Data hacks: clean puiss_max and code_insee

[AntoineAugusti]

A common error made by producers is badly formatted data. There are already a handful of hacks in place to clean the data. After looking at a failed validation report, I fixed more common mistakes:

• puiss_max: expected format is 3.7. People often input 3,7 or 3,7kW
• code_insee: expect format is 76100. People often add a space after the department (why?) or miss a space at the beginning if the department code is a single digit.

Disclaimer: I didn't run the script locally, you may want to do this after reviewing the changes.

[abulte]

The script crashes ;-) Those columns are not always defined, cf what I've done for lat long:

    unique['Xlongitude'] = unique['Xlongitude'].replace(',', '.') if unique['Xlongitude'] else ''

[AntoineAugusti]

Oh so, these conditions matter! Right, updating now.

[abulte]