.column_name for col in self.cursor.columns(table=table)]
        if len(source_insp_cols) == 0:
            q = self.cursor.execute(
                "SELECT COLUMN_NAME as column_name FROM information_schema.columns "
                + f"WHERE TABLE_NAME = '{self.convert_table(table)}'"
            )
            source_insp_cols = [col[0] for col in q.fetchall()]
        return source_insp_cols

    def get_columns_is_typeof_str(self, table):
        source_insp_cols = [
            col.data_type in [pyodbc.SQL_CHAR, pyodbc.SQL_VARCHAR] for col in self.cursor.columns(table=table)
        ]
        if len(source_insp_cols) == 0:
            q = self.cursor.execute(
                "SELECT COLLATION_NAME as column_collation FROM information_schema.columns "
                + f"WHERE TABLE_NAME = '{self.convert_table(table)}'"
            )
            source_insp_cols = [len(col[0]) > 0 for col in q.fetchall()]
        return source_insp_cols

    def get_pkey(self, table):
        source_insp_cols = [col.column_name for col in self.cursor.primaryKeys(table=table)]
        if len(source_insp_cols) == 0:
            q = self.cursor.execute(
                "SELECT COLUMN_NAME "
                "FROM INFORMATION_SCHEMA.KEY_COLUMN_USAGE "
                "WHERE OBJECTPROPERTY(OBJECT_ID(CONSTRAINT_SCHEMA + '.' + QUOTENAME(CONSTRAINT_NAME)), 'IsPrimaryKey') = 1 "
                f"AND TABLE_NAME = '{self.convert_table(table)}' "  # AND TABLE_SCHEMA = 'dbo'"
            )
            source_insp_cols = [col[0] for col in q.fetchall()]
        return source_insp_cols

    def convert_table(self, table):
        if "." in table:
            table = table.split(".")[-1]
        if "[" in table:
            table = table[1:-1]
        return table


def load_config(config_file: str):
    cfg_import = json.load(open(config_file, "r", encoding="latin-1"))
    base_dir = Path(config_file).resolve().parent
    cfg_import["name"] = Path(config_file).stem
    if "logs_dir" not in cfg_import:
        cfg_import["logs_dir"] = "..\\logs"

    for folder in ["stage_dir", "batch_dir", "logs_dir"]:
        if cfg_import[folder].startswith(".."):
            cfg_import[folder] = str(base_dir.joinpath(cfg_import[folder]).resolve())
        os.makedirs(cfg_import[folder], exist_ok=True)
    cfg_import["source_dsn"] = DsnConfig(**cfg_import["source_dsn"])
    cfg_import["target_dsn"] = DsnConfig(**cfg_import["target_dsn"])
    return DbCreateConfig(**cfg_import)


def get_import_config(filename: str, db_name: str):
    df = pd.read_csv(filename, sep=";", encoding="latin-1")
    if "cols" not in df.columns:
        df["target_db"] = db_name
        df["cols"] = ""
        df[["source", "target", "target_db", "filter", "query", "iterative", "cols"]].to_csv(
            filename, sep=";", encoding="latin-1", index=False
        )
    return df[df["target"].notnull()]


def create(config_file: str = "database/CARLO.json"):
    cfg = load_config(config_file)
    base_dir = str(Path(cfg.batch_dir).parent)
    config = get_import_config(f"{base_dir}/config/{cfg.csv_file}", cfg.target_dsn.database)

    source_db = database_inspect(cfg.source_dsn, source=True)
    source_tables = source_db.get_tables()
    print(json.dumps(source_db.get_prefix(), indent=2))

    target_db = database_inspect(cfg.target_dsn)
    target_tables = target_db.get_tables()

    for _, current_table in config.iterrows():
        with open(f"{cfg.batch_dir}/{current_table['target']}.bat", "w", encoding="cp850") as f:
            f.write("@echo off \n")
            f.write("rem ==" + current_table["target"] + "==\n")

            if not current_table["target"] in target_tables:
                f.write(f"echo Ziel-Tabelle '{current_table['target']}' existiert nicht!\n")
                print(f"Ziel-Tabelle '{current_table['target']}' existiert nicht!")
                continue

            f.write(f"del {cfg.stage_dir}\\{current_table['target']}*.* /Q /F >nul 2>nul \n")
            f.write(
                f"sqlcmd.exe {target_db.bcp_conn_params} -p "
                + f"-Q \"TRUNCATE TABLE [{cfg.target_dsn.schema}].[{current_table['target']}]\" \n"
            )

            target_columns_list = target_db.get_columns(current_table["target"])
            target_column_types = target_db.get_columns_is_typeof_str(current_table["target"])

            if "CLIENT_DB" in target_columns_list:
                target_columns_list.remove("CLIENT_DB")
                target_columns_list.append("Client_DB")
            target_columns = set(target_columns_list)

