umbreit-py/data_analysis/02-3_oracle_workflow_test.py

# %%
from __future__ import annotations

import datetime
import json
import time
import typing
from collections.abc import Sequence
from pathlib import Path
from pprint import pprint

import dopt_basics.datetime as dt
import oracledb
import polars as pl
import sqlalchemy as sql
from dopt_basics import configs, io
from sqlalchemy import event

from umbreit import db, types

# %%
# import importlib
# types = importlib.reload(types)
# db = importlib.reload(db)


# %%
p_cfg = io.search_file_iterative(
    starting_path=Path.cwd(),
    glob_pattern="CRED*.toml",
    stop_folder_name="umbreit-py",
)
assert p_cfg is not None
CFG = configs.load_toml(p_cfg)
HOST = CFG["server"]["host"]
PORT = CFG["server"]["port"]
SERVICE = CFG["server"]["service"]
USER_NAME = CFG["user"]["name"]
USER_PASS = CFG["user"]["pass"]
# %%
# !! init thick mode
# p_oracle_client = Path(r"C:\Databases\Oracle\instantclient_19_29")
# assert p_oracle_client.exists()
# assert p_oracle_client.is_dir()
# oracledb.init_oracle_client(lib_dir=str(p_oracle_client))
# %%
conn_string = (
    f"oracle+oracledb://{USER_NAME}:{USER_PASS}@{HOST}:{PORT}?service_name={SERVICE}"
)
# engine = sql.create_engine(conn_string)
engine = sql.create_engine(conn_string, execution_options={"stream_results": True})


@event.listens_for(engine, "after_cursor_execute")
def set_fetch_sizes(conn, cursor, statement, parameters, context, executemany):
    cursor.arraysize = 1000
    cursor.prefetchrows = 1000


# %%
########### RESULTS ###########
# temporary
res_engine = sql.create_engine("sqlite://")
db.metadata.create_all(res_engine, tables=(db.results,))


# %%
# delete existing results
def delete_results(
    res_engine: sql.Engine,
) -> None:
    with res_engine.begin() as conn:
        res = conn.execute(sql.delete(db.results))

    print("Rows deleted: ", res.rowcount)


delete_results(res_engine)
stmt = sql.select(db.results.c.bedarf_nr, db.results.c.bedarf_sequenz)
with res_engine.connect() as conn:
    res = conn.execute(stmt)
    print(res.all())

# %%
# define starting date for 3 month interval
# returns UTC time
current_dt = dt.current_time_tz(cut_microseconds=True)
print("Current DT: ", current_dt)
td = dt.timedelta_from_val(90, dt.TimeUnitsTimedelta.DAYS)
print("Timedelta: ", td)

start_date = (current_dt - td).date()
print("Starting date: ", start_date)

# %%
# // ---------- LIVE DATA -----------

# TODO find way to filter more efficiently
# WF-200: filter for relevant orders with current BEDP set
# missing: order types which are relevant
filter_K_rech = (608991, 260202)
join_condition = sql.and_(
    db.ext_bedpbed.c.BEDP_TITELNR == db.EXT_AUFPAUF.c.TITELNR,
    db.ext_bedpbed.c.BEDP_MAN == db.EXT_AUFPAUF.c.MANDANT,
)
join_condition = sql.and_(
    db.ext_bedpbed.c.BEDP_TITELNR == db.EXT_AUFPAUF.c.TITELNR,
)
where_condition = sql.and_(
    db.EXT_AUFPAUF.c.AUFTRAGS_DATUM > start_date,
    db.EXT_AUFPAUF.c.KUNDE_RECHNUNG.not_in(filter_K_rech),
)

stmt = (
    sql.select(
        db.ext_bedpbed.c.BEDARFNR,
        db.ext_bedpbed.c.BEDP_SEQUENZ,
        db.ext_bedpbed.c.BEDP_TITELNR,
        db.ext_bedpbed.c.BEDP_MAN,
        db.ext_bedpbed.c.BEDP_MENGE_BEDARF_VM,
        db.EXT_AUFPAUF,
    )
    .select_from(db.ext_bedpbed.join(db.EXT_AUFPAUF, join_condition))
    .where(where_condition)
    .limit(100)  # full query really slow
)

# %%
print(stmt.compile(engine))
# %%
df_order = pl.read_database(stmt, engine, schema_overrides=db.raw_data_query_schema_map)
df_order

# %%
# AUFPAUF
# stmt = sql.select(db.EXT_AUFPAUF)
# df_aufpauf = pl.read_database(stmt, engine, schema_overrides=db.raw_data_query_schema_map)
# df_aufpauf
# df_aufpauf.filter(pl.col("TITELNR") == 6315273)

# prefilter amount columns for invalid entries

# // tests with ext_bedpbed
# print("--------------- ext_bedpbed --------------")
# t1 = time.perf_counter()
# AMOUNT_COLS = frozenset(
#     (
#         "BEDP_MENGE_BEDARF",
#         "BEDP_MENGE_VERKAUF",
#         "BEDP_MENGE_ANFRAGE",
#         "BEDP_MENGE_BESTELLUNG",
#         "BEDP_MENGE_FREI",
#         "BEDP_MENGE_BEDARF_VM",
#     )
# )

# case_stmts = []
# for col in AMOUNT_COLS:
#     case_stmts.append(
#         sql.case(
#             (db.ext_bedpbed.c[col] <= -1, sql.null()),
#             else_=db.ext_bedpbed.c[col],
#         ).label(col)
#     )

# stmt = sql.select(
#     *[c for c in db.ext_bedpbed.c if c.name not in AMOUNT_COLS],
#     *case_stmts,
# )
# df = pl.read_database(stmt, engine, schema_overrides=db.ext_bedpbed_schema_map)
# t2 = time.perf_counter()
# elapsed = t2 - t1

# %%
# df.select(pl.col("BEDP_MENGE_BEDARF").is_null().sum())
print(f"Query duration: {elapsed:.4f} sec")
print("Number of entries: ", len(df))
print(f"Estimated size in memory: {df.estimated_size(unit='mb')} MB")

