"""Analyse la structure du classeur de masse salariale sans afficher d'identité ni de montant."""
from collections import Counter
from pathlib import Path
import re

from openpyxl import load_workbook


workbook_path = Path("/home/ubuntu/upload/TableaurecapMassesalarialeSANTINOVA-28.08.26.xlsx")
workbook = load_workbook(workbook_path, data_only=False)
values_workbook = load_workbook(workbook_path, data_only=True)

for sheet in workbook.worksheets:
    print(f"sheet={sheet.title}|rows={sheet.max_row}|cols={sheet.max_column}")
    for row_index in range(1, min(sheet.max_row, 8) + 1):
        values = []
        for column_index in range(1, sheet.max_column + 1):
            value = sheet.cell(row_index, column_index).value
            if isinstance(value, str) and value.startswith("="):
                value = f"FORMULA:{value}"
            if value is not None:
                values.append(f"C{column_index}={value}")
        if values:
            print(f"row={row_index}|{' ; '.join(values)}")
    formula_count = sum(
        1
        for row in sheet.iter_rows()
        for cell in row
        if isinstance(cell.value, str) and cell.value.startswith("=")
    )
    print(f"formula_count={formula_count}")

salary_sheet = workbook["Masse salariale"]
data_rows = [row for row in range(6, salary_sheet.max_row + 1) if salary_sheet.cell(row, 2).value]
print(f"salaries_with_identifier={len(data_rows)}")
for column, label in {
    8: "2025_brut_avec_primes",
    9: "2025_astreintes",
    11: "2025_taux_charges",
    12: "2025_evolution",
    15: "2026_astreintes",
    17: "2026_taux_charges",
    18: "2026_evolution",
    21: "2027_astreintes",
    23: "2027_evolution",
    24: "2027_prime",
}.items():
    populated = sum(salary_sheet.cell(row, column).value is not None for row in data_rows)
    formulas = sum(
        isinstance(salary_sheet.cell(row, column).value, str)
        and salary_sheet.cell(row, column).value.startswith("=")
        for row in data_rows
    )
    print(f"field={label}|populated={populated}|formulas={formulas}")

for column, label in {7: "2025_hors_astreinte", 14: "2026_avec_primes", 20: "2027_avec_primes"}.items():
    formulas = Counter()
    for row in data_rows:
        value = salary_sheet.cell(row, column).value
        if isinstance(value, str) and value.startswith("="):
            formulas[re.sub(r"\d+", "#", value)] += 1
    for formula, count in formulas.items():
        print(f"formula_pattern={label}|count={count}|formula={formula}")

values_sheet = values_workbook["Masse salariale"]
for column, label in {7: "2025_hors_astreinte", 10: "2025_mensuel", 13: "2026_hors_astreinte", 14: "2026_avec_primes", 16: "2026_mensuel", 19: "2027_hors_astreinte", 20: "2027_avec_primes", 22: "2027_mensuel"}.items():
    calculated = sum(values_sheet.cell(row, column).value is not None for row in data_rows)
    print(f"cached_value={label}|populated={calculated}")