JP.fetch_price is failing due to a UnicodeDecodeError in pandas.read_csv, the CSV file contains Japanese characters. I don't get this issue locally and saw it for the first time on the Kibana logs.
Maybe it's as simple as passing an encoding parameter to pandas.read_csv. But I'm using the same version of pandas that's specified in the requirements.txt file, so maybe that's not it. I can't replicate the issue locally. Any ideas?
Hmm maybe. @maxbellec any ideas?
I was able to replicate the UnicodeDecodeError on my other Debian installation. Giving read_csv encoding='shift-jis' seemed to have fixed it. Sorry, I can't debug the locale/encoding settings more right now.
Most helpful comment
I was able to replicate the UnicodeDecodeError on my other Debian installation. Giving
read_csvencoding='shift-jis'seemed to have fixed it. Sorry, I can't debug the locale/encoding settings more right now.