Feature/monadic api tests 2

[amiri]

I am having some strangeness with this test. I am testing the sampleM function, seeing whether the expected sample rate is the real sample rate, but there are 2 problems.

1. Sometimes I get a non-number from D.nRows:

Rows in initialDf: 705)

It looks like I may be getting generated arbitrary data in there, but GenDataFrame:genDataFrame says that numRows should always be an Int.

2. And sometimes I get no rows sampled, which doesn't make sense to me:

Rows in initialDf: 204d)
Rows in finalDf: 0; expectedRate is 0.0 where 0.22 was expected
Diff is 0.22

This is always the last test, which always fails. I am doing something wrong, but I don't know what.

[mchav]

Investigating trace.

There's this message:

-- | The 'traceIO' function outputs the trace message from the IO monad.
-- This sequences the output with respect to other IO actions.
--
-- @since base-4.5.0.0

Maybe the right approach is getting out the I/O monad.

[mchav]

Seems like a weird interaction from generating zero columns dataframes:

Change GenDataFrame to:

genDataFrame :: Gen DataFrame
genDataFrame = do
    numRows <- choose (100, 1000)
    numCols <- choose (1, 10)
    colNames <- V.fromList <$> vectorOf numCols genUniqueColName
    cols <- V.fromList <$> vectorOf numCols (genColumn numRows)
    let indices = M.fromList $ zip (V.toList colNames) [0 ..]
    pure $
        DataFrame
            { columns = cols
            , columnIndices = indices
            , dataframeDimensions = (numRows, numCols)
            , derivingExpressions = M.empty
            }

[amiri]

Doh! I suppose my pre condition didn't work right... Still working on this.

[mchav]

The precondition checks rows not columns. So might be worth checking both. Even with a passing test the trace output is still weird.