Test for expected data rectangles and file contents

[wibeasley]

As part of the test suite (#4), the retrieved data & files should be compared to their expected values/file/data. Some of the options are:

• the original data file (compared as text) --this is the most portable across client libraries, like pyDataverse and dataverse-client-python. See Steal tests from pyDataverse #29
• the tsv from the ingested tab (maybe compared as text)
• the data.frame converted from the ingested tab (maybe compared as an rds)

[kuriwaki]

What's the difference between the second and third items? Once a tabular datafile is read into R, it's already a dataframe.

If we want to go for extensive tests, I think we want tests for:

• ingested datafiles read in in their ingested (tab plain text) form
• ingested datafiles read in their original form (which may be csv, dta, sav, xlsx - see supported formats).
• non -ingested datafiles (e.g. a rds file nlsw_rds-export.rds) read in their original form

[kuriwaki]

The above commits give decent tests for rds and rda.