# %%
# try title_info parsing
stmt = sql.select(db.ext_titel_info)
print(stmt.compile(engine))


# %%
# raw data query
# TODO look for entries which do not have an associated title number

print("--------------- raw data query --------------")
t1 = time.perf_counter()
# join_condition = sql.and_(
#     db.ext_bedpbed.c.BEDP_TITELNR == db.ext_titel_info.c.TI_NUMMER,
#     db.ext_bedpbed.c.BEDP_MAN == db.ext_titel_info.c.MANDFUEHR,
# )
join_condition = sql.and_(
    db.ext_bedpbed.c.BEDP_TITELNR == db.ext_titel_info.c.TI_NUMMER,
)
stmt = sql.select(
    db.ext_bedpbed.c.BEDARFNR,
    db.ext_bedpbed.c.BEDP_SEQUENZ,
    db.ext_bedpbed.c.BEDP_TITELNR,
    db.ext_bedpbed.c.BEDP_MAN,
    sql.case(
        (db.ext_bedpbed.c.BEDP_MENGE_BEDARF_VM <= -1, sql.null()),
        else_=db.ext_bedpbed.c.BEDP_MENGE_BEDARF_VM,
    ).label("BEDP_MENGE_BEDARF_VM"),
    db.ext_titel_info.c.MELDENUMMER,
    db.ext_titel_info.c.VERLAGSNR,
    db.ext_titel_info.c.MENGE_VORMERKER,
    db.ext_titel_info.c.MANDFUEHR,
).select_from(db.ext_bedpbed.join(db.ext_titel_info, join_condition, isouter=True))

print(stmt.compile(engine))
df = pl.read_database(
    stmt,
    engine,
    schema_overrides=db.raw_data_query_schema_map,
)
t2 = time.perf_counter()
elapsed = t2 - t1
# %%
print(f"Query duration: {elapsed:.4f} sec")
print("Number of entries: ", len(df))
print(f"Estimated size in memory: {df.estimated_size(unit='mb')} MB")
# %%
df.head()

# %%
# // NO LIVE DATA NEEDED
# SAVING/LOADING
p_save = Path.cwd() / "raw_data_from_sql_query_20260114-1.arrow"
# df.write_ipc(p_save)
df = pl.read_ipc(p_save)
# %%
print(len(df))
df.head()
# %%
df.head()
# %%
# ** CHECK: duplicates
temp = df.fill_null(0)
mask = temp.select(pl.exclude(("BEDARFNR", "BEDP_SEQUENZ"))).is_duplicated()
temp.filter(mask)
# %%
df.filter(pl.col.BEDP_TITELNR.is_duplicated()).sort("BEDP_TITELNR", descending=False)
# %%
# ** CHECK: positions without titlenumber
df.filter(pl.col.VERLAGSNR.is_null())["BEDP_MAN"].unique()
# %%
# ** CHECK: unique title number?
df.group_by("BEDP_TITELNR").agg(
    pl.col("BEDP_TITELNR").len().alias("count"),
    pl.col.BEDP_MAN.unique().alias("unique_bedp_man"),
    pl.col.MANDFUEHR.unique().alias("unique_man_fuehr"),
).unique().filter(pl.col("count") > 1)
# %%
df.filter(pl.col.BEDP_TITELNR == 8679893)
# %%
df.with_columns(
    pl.col("BEDP_TITELNR").count().over("BEDP_TITELNR").alias("titlenumber_count")
).select(["BEDP_TITELNR", "titlenumber_count"]).unique().filter(
    pl.col("titlenumber_count") > 1
)
# %%
# ** CHECK: distribution of MELDENUMMER
temp = df.filter(pl.col.BEDP_MAN.is_in((1, 90)))
sum_entries = len(temp)
temp = (
    temp.group_by("MELDENUMMER")
    .agg(pl.col("MELDENUMMER").len().alias("count"))
    .sort("count", descending=True)
)

temp = temp.with_columns((pl.col.count / sum_entries).alias("proportion"))
temp = temp.with_columns(pl.col.proportion.cum_sum().alias("cum"))
temp
# df.filter(pl.col("MELDENUMMER").is_not_null() & pl.col("MELDENUMMER").is_in((17, 18))).select(
#     pl.len()
# )
# p_save = Path.cwd() / "meldenummer_anteile_20260114-2.xlsx"
# temp.write_excel(p_save)
# %%
# ** CHECK: differences MANDANT in BEDP and in TINFO
# 4591588: in title database with different MANDANT (are MANDANTFUEHR and BEDP_MAN feasible for matching?)
df.filter(pl.col("BEDP_MAN") != pl.col("MANDFUEHR")).select(pl.col("BEDP_MAN").unique())

# %%
df.group_by("BEDP_MAN").agg(pl.col("MANDFUEHR").unique())
# %%
df.filter(pl.col("MANDFUEHR").is_null()).filter(pl.col("BEDP_MAN") == 1)

# %%
# df.filter(pl.col("BEDP_MAN") != pl.col("MANDFUEHR")).filter(pl.col("BEDP_MAN") == 5)
df.filter(pl.col("BEDP_MAN") == 60).filter(pl.col("MANDFUEHR").is_null())
# %%
# ** CHECK: different MANDANTEN
# check for valid entries for unknown MANDANTEN
# MANDANTEN others than (1, 90) do not possess relevant properties such as
# "MELDENUMMER" and others --> conclusion: not relevant