            for client_db, prefix in cfg.clients.items():
                source_table = current_table["source"].format(prefix)
                if source_table not in source_tables:
                    source_table2 = source_db.convert_table(source_table)
                    if source_table2 not in source_tables:
                        f.write(f"echo Quell-Tabelle '{source_table}' existiert nicht!\n")
                        print(f"Quell-Tabelle '{source_table}' existiert nicht!")
                        continue

                source_columns = set(source_db.get_columns(source_table))

                intersect = source_columns.intersection(target_columns)
                # print("Auf beiden Seiten: " + ";".join(intersect))
                diff1 = source_columns.difference(target_columns)
                if len(diff1) > 0:
                    f.write("rem Nur in Quelle: " + ";".join(diff1) + "\n")
                diff2 = target_columns.difference(source_columns)
                if "Client_DB" not in diff2:
                    f.write("echo Spalte 'Client_DB' fehlt!\n")
                    print(f"Ziel-Tabelle '{current_table['target']}' Spalte 'Client_DB' fehlt!")
                    continue
                diff2.remove("Client_DB")
                if len(diff2) > 0:
                    f.write("rem Nur in Ziel:   " + ";".join(diff2) + "\n")

                if not pd.isnull(current_table["query"]):
                    select_query = current_table["query"].format(prefix, cfg.filter[0], cfg.filter[1])
                elif "." in source_table or cfg.source_dsn.schema == "":
                    if source_table[0] != "[":
                        source_table = f"[{source_table}]"
                    select_query = f"SELECT T1.* FROM {source_table} T1 "
                else:
                    select_query = f"SELECT T1.* FROM [{cfg.source_dsn.schema}].[{source_table}] T1 "

                if not pd.isnull(current_table["filter"]):
                    select_query += " WHERE " + current_table["filter"].format("", cfg.filter[0], cfg.filter[1])
                # select_columns = "T1.[" + "], T1.[".join(intersect) + "],"
                select_columns = ""
                for col, is_char_type in zip(target_columns_list, target_column_types):
                    if col in intersect:
                        if is_char_type:
                            select_columns += f"dbo.cln(T1.[{col}]), "
                        else:
                            select_columns += f"T1.[{col}], "
                    elif col == "Client_DB":
                        select_columns += f"'{client_db}' as \\\"Client_DB\\\", "
                    else:
                        select_columns += "'' as \\\"" + col + '\\", '

                select_query = select_query.replace("T1.*", select_columns[:-2])
                if "timestamp" in source_columns:
                    select_query += " ORDER BY T1.[timestamp] "
                else:
                    print(current_table["target"] + " hat kein timestamp-Feld")
                pkey = target_db.get_pkey(current_table["target"])
                if len(pkey) == 0:
                    print(current_table["target"] + " hat keinen Primaerschluessel")
                select_query = select_query.replace("%", "%%")  # batch-Problem

                stage_csv = f"{cfg.stage_dir}\\{current_table['target']}_{client_db}.csv"
                logfile = f"{cfg.logs_dir}\\{current_table['target']}_{client_db}"
                # insert_query = f"LOAD DATA INFILE '{stage_csv}' INTO TABLE {current_table['target']} FIELDS TERMINATED BY ','
                # ENCLOSED BY '\"' LINES TERMINATED BY '\n';"

                # print(select_query)
                bulk_copy = "bcp" if cfg.source_dsn.driver == "mssql" else "cet"
                f.write(
                    f'{bulk_copy} "{select_query}" queryout "{stage_csv}" {source_db.bcp_conn_params} -c -C 65001 -m 1000 '
                    + f'-e "{logfile}.queryout.log" > "{logfile}.bcp1.log" \n'
                )
                f.write(f'type "{logfile}.bcp1.log" | findstr -v "1000" \n')
                f.write(
                    f"bcp [{cfg.target_dsn.schema}].[{current_table['target']}] in \"{stage_csv}\" {target_db.bcp_conn_params} "
                    + f'-c -C 65001 -m 1000 -e "{logfile}.in.log" > "{logfile}.bcp2.log" \n'
                )
                f.write(f'type "{logfile}.bcp2.log" | findstr -v "1000" \n')

                f.write(f'del "{stage_csv}" /F >nul 2>nul \n')

    with open(f"{cfg.batch_dir}/_{cfg.name}.bat", "w", encoding="cp850") as f:
        f.write("@echo off & cd /d %~dp0 \n")
        f.write(f"del {cfg.stage_dir}\\*.* /Q /F >nul 2>nul \n\n")
        for index, current_table in config.iterrows():
            f.write(f"echo =={current_table['target']}==\n")
            f.write(f"echo {current_table['target']} >CON \n")
            f.write(f"call {current_table['target']}.bat\n\n")


if __name__ == "__main__":
    create()