fix: results containing \u0000 break run_query_dataframe

[gosuto-inzasheru]

pandas' read_csv does not properly read the string if it contains these unicodes and the string is discarded instead

it either should be stripped on dune side or maybe there is a cleaner way than my current hacky workaround:

raw = dune.run_query_csv(query).data
with open('dirty.tmp', 'wb') as f:
    f.write(raw.getbuffer())
with open('clean.tmp', 'w') as f:
    buffer = open('dirty.tmp').read()
    for sanitise in ['\x00', '\x1a', '\x1b', '\x1c', '\x1d', '\x1e', '\x1f']:
        buffer = buffer.replace(sanitise, '')
    f.write(buffer)
df = pd.read_csv('clean.tmp')

for clarity, the data point in question is a result of

FROM_UTF8(VARBINARY_LTRIM (data))