# MANDANT = 80

# print(f"Mandant: {MANDANT}")
# print(
#     df.filter(pl.col("BEDP_MAN") == MANDANT).select(
#         ["BEDP_MENGE_BEDARF_VM", "MELDENUMMER", "MENGE_VORMERKER"]
#     )
# )
# print(
#     df.filter(pl.col("BEDP_MAN") == MANDANT).select(
#         ["BEDP_MENGE_BEDARF_VM", "MELDENUMMER", "MENGE_VORMERKER"]
#     ).null_count()
# )
# print("Unique value counts: ", df.select(pl.col("BEDP_MAN").value_counts()))
# %%
df.filter(pl.col("MELDENUMMER").is_null()).filter(pl.col("MANDFUEHR").is_not_null())
# %%
# ** PREFILTER
# always needed, entries filtered out are to be disposed
filter_meldenummer_null = pl.col("MELDENUMMER").is_not_null()
filter_mandant = pl.col("MANDFUEHR").is_in((1, 90))
df.filter(filter_meldenummer_null).filter(filter_mandant)

# df = df.filter(pl.col("BEDP_MAN").is_in((1, 90))).filter(pl.col("MELDENUMMER") != 26)
# %%
len(df)
# %%
# ** CHECK: null values set in the query with CASE statement
# not known if NULL because of CASE statement or already set in table
# unknown consequences: Are they relevant? How does it relate to "MENGE_VORMERKER"?
# from the title DB
df.filter(pl.col("BEDP_MENGE_BEDARF_VM").is_null())
df.filter(pl.col("BEDP_MENGE_BEDARF_VM") == 0)
# %%
df.select("MELDENUMMER").unique()
# %%
# ** CHECK: null values for "MENGE_VORMERKER"
df.filter(pl.col("MENGE_VORMERKER").is_null())
# df.filter(pl.col("BEDP_MENGE_BEDARF_VM") == 0)

agg_t = (
    df.group_by(["MELDENUMMER"]).agg(
        # pl.count("MENGE_VORMERKER").alias("pos_count").n_unique(),
        pl.col("MENGE_VORMERKER").alias("VM_count").unique(),
    )
    # .filter(pl.col("count_customer") >= 0)  # !! should be 3
)  # .filter(pl.col("MELDENUMMER") == 18)
agg_t
# %%
df.filter(pl.col("MELDENUMMER") == 18).select(pl.col("MENGE_VORMERKER").is_null().sum())

# %%
# ** CHECK: relationship between "BEDP_MENGE_BEDARF_VM" and "MENGE_VORMERKER"
# ** not known at this point
# there are entries where BEDP_MENGE_BEDARF_VM > MENGE_VORMERKER -->
# BEDP_MENGE_BEDARF_VM as reference or ground truth not suitable
df_diff_VM_bedp_tinfo = df.filter(pl.col("BEDP_MENGE_BEDARF_VM") > pl.col("MENGE_VORMERKER"))
p_save_diff_VM_bedp_tinfo = (
    Path.cwd() / "diff_BEDP-MENGE-BEDARF-VM_TINF-MENGE-VORMERKER_20251211-1.xlsx"
)
from polars.datatypes import classes as pl_dtypes

df_diff_VM_bedp_tinfo.to_pandas().to_excel(p_save_diff_VM_bedp_tinfo, index=False)
# why are there entries where "BEDP_MENGE_BEDARF_VM" > "MENGE_VORMERKER"?
# %%
# ** CHECK: titles with request where no title information is found
# result: there were entries found on 02.12., but not on 03.12.2025
not_in_title_table = df.filter(pl.col("MELDENUMMER").is_null())
EXPORT_FEAT = "BEDP_TITELNR"
to_save = {EXPORT_FEAT: not_in_title_table.select(EXPORT_FEAT).to_series().to_list()}
p_save_not_in_title_table = Path.cwd() / "not_in_title_table_20251211-1.json"
print(to_save)
# with open(p_save_not_in_title_table, "w") as file:
#     json.dump(to_save, file, indent=4)
# %%
df.group_by("BEDP_MAN").agg(pl.len())
# %%
df.filter(pl.col("MELDENUMMER").is_null()).group_by("BEDP_MAN").agg(pl.len().alias("count"))
# %%
print(len(df.filter(pl.col("MELDENUMMER") == 18)))
# df.filter(pl.col("MELDENUMMER") == 18).filter((pl.col("BEDP_MENGE_BEDARF_VM").is_not_null()) & (pl.col("BEDP_MENGE_BEDARF_VM") > 0))
# %%
# VM_CRITERION = "MENGE_VORMERKER"
VM_CRITERION = "BEDP_MENGE_BEDARF_VM"
MANDANT_CRITERION = "BEDP_MAN"


def get_starting_date(
    days: int,
) -> datetime.date:
    current_dt = dt.current_time_tz(cut_microseconds=True)
    td = dt.timedelta_from_val(days, dt.TimeUnitsTimedelta.DAYS)

    return (current_dt - td).date()


# TODO exchange to new query focusing on TINFO table
def get_raw_data() -> pl.DataFrame:
    join_condition = sql.and_(
        db.ext_bedpbed.c.BEDP_TITELNR == db.ext_titel_info.c.TI_NUMMER,
        db.ext_bedpbed.c.BEDP_MAN == db.ext_titel_info.c.MANDFUEHR,
    )
    stmt = sql.select(
        db.ext_bedpbed.c.BEDARFNR,
        db.ext_bedpbed.c.BEDP_SEQUENZ,
        db.ext_bedpbed.c.BEDP_TITELNR,
        db.ext_bedpbed.c.BEDP_MAN,
        sql.case(
            (db.ext_bedpbed.c.BEDP_MENGE_BEDARF_VM <= -1, sql.null()),
            else_=db.ext_bedpbed.c.BEDP_MENGE_BEDARF_VM,
        ).label("BEDP_MENGE_BEDARF_VM"),
        db.ext_titel_info.c.MELDENUMMER,
        db.ext_titel_info.c.MENGE_VORMERKER,
    ).select_from(db.ext_bedpbed.join(db.ext_titel_info, join_condition, isouter=True))

    return pl.read_database(
        stmt,
        engine,
        schema_overrides=db.raw_data_query_schema_map,
    )


