gusutabopb/pytrthree

Wrapper and utils for the Thomson Reuters Tick History API

pytrthree's Issues

Validate input to TRTHIterator

1. the code throws when files is actually a file.
   self.files = files should be moved up front like below:

    def __init__(self, files, chunksize=10 ** 6):
      ->self.files = files
        if isinstance(files, str):
            self.files = [files]
        if not isinstance(files, Sequence):
            raise ValueError('Invalid input')
        self.chunksize = chunksize

2. It would be very nice if make_next(self) skips a file name that has confirmation or report in it.
   Or, at least it informs users of a file other than result ones are getting fed and throwing.

    def make_next(self):
        for file in self.files:
            if re.search('confirmation|report', file):
                raise
            else:
                chunks = pd.read_csv(file, iterator=True, chunksize=self.chunksize)
                for chunk in chunks:
                    for _, df in chunk.groupby('#RIC'):
                        yield self.pre_process(df.copy())

Retrying submission in request_sender

It would be nice if we can put a large number of requests at once.
If we reach max number of simultaneous requests for reuters server, then the tool creates an event listener (waiting for job complete notification) to kick a callback which in turn sends another request via api.