CONTEXT

My project is about implementing the “conventional method for house-energy-efficiency assessment” in France (120 pages).

This method, comes with a huge .xls file providing definition for 10’s of enum-like data structure (see sample table below)

I want to model this data as a Python Enum with multiple attributes. Each table always has an id column (int|str) and a label column (str). Possibly several other columns are present as in the sample table (see below the hors_methode column).

I already know how to achieve that using a modified aenum.Enum :

from aenum import Enum

class BaseEnumWithProps(Enum):

    def __str__(self):
        return self.label
    
    @classmethod
    def _missing_value_(cls, value):
        """
        Permet d'instancier une Enum à partir de son label (str) ou de son id (int)
        """
        if isinstance(value, int):
            for member in cls:
                if member.id == value:
                    return member
        if isinstance(value, str):
            for member in cls:
                if member.label == value:
                    return member 
    

class TypePlancherBas(BaseEnumWithProps):
  
  _init_ = "id, label, hors_methode"
  
  ID_1 = 1, "Inconnu", 0
  """Inconnu"""
  ID_2 = 2, "Plancher avec ou sans remplissage", 0
  """Plancher avec ou sans remplissage"""
  # ...
  ID_13 = 13, "Autre type de plancher non répertorié", 1
  """Autre type de plancher non répertorié"""

pb_1 = TypePlancherBas.ID_1
pb_2 = TypePlancherBas.ID_2

print(pb_1.label)
print(pb_1.hors_methode)

The Problem

The problem with this way of doing things is because TypePlancherBas is dynamically created (through metaclass), there is no way to get IntelliSense working on Enums. vscose will never tell me that label or id is an attribute of TypePlancherBas.ID_XX. This is bad for code discovery and usability of the API I am building (remember there are 10’s of such Enums …).

The Question

I am thinking about creating all the Enums through code generation : a python script (run once for ever) coud process the .xls file and build a enums.py module gathering all the Enum’s (through templating with jinja2).

What would be the proper class structure for such an Enum (see sample below) to make the code discoverable by IntelliSense ?

I’ve tried countless options but could not reach my goal …

Related posts :
/a/59417442/7060811
Python Enum: How to get enum values with multiple attributes

Sample Table : type_plancher_bas