# def get_empty_pipeline_result(
#     data: pl.DataFrame,
# ) -> PipelineResult:
#     return PipelineResult(data)


def _apply_several_filters(
    df: pl.DataFrame,
    filters: Sequence[pl.Expr],
) -> types.FilterResult:
    df_current = df
    removed_rows: list[pl.DataFrame] = []

    for filter in filters:
        removed = df_current.filter(~filter)
        removed_rows.append(removed)

        df_current = df_current.filter(filter)

    df_removed = pl.concat(removed_rows)

    return types.FilterResult(in_=df_current, out_=df_removed)


class PipelineResult:
    __slots__ = ("_results", "_open", "_subtracted_indices")
    _index_cols: tuple[str, ...] = ("BEDARFNR", "BEDP_SEQUENZ")

    def __init__(
        self,
        data: pl.DataFrame,
    ) -> None:
        self._open = data
        schema = db.results_schema_map.copy()
        del schema["id"]
        self._results = pl.DataFrame(schema=schema)

        schema = {}
        for col in self._index_cols:
            schema[col] = db.raw_data_query_schema_map[col]
        self._subtracted_indices = pl.DataFrame(schema=schema)

    def __len__(self) -> int:
        return len(self._results) + len(self._open)

    @property
    def open(self) -> pl.DataFrame:
        return self._open

    @property
    def results(self) -> pl.DataFrame:
        return self._results

    @property
    def subtracted_indices(self) -> pl.DataFrame:
        return self._subtracted_indices

    def update_open(
        self,
        data: pl.DataFrame,
    ) -> None:
        self._open = data

    def _subtract_data(
        self,
        data: pl.DataFrame,
    ) -> None:
        self._open = self._open.join(data, on=self._index_cols, how="anti")
        self._subtracted_indices = pl.concat(
            (self._subtracted_indices, data[self._index_cols])
        )

    def _add_results(
        self,
        data: pl.DataFrame,
    ) -> None:
        self._results = pl.concat([self._results, data])

    def merge_pipeline(
        self,
        pipeline: PipelineResult,
    ) -> None:
        self._subtract_data(pipeline.subtracted_indices)
        self._add_results(pipeline.results)

    def write_results(
        self,
        data: pl.DataFrame,
        vorlage: bool,
        wf_id: int,
        freigabe_auto: types.Freigabe,
        is_out: bool,
    ) -> None:
        # TODO move to other position
        ORDER_QTY_CRIT: typing.Final[str] = "BEDP_MENGE_BEDARF_VM"

        results = data.rename(db.map_to_result)
        order_qty_expr: pl.Expr
        if is_out:
            order_qty_expr = (
                pl.lit(0)
                .alias("ORDER_QTY_CRIT")
                .alias("best_menge")
                .cast(db.results_schema_map["best_menge"])
            )
        else:
            order_qty_expr = pl.col(ORDER_QTY_CRIT).alias("best_menge")

        results = results.with_columns(
            [
                pl.lit(vorlage).alias("vorlage").cast(db.results_schema_map["vorlage"]),
                pl.lit(wf_id).alias("wf_id").cast(db.results_schema_map["wf_id"]),
                order_qty_expr,
                pl.lit(freigabe_auto.value)
                .alias("freigabe_auto")
                .cast(db.results_schema_map["freigabe_auto"]),
            ]
        )
        results = results.drop(
            [
                "BEDP_TITELNR",
                "BEDP_MAN",
                "BEDP_MENGE_BEDARF_VM",
                "MELDENUMMER",
                "VERLAGSNR",
                "MENGE_VORMERKER",
                "MANDFUEHR",
            ]
        )

        self._subtract_data(data)
        self._add_results(results)


# post-processing the results
# TODO: order quantity not always necessary
# TODO: change relevant criterion for order quantity
# def _write_results(
#     pipe_result: PipelineResult,
#     data: pl.DataFrame,
#     vorlage: bool,
#     wf_id: int,
#     freigabe_auto: types.Freigabe,
#     is_out: bool,
# ) -> PipelineResult:
#     ORDER_QTY_CRIT: typing.Final[str] = "BEDP_MENGE_BEDARF_VM"

#     results = data.rename(db.map_to_result)
#     order_qty_expr: pl.Expr
#     if is_out:
#         order_qty_expr = (
#             pl.lit(0)
#             .alias("ORDER_QTY_CRIT")
#             .alias("best_menge")
#             .cast(db.results_schema_map["best_menge"])
#         )
#     else:
#         order_qty_expr = pl.col(ORDER_QTY_CRIT).alias("best_menge")

#     results = results.with_columns(
#         [
#             pl.lit(vorlage).alias("vorlage").cast(db.results_schema_map["vorlage"]),
#             pl.lit(wf_id).alias("wf_id").cast(db.results_schema_map["wf_id"]),
#             order_qty_expr,
#             pl.lit(freigabe_auto.value)
#             .alias("freigabe_auto")
#             .cast(db.results_schema_map["freigabe_auto"]),
#         ]
#     )
#     results = results.drop(
#         [
#             "BEDP_TITELNR",
#             "BEDP_MAN",
#             "BEDP_MENGE_BEDARF_VM",
#             "MELDENUMMER",
#             "VERLAGSNR",
#             "MENGE_VORMERKER",
#             "MANDFUEHR",
#         ]
#     )

#     pipe_result.subtract_from_open(data)
#     pipe_result.add_results(results)

#     return pipe_result


def workflow_900(
    pipe_result: PipelineResult,
) -> PipelineResult:
    """filter 'Meldenummer' and fill non-feasible entries"""

    filter_meldenummer_null = pl.col("MELDENUMMER").is_not_null()
    filter_mandant = pl.col(MANDANT_CRITERION).is_in((1, 90))
    res = _apply_several_filters(
        pipe_res.open,
        (
            filter_meldenummer_null,
            filter_mandant,
        ),
    )

    pipe_result.write_results(
        data=res.out_,
        vorlage=False,
        wf_id=900,
        freigabe_auto=types.Freigabe.WF_900,
        is_out=True,
    )

    pipe_result.update_open(res.in_.with_columns(pl.col("MENGE_VORMERKER").fill_null(0)))
    pipe_result.update_open(res.in_.with_columns(pl.col("BEDP_MENGE_BEDARF_VM").fill_null(0)))

    return pipe_result


def workflow_910(
    pipe_result: PipelineResult,
) -> PipelineResult:
    # TODO check if necessary because of WF-900
    filter_mandant = pl.col(MANDANT_CRITERION).is_in((1, 90))

    filter_ignore_MNR26 = pl.col("MELDENUMMER") != 26

    res = _apply_several_filters(
        pipe_result.open,
        filters=(
            filter_mandant,
            filter_ignore_MNR26,
        ),
    )
    # write results for entries which were filtered out
    pipe_result.write_results(
        data=res.out_,
        vorlage=False,
        wf_id=910,
        freigabe_auto=types.Freigabe.WF_910,
        is_out=True,
    )

    return pipe_result


# this a main routine:
# receives and gives back result objects
def workflow_100_umbreit(
    pipe_result: PipelineResult,
    vm_criterion: str,
) -> PipelineResult:
    filter_meldenummer = pl.col("MELDENUMMER") == 18
    filter_mandant = pl.col(MANDANT_CRITERION) == 1
    filter_number_vm = pl.col(vm_criterion) > 0

    res = _apply_several_filters(
        pipe_result.open,
        (
            filter_meldenummer,
            filter_mandant,
            filter_number_vm,
        ),
    )
    pipe_result.write_results(
        data=res.in_,
        vorlage=True,
        wf_id=100,
        freigabe_auto=types.Freigabe.WF_100,
        is_out=False,
    )

    return pipe_result


def workflow_100_petersen(
    pipe_result: PipelineResult,
    vm_criterion: str,
) -> PipelineResult:
    # difference WDB and others

    # // WDB branch
    # order quantity 0, no further action in other WFs
    filter_meldenummer = pl.col("MELDENUMMER") == 18
    filter_mandant = pl.col(MANDANT_CRITERION) == 90
    filter_WDB = pl.col("VERLAGSNR").is_in((76008, 76070))
    filter_number_vm = pl.col(vm_criterion) == 0
    res = _apply_several_filters(
        pipe_result.open,
        (
            filter_meldenummer,
            filter_mandant,
            filter_WDB,
            filter_number_vm,
        ),
    )
    pipe_result.write_results(
        data=res.in_,
        vorlage=False,
        wf_id=100,
        freigabe_auto=types.Freigabe.WF_100,
        is_out=False,
    )

    # TODO add check for orders or quantity to transform
    # TODO show them
    filter_meldenummer = pl.col("MELDENUMMER") == 18
    filter_mandant = pl.col(MANDANT_CRITERION) == 90
    filter_WDB = pl.col("VERLAGSNR").is_in((76008, 76070))
    filter_number_vm = pl.col(vm_criterion) > 0
    res_candidates = _apply_several_filters(
        pipe_result.open,
        (
            filter_meldenummer,
            filter_mandant,
            filter_WDB,
            filter_number_vm,
        ),
    )
    wdb_sub_pipe = PipelineResult(res_candidates.in_)
    wdb_sub_pipe = wf100_petersen_wdb_sub1(wdb_sub_pipe)
    assert len(wdb_sub_pipe.open) == 0
    pipe_result.merge_pipeline(wdb_sub_pipe)

    # // other branch
    # show always entries with #VM > 1
    filter_meldenummer = pl.col("MELDENUMMER") == 18
    filter_mandant = pl.col(MANDANT_CRITERION) == 90
    filter_number_vm = pl.col(vm_criterion) > 1
    res = _apply_several_filters(
        pipe_result.open,
        (
            filter_meldenummer,
            filter_mandant,
            filter_number_vm,
        ),
    )
    pipe_result.write_results(
        data=res.in_,
        vorlage=True,
        wf_id=100,
        freigabe_auto=types.Freigabe.WF_100,
        is_out=False,
    )

    filter_meldenummer = pl.col("MELDENUMMER") == 18
    filter_mandant = pl.col(MANDANT_CRITERION) == 90
    filter_number_vm = pl.col(vm_criterion) > 0
    res = _apply_several_filters(
        pipe_result.open,
        (
            filter_meldenummer,
            filter_mandant,
            filter_number_vm,
        ),
    )
    pipe_result.write_results(
        data=res.in_,
        vorlage=True,
        wf_id=100,
        freigabe_auto=types.Freigabe.WF_100,
        is_out=False,
    )

    return pipe_result


def wf100_petersen_wdb_sub1(
    pipe_result: PipelineResult,
) -> PipelineResult:
    # input: pre-filtered entries (WDB titles and #VM > 0)
    # more then 1 VM
    # !! show these entries
    filter_number_vm = pl.col(VM_CRITERION) > 1
    res = _apply_several_filters(
        pipe_result.open,
        (filter_number_vm,),
    )
    pipe_result.write_results(
        data=res.in_,
        vorlage=True,
        wf_id=100,
        freigabe_auto=types.Freigabe.WF_100,
        is_out=False,
    )

