You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Based on #472, I noticed that the General Info class could expose a few more attributes that are useful. Here are a couple I was surprised to see weren't provided:
dtype: Weird this isn't explicitly available as complex_data is provided and, from reading the code in base/base.py, it appears that dtype is inferred via np.complex64 if self.complex_data else np.float32
sample_shape: Since shape is provided, shouldn't sample_shape be provided as well?
The text was updated successfully, but these errors were encountered:
theXYZT
changed the title
Dtype in Info object
Dtype in Info object (and other useful attributes to have?)
Nov 28, 2020
Going over old issues: yes, this makes sense. I thought initially that info should just provide a minimal set, but really dtype is useful. And easy to do, just add some parent attributes to base.file_info.StreamReaderInfo...
p.s. I'd also like to add some more control over the exact dtype. Perhaps float16 for some cases, or even allow integer representation to be used. For incoherent dedispersion plus gating and then writing out, one does not need the transformation to float.
Based on #472, I noticed that the General Info class could expose a few more attributes that are useful. Here are a couple I was surprised to see weren't provided:
dtype
: Weird this isn't explicitly available ascomplex_data
is provided and, from reading the code in base/base.py, it appears thatdtype
is inferred vianp.complex64 if self.complex_data else np.float32
sample_shape
: Sinceshape
is provided, shouldn'tsample_shape
be provided as well?The text was updated successfully, but these errors were encountered: