Issue with case sensitivity in enums

[robbertvanwaveren]

Describe the bug
python Fields that represent enums values are translated directly to UPPERCASE regardless of their original casing.
This results in duplicate fields if original values from the schema only differ in casing.

I encounted this is issue in our from standards derived avro models that use SI units that have quite some cases where this is an issue like M(ega) vs m(illi), M(onth) vs m(inute) etc.

{
  "name": "UnitMultiplier",
  "type": "enum",
  "symbols": [
    "p",
    "P",
     ....
  ]
}

results in

class UnitMultiplier(enum.Enum):
    P = "p"
    P = "P"
    ... 

Perhaps rename the non-original uppercase P field to 'P_' or something when duplicates are found.

To Reproduce
See above

Expected behavior
Support for case-sensitive uniqueness such that any valid avro schema results in a valid generated python model.

[marcosschroh]

Hi @robbertvanwaveren

It makes sense. I think we should use the values that are in the symbols directly without any transformation. Currently what is happening is that we are doing an uppercase of the key

[marcosschroh]

Looking at the enum module documentation the keys are always in uppercase (which makes sense) so we should keep doing the same. If we go with your solution of adding _ we can have the situations like:

{
  "name": "UnitMultiplier",
  "type": "enum",
  "symbols": [
    "p",
    "P",
    "P_",
     ....
  ]
}

and the result should be:

class UnitMultiplier(enum.Enum):
    P = "p"
    P_ = "P"
    P__ = "P_"
    ... 

which is confusing. Ideally symbols should be proper defined and descriptive but sometimes we can not do anything.
I will think another strategy.

[marcosschroh]

After taking a look the only way to solve this is using the symbol as it is in the schema when it is repeated. Following your example the generated enum will be:

class UnitMultiplier(enum.Enum):
    P = "P"
    p = "p"
    ... 

For future references to other uses: Avoid this because it is hard to understand and the enum won't generate the proper uppercase

The documentation: https://marcosschroh.github.io/dataclasses-avroschema/pr-preview/pr-402/model_generator/#enums-and-case-sensitivity

[robbertvanwaveren]

I'm afraid your fix only works if the uppercase symbol comes before the lowercase.

[marcosschroh]

@robbertvanwaveren The next release is fixing it! Thanks for reporting it!