    # filtered out entries (WDB with #VM == 1) must be analysed for orders in the
    # past 6 months
    start_date = get_starting_date(180)
    filter_ = sql.and_(
        db.EXT_BESPBES_INFO.c.BESP_TITELNR.in_(res.out_["BEDP_TITELNR"].to_list()),
        db.EXT_BESPBES_INFO.c.BES_DATUM >= start_date,
    )
    stmt = sql.select(db.EXT_BESPBES_INFO).where(filter_)
    df_order = pl.read_database(stmt, engine, schema_overrides=db.EXT_BESPBES_INFO_schema_map)
    entries_show = (
        df_order.group_by("BESP_TITELNR")
        .agg(pl.col("BESP_TITELNR").count().alias("count"))
        .filter(pl.col("count") > 1)
    )
    filter_titleno = pl.col("BEDP_TITELNR").is_in(entries_show["BESP_TITELNR"].to_list())
    res = _apply_several_filters(pipe_result.open, (filter_titleno,))
    pipe_result.write_results(
        data=res.in_,
        vorlage=True,
        wf_id=100,
        freigabe_auto=types.Freigabe.WF_100,
        is_out=False,
    )
    pipe_result.write_results(
        data=pipe_result.open,
        vorlage=False,
        wf_id=100,
        freigabe_auto=types.Freigabe.WF_100,
        is_out=False,
    )

    return pipe_result


# %%
# SAVING/LOADING
p_save = Path.cwd() / "raw_data_from_sql_query_20260114-1.arrow"
df = pl.read_ipc(p_save)
print(f"Number of entries: {len(df)}")

# %%
df.head()


# %%
# removed_rows = []

# raw_data = df.clone()
# print(f"Length raw data: {len(raw_data)}")
# filter_mandant = pl.col("BEDP_MAN").is_in((1, 90))
# filter_ignore_MNR26 = pl.col("MELDENUMMER") != 26


# filtered = raw_data.filter(filter_mandant)
# filtered_n = raw_data.filter(~filter_mandant)
# num_filter = len(filtered)
# num_filter_n = len(filtered_n)
# removed_rows.append(filtered_n)
# print(f"Length filtered: {num_filter}")
# print(f"Length filtered out: {num_filter_n}")
# print(f"Length all: {num_filter + num_filter_n}")
# raw_data = filtered
# out = pl.concat(removed_rows)
# print(f"Length out: {len(out)}")

# # %%
# print("---------------------------------------")
# filtered = raw_data.filter(filter_ignore_MNR26)
# filtered_n = raw_data.filter(~filter_ignore_MNR26)
# num_filter = len(filtered)
# num_filter_n = len(filtered_n)
# len(filtered_n)
# # %%
# removed_rows.append(filtered_n)
# print(f"Length filtered: {num_filter}")
# print(f"Length filtered out: {num_filter_n}")
# print(f"Length all: {num_filter + num_filter_n}")
# out = pl.concat(removed_rows)
# print(f"Length out: {len(out)}")

# %%
raw_data = df.clone()
# pipe_res = get_empty_pipeline_result(raw_data)
pipe_res = PipelineResult(raw_data)
pipe_res.results
pipe_res = workflow_900(pipe_res)
print(f"Length of base data: {len(raw_data):>18}")
print(f"Number of entries pipe data: {len(pipe_res):>10}")
print(f"Number of entries result data: {len(pipe_res.results):>8}")
print(f"Number of entries open data: {len(pipe_res.open):>10}")

# %%
pipe_res.results
# raw_data.filter(pl.col("BEDARFNR") == 166982).filter(pl.col("BEDP_SEQUENZ") == 1)
# %%
# pipe_res.open.filter(pl.col("BEDP_MENGE_BEDARF_VM") > pl.col("MENGE_VORMERKER"))
# %%
pipe_res = workflow_910(pipe_res)
print(f"Length of base data: {len(raw_data):>18}")
print(f"Number of entries pipe data: {len(pipe_res):>10}")
print(f"Number of entries result data: {len(pipe_res.results):>8}")
print(f"Number of entries open data: {len(pipe_res.open):>10}")
# %%
# pipe_res.results.select(pl.col("vorlage").value_counts())
# %%
pipe_res = workflow_100_umbreit(pipe_res, VM_CRITERION)
print(f"Length of base data: {len(raw_data):>18}")
print(f"Number of entries pipe data: {len(pipe_res):>10}")
print(f"Number of entries result data: {len(pipe_res.results):>8}")
print(f"Number of entries open data: {len(pipe_res.open):>10}")
# %%
pipe_res = workflow_100_petersen(pipe_res, VM_CRITERION)
print(f"Length of base data: {len(raw_data):>18}")
print(f"Number of entries pipe data: {len(pipe_res):>10}")
print(f"Number of entries result data: {len(pipe_res.results):>8}")
print(f"Number of entries open data: {len(pipe_res.open):>10}")
# %%
pipe_res.open.filter(pl.col.MELDENUMMER == 18).filter(pl.col.BEDP_MENGE_BEDARF_VM > 0)

# %%
pipe_res.results.select(pl.col("vorlage").value_counts())
# %%
pipe_res.results.filter(pl.col("vorlage") == True)
# %%
raw_data.filter(pl.col("BEDARFNR") == 922160).filter(pl.col("BEDP_SEQUENZ") == 3)
# %%
raw_data.head()

# %%
filt_out


# %%
# ---------------------------------------------------------------------------- #
# Workflow 200 (Umbreit only)
# ---------------------------------------------------------------------------- #
# %%
wf_200_start_data = pipe_res.open.clone()
wf_200_start_data


# %%
def workflow_200_umbreit(
    pipe_result: PipelineResult,
) -> PipelineResult:
    relevant_mnr: tuple[int, ...] = (17, 18)
    filter_meldenummer = pl.col("MELDENUMMER").is_in(relevant_mnr)
    filter_mandant = pl.col("BEDP_MAN") == 1
    # not relevant, because already done in WF-100
    # filter_number_vm = pl.col(vm_criterion) == 0

    res = _apply_several_filters(
        pipe_result.open,
        (filter_meldenummer, filter_mandant),
    )


relevant_mnr: tuple[int, ...] = (17, 18)
filter_meldenummer = pl.col("MELDENUMMER").is_in(relevant_mnr)
filter_mandant = pl.col("BEDP_MAN") == 1
res = _apply_several_filters(
    pipe_res.open,
    (filter_meldenummer, filter_mandant),
)
# %%
# these entries must be checked for relevant orders
# therefore, a temp table must be created in the database to execute efficient
# queries, other approaches are just hacks
# SOLUTION:
# - save these entries to a temp table 'temp'
# - look up the order history of the past 3 months
# -- JOIN ON temp.BEDP_TITELNR = EXT_AUFPAUF.TITELNR
# -- WHERE EXT_AUFPAUF.AUFTRAGS_DATUM > (CURRENT_DATE - 3 months) AND
# -- EXT_AUFPAUF.KUNDE_RECHNUNG NOT IN (608991, 260202) AND
# --
res.in_
# %%
# // demo query with IN statement
data = res.in_.clone()
title_sub_choice = data["BEDP_TITELNR"][:300].to_list()
rel_date = get_starting_date(90)
rel_date

