Bit late probably, but have you tried out my workaround (as discussed here)? It might do the trick fetching the right macro value, but you will have to export/link your custom transformer as an FMX.

In your case, you could try adding the following function at the "global" level of your Python code:

def get_macro_value(param_name, default=None): context = FME_MacroValues.get('WB_CURRENT_CONTEXT') for key, value in FME_MacroValues.items(): if key.endswith(param_name): if not context or (context and key.startswith(context)): return value return default

Then, you can replace the first 4 lines in your code above with:

parameter_name = 'codeLength' mv_codeLength = get_macro_value(parameter_name)

Specifying the custom transformer name is irrelevant here, since the function fetches that from the WB_CURRENT_CONTEXT macro value, which always exists within a custom transformer and returns the current name of the transformer. This is why it only works for linked custom transformers: WB_CURRENT_CONTEXT is empty for embedded transformers.

Let me know if that works out for you!

Btw, I would replace lines 6-13 in your code with (to make it a little more robust):

codeLength = 0 try: if mv_codeLength.startswith(('@Value', '@Evaluate')): codeLength = int(feature.performFunction(mv_codeLength)) elif mv_codeLength.isdigit(): codeLength = int(mv_codeLength) except (TypeError, AttributeError, ValueError) as e: # Can happen when mv_codeLength is not a string or # when the feature function output can not be cast to an integer. # You could log this error or do something else here. pass

Finally, there actually is a way to fetch the name of the current custom transformer name from within (even if it is embedded), but you will have to use a PythonCaller with a FeatureProcessor class. When instantiated, a property called self.factory_name is set, which starts with the name of the custom transformer, followed by the name of the PythonCaller (and concatenated by an underscore)!

This is a known issue / challenge / problem (depending on how you look at it). Your workaround using e.g. an AttributeCreator + AttributeRemover is currently the simplest and most robust solution, so that's what I'd recommend.