Field names get capitalized implicitly..

[goodboy]

I'm only writing this up so it hopefully doesn't bite any other user 😺.

This shamefully took me a while to figure out because of 2 problems:

1. the pymarketstore.Client.write() method doesn't raise errors returned from the server by default which meant I didn't see the column mismatch error returned from the server (at least errors of a certain class).

2. The server actually capitalizes all element names (numpy.dtype field names from the pymarketstore.Client perspective) when loading timebucket info. So, if you write non-capitalized field names in your numpy` structured array field names, you'll get an error when you try to append to that time bucket. So for example the following client code raises an error:

import pymarketstore as pymkts
client = pymkts.Client('http://localhost:5993/rpc')
symbol = 'SPY'

# Constants (notice not capitalized!)
DATA_TYPE_CANDLE = [
    ('Epoch', 'i8'), ('open', 'f8'), 
    ('high', 'f8'), ('low', 'f8'), 
    ('close', 'f8'), ('volume', 'f8')
]

data = [(timestamp('2019-01-01 00:00:00'), 1.0, 2.0, 3.0, 4.0, 5.0 ),  
          (timestamp('2019-01-01 00:00:01'), 1.0, 2.0, 3.0, 4.0, 5.0 ),
          (timestamp('2019-12-31 23:59:59'), 6.0, 7.0, 8.0, 9.0, 10.0)]

# works
client.write(np.array(data, dtype=DATA_TYPE_CANDLE), "{}/1Sec/OHLCV".format(symbol)))

# errors in the response as per 1. above.
client.write(np.array(data, dtype=DATA_TYPE_CANDLE), "{}/1Sec/OHLCV".format(symbol)))

I'll probably make an issue on pymarketstore to at least document that structarray fields must be caps.