When creating a geometamaker yaml file for a geotiff, frictionless comes back with an encoding attribute inferred from chardet. The thing is, sometimes chardet returns a surprising result when we pass it a binary file like a GeoTiff. In a recent example, the encoding attribute provided by geometamaker was cp1254, which is the Turkish codepage on Windows!
The thing that's a bit confusing to me about this attribute is that the encoding in this context refers to how text is encoded ... which doesn't really make sense for binary files like GeoTiffs. For text files analyzed by geometamaker, the encoding definitely does make sense.
For binary files, maybe the existing format tag is sufficient? Or maybe this is a use case for MIMEtypes or something else that describes the binary format?
When creating a geometamaker yaml file for a geotiff,
frictionlesscomes back with anencodingattribute inferred fromchardet. The thing is, sometimeschardetreturns a surprising result when we pass it a binary file like a GeoTiff. In a recent example, the encoding attribute provided by geometamaker wascp1254, which is the Turkish codepage on Windows!The thing that's a bit confusing to me about this attribute is that the encoding in this context refers to how text is encoded ... which doesn't really make sense for binary files like GeoTiffs. For text files analyzed by geometamaker, the encoding definitely does make sense.
For binary files, maybe the existing
formattag is sufficient? Or maybe this is a use case for MIMEtypes or something else that describes the binary format?