BBIRead.values() returning all nans when missing is not default

[mtvector]

Hi there,
Using the pip install of pybigtools-0.2.3, I am encountering the following strange behavior:

with pybigtools.open('/allen/programs/celltypes/workgroups/rnaseqanalysis/EvoGen/Team/Matthew/data/cortex_data/internal_mouse/atac/subclass_bigwig/test/L5_ET.bw') as bw:
    x = bw.values("chr1", 140254337, 140256551)
print(x)

array([0.00977104, 0.00977104, 0.00977104, ..., 0. , 0. , 0.00977104])

with pybigtools.open('/allen/programs/celltypes/workgroups/rnaseqanalysis/EvoGen/Team/Matthew/data/cortex_data/internal_mouse/atac/subclass_bigwig/test/L5_ET.bw') as bw:
    x = bw.values("chr1", 140254337, 140256551,missing=np.nan,)
print(x)

array([nan, nan, nan, ..., nan, nan, nan])

Maybe I'm not understanding something, but it seems like a bug as these values shouldn't be missing! oob seems to behave as expected.

Thanks for looking into this

Matthew