Fix encoding issues if possible

[SvenLieber]

Some input files are encoded as UTF-8, but include wrongly encoded data that have to be interpreted correctly.
We internally faced this issue already in the past, we should apply the same fix here too.

Working solution from the geoname_enrichment script:

https://github.com/MetaBelgica/geoname-enrichment/blob/d94851e9ae81e06a45ee2f53d4b8e952fe0182f7/geoname_enrichment/utils.py#L2-L23

[SvenLieber]

The current implementation has insufficient validity checks and throws errors:

    encoded = text.encode('latin1')
AttributeError: 'NoneType' object has no attribute 'encode'