Guilherme Castelão

@OlyDLG raised some good points at #12, and we should extract and save corrupted lines, like blank lines and incomplete lines. This is different than #32 , since with the...

load_data, empty line

4

The load_data was not able to handle an empty line on the raw_data. A temporary solution is to remove the empty line with re.sub()

I found a new case with minutes with only one digit. Build tests for all these formats. Check the TSG_PIR_001.cnv

Maybe use nose to apply the test on the parsed data.

Check my comment on the cnv.yaml itself

set bad_flag to a very small value, like 1e-30, and check if seabird can read a valid 0.

In the rules, it uses only . as separator for the decimals of the minutes. Might be a good idea to allow .|\, but in that case I would need...

The load_data() loads the raw data into a masked array. It uses the bad_flag to identify the non valid data to be masked. The default tolerance for the masked_value is...

The CNV class should have an entry for default attributes. For each key, if don't exists in self.attributes, include it. It would make easier to read several stations from the...

The `bias_calc` was already getting too large, so let's split and reorganize it. This PR is meant to simplify the review of the following `PresRat`. Here, it is mostly about...