# %%
filter_ = sql.and_(
    db.EXT_AUFPAUF.c.TITELNR.in_(title_sub_choice),
    db.EXT_AUFPAUF.c.AUFTRAGS_DATUM >= rel_date,
    db.EXT_AUFPAUF.c.KUNDE_RECHNUNG.not_in((608991, 260202)),
)
stmt = sql.select(db.EXT_AUFPAUF).where(filter_)

print(stmt.compile(engine))

# %%
demo = pl.read_database(
    stmt,
    engine,
    schema_overrides=db.EXT_AUFPAUF_schema_map,
)
# %%
demo.head()
# %%
demo_2 = demo.clone()
demo_2.head()
print(f"Number of titles before filtering: {len(demo_2)}")
demo_2 = demo_2.filter(pl.col.AUFTRAGS_ART.is_in((1, 99)))
demo_2 = (
    demo_2.group_by("TITELNR", maintain_order=True)
    .agg(
        pl.len().alias("count"),
        pl.col.KUNDE_RECHNUNG.n_unique().alias("customer_count"),
    )
    .filter(pl.col.customer_count >= 3)
)

# these remaining titles are relevant and should be shown
# the others should be disposed
print(f"Number of titles which are relevant: {len(demo_2)}")
print(f"Number of titles which are to be disposed: {len(demo) - len(demo_2)}")
demo_2
# %%
df, filt_out = _init_workflow_200_umbreit(results, wf_200_start_data, VM_CRITERION)
df

# %%
df.filter(pl.col("BEDARFNR") == 884607)

# %%
df_order.filter(pl.col("BEDARFNR") == 884607)

# %%
# now obtain order data for entries
t = df.join(df_order, on=["BEDARFNR", "BEDP_SEQUENZ"], how="inner")
t = t.with_columns(pl.col("AUFP_POSITION").fill_null(0))
t
# %%
agg_t = (
    t.group_by(["BEDARFNR", "BEDP_SEQUENZ"])
    .agg(
        pl.count("AUFP_POSITION").alias("pos_count"),
        pl.col("KUNDE_RECHNUNG").alias("count_customer").n_unique(),
    )
    .filter(pl.col("count_customer") >= 0)  # !! should be 3
)
agg_t

# %%
df_order.filter((pl.col("BEDARFNR") == 883608) & (pl.col("BEDP_SEQUENZ") == 65))

# %%
# ---------------------------------------------------------------------------- #
# Writing results in DB
# ---------------------------------------------------------------------------- #

delete_results()
pipe_post.write_database(db.results.fullname, engine, if_table_exists="append")

stmt = sql.select(db.results)
db_results = pl.read_database(stmt, engine)
db_results

# ---------------------------------------------------------------------------- #
# Further Data Analysis
# ---------------------------------------------------------------------------- #
# %%
stmt = sql.select(db.ext_bedpbed)
df = pl.read_database(
    stmt,
    engine,
    schema_overrides=db.ext_bedpbed_schema_map,
)
# %%
df.group_by("BEDP_TITELNR").agg(
    pl.col("BEDP_MAN").n_unique().alias("unique_BEDP_MAN")
).filter(pl.col("unique_BEDP_MAN") > 1)
# %%
df["BEDP_MAN"].unique()
# %%
df.estimated_size(unit="mb")
# %%
target_bednr = df_raw["BEDARFNR"].to_list()
target_seq = df_raw["BEDP_SEQUENZ"].to_list()
# %%
stmt = (
    sql.select(
        db.ext_bedpbed.c.BEDARFNR,
        db.ext_bedpbed.c.BEDP_SEQUENZ,
        db.ext_bedpbed.c.BEDP_TITELNR,
        db.ext_bedpbed.c.BEDP_MENGE_BEDARF_VM,
    )
    .where(db.ext_bedpbed.c.BEDARFNR.in_(target_bednr))
    .where(db.ext_bedpbed.c.BEDP_SEQUENZ.in_(target_seq))
)
df_targets = pl.read_database(stmt, engine)
# %%
# df_targets.filter(pl.col("BEDARFNR") == 884174)
df_targets.filter(pl.col("BEDP_MENGE_BEDARF_VM") > 0)

# %%
# interesting order: 883697, 1, titleno: 7945981, 9964027
TITLE_NO = 7945981
# TITLE_NO = 9964027
stmt = sql.select(db.EXT_BESPBES_INFO).where(db.EXT_BESPBES_INFO.c.BESP_TITELNR == TITLE_NO)
title_buy = pl.read_database(stmt, engine)
# %%
title_buy

# %% when were the orders placed
stmt = sql.select(db.EXT_AUFPAUF).where(db.EXT_AUFPAUF.c.TITELNR == 7945981)
title_order = pl.read_database(stmt, engine)
# %%
title_order

# -------------------------------------------------------------------------------------------

# %%
# title DB complete?
# - includes only titles which are deliverable since 01.06.2025 and who are assigned to
# buyer "Fröhlich"
stmt = sql.select(db.ext_titel_info)  # .where(db.ext_titel_info.c.TI_NUMMER == 2928800)
titles = pl.read_database(stmt, engine, schema_overrides=db.ext_titel_info_schema_map)
# %%
titles["MANDFUEHR"].unique()
# %%
unique_titles = set(titles["TI_NUMMER"].to_list())
len(unique_titles)

