The files published in recent days have been successfully converted using the file check in the bash script, but the file of the 27th of April was detected as utf-8 and still had the wrong encoding that crashed pandas.read_csv.
Fixed in 0bf8552. The problem is a mix of encoding so iconv is not the tool that can fix the files. We rely on a perl script that correctly converts all of latin-1 into UTF-8.
The files published in recent days have been successfully converted using the
file
check in the bash script, but the file of the 27th of April was detected as utf-8 and still had the wrong encoding that crashedpandas.read_csv
.