# %%
# requirements?
# - includes only order since 05.11.2025
stmt = sql.select(db.ext_bedpbed)  # .where(db.ext_titel_info.c.TI_NUMMER == 2928800)
reqs = pl.read_database(stmt, engine, schema_overrides=db.ext_bedpbed_schema_map)
# %%
reqs

# %%
reqs["BEDP_MAN"].unique()

# %%
# intersection between all titles and the titles contained in the requirements table
unique_titles_req = set(reqs["BEDP_TITELNR"].to_list())
len(unique_titles_req)
# %%
intersection = unique_titles & unique_titles_req
len(intersection)
# %%
# orders?
# - includes only order since 05.11.2025
stmt = sql.select(db.EXT_AUFPAUF)
orders = pl.read_database(stmt, engine, schema_overrides=db.EXT_AUFPAUF_schema_map)

# %%
orders.estimated_size(unit="mb")

# %%
with engine.connect() as conn:
    res = conn.execute(stmt)
    print(res.all())

# %%
stmt = sql.text("SELECT * FROM EXT_AUFPAUF WHERE AUFTRAGSNUMMER=37847548 and TITELNR=6315273")
with engine.connect() as conn:
    res = conn.execute(stmt)
    print(res.all())

# %%
stmt = sql.text("SELECT * FROM ext_bedpbed WHERE BEDARFNR=859131 and BEDP_SEQUENZ=2")
with engine.connect() as conn:
    res = conn.execute(stmt)
    print(res.all())
# %%
stmt = sql.text("SELECT * FROM EXT_BESPBES_INFO WHERE BESP_TITELNR=6312977")
with engine.connect() as conn:
    res = conn.execute(stmt)
    print(res.all())

# %%
df = dataframes[1]
# %%
col_dtype = {}
for col, dtype in zip(df.columns, df.dtypes):
    col_dtype[col] = dtype

print("dtypes of DF...")
pprint(col_dtype)
# %%
len(df)
# %%
df.filter((pl.col("BEDP_MENGE_BEDARF_VM") != "") & (pl.col("BEDP_MENGE_BEDARF_VM") != "0"))
# %%
stmt = sql.text("SELECT * FROM ext_bedpbed")
df = pl.read_database(stmt, engine)

# %%
df
# %%
# %%
col_dtype = {}
for col, dtype in zip(df.columns, df.dtypes):
    col_dtype[col] = dtype

print("dtypes of DF...")
pprint(col_dtype)
# %%
# ** Petersen WDB
filter_meldenummer = pl.col("MELDENUMMER") == 18
filter_mandant = pl.col(MANDANT_CRITERION) == 90
filter_WDB = pl.col("VERLAGSNR").is_in((76008, 76070))
filter_number_vm = pl.col(VM_CRITERION) > 0

res = _apply_several_filters(
    df,
    (
        filter_meldenummer,
        filter_mandant,
        filter_WDB,
        filter_number_vm,
    ),
)
# %%
res.in_
# %%
# !! show these entries
filter_number_vm = pl.col(VM_CRITERION) > 1

res_vm_crit = _apply_several_filters(
    res.in_,
    (filter_number_vm,),
)
# %%
res_vm_crit.out_
# %%
# filtered out entries (WDB with #VM == 1) must be analysed for orders in the past 6 months
title_nos = res_vm_crit.out_["BEDP_TITELNR"].to_list()
len(title_nos)
# %%
title_nos

# %%
# define starting date for 6 month interval
# returns UTC time
start_date = get_starting_date(180)
filter_ = sql.and_(
    db.EXT_BESPBES_INFO.c.BESP_TITELNR.in_(title_nos),
    db.EXT_BESPBES_INFO.c.BES_DATUM >= start_date,
)
stmt = sql.select(db.EXT_BESPBES_INFO).where(filter_)
df_order = pl.read_database(stmt, engine, schema_overrides=db.EXT_BESPBES_INFO_schema_map)
df_order
# %%
# filter entries which have
df_show = (
    df_order.group_by("BESP_TITELNR")
    .agg(pl.col("BESP_TITELNR").count().alias("count"))
    .filter(pl.col("count") > 1)
)
df_show
# %%
# !! show these entries
# !! do not show others
entries_to_show = df_show["BESP_TITELNR"].to_list()
print(f"Number of entries relevant: {len(entries_to_show)}")
# %%
res_vm_crit.out_
# %%
filter_titleno = pl.col("BEDP_TITELNR").is_in(df_show["BESP_TITELNR"].implode())

res_wdb = _apply_several_filters(res_vm_crit.out_, (filter_titleno,))

# %%
res_wdb.in_
# %%
res_wdb.out_


# %%


# %%
# %%
# %%
schema = {}
for col in ("BEDARFNR", "BEDP_SEQUENZ"):
    schema[col] = db.raw_data_query_schema_map[col]
base = pl.DataFrame(schema=schema)
# %%
data = {"BEDARFNR": list(range(10)), "BEDP_SEQUENZ": list(range(10))}
orig_data = pl.DataFrame(data, schema=schema)
data = orig_data[:5].clone()
# %%
pl.concat([base, data])
# %%
orig_data.join(data, on=["BEDARFNR", "BEDP_SEQUENZ"], how="anti")
# %%
orig_data[("BEDARFNR", "BEDP_SEQUENZ")]
# %%
raw_data = df.clone()
pipe_res = PipelineResult(raw_data)
pipe_res.open
# %%
pipe_res.results
# %%
sub_data = pipe_res.open[:20].clone()
sub_data
# %%
pipe_res.write_results(
    sub_data,
    vorlage=True,
    wf_id=30,
    freigabe_auto=types.Freigabe.WF_100,
    is_out=True,
)
# %%
pipe_res.open
# %%
pipe_res.results
# %%
raw_data = df.clone()
pipe_res_main = PipelineResult(raw_data)
pipe_res_main.open
# %%
pipe_res_main.merge_pipeline(pipe_res)
# %%
pipe_res_main.open
# %%
pipe_res.results
# %%