DEV: _get_images can take header or list of headers as input #61
Conversation
Codecov Report
@@ Coverage Diff @@
## master #61 +/- ##
==========================================
- Coverage 28.92% 28.72% -0.21%
==========================================
Files 10 10
Lines 280 282 +2
==========================================
Hits 81 81
- Misses 199 201 +2
Continue to review full report at Codecov.
|
| @@ -178,7 +178,10 @@ def get_images_to_3D(images, dtype=None): | |||
|
|
|||
| def _get_images(header, tag, roi=None): | |||
| t = ttime.time() | |||
| images = header.db.get_images(header, tag) | |||
| if isinstance(header, (list, tuple)): | |||
There was a problem hiding this comment.
Can you change this to use header.data instead?
There was a problem hiding this comment.
thanks for suggestion on new api. updated
csxtools/utils.py
Outdated
| @@ -178,7 +178,10 @@ def get_images_to_3D(images, dtype=None): | |||
|
|
|||
| def _get_images(header, tag, roi=None): | |||
| t = ttime.time() | |||
| images = header.db.get_images(header, tag) | |||
| if isinstance(header, (list, tuple)): | |||
| images = header[0].data(tag) | |||
There was a problem hiding this comment.
Are we only interested in the first image here?
There was a problem hiding this comment.
@mrakitin is correct -- we don't want to throw away headers in the case of a list input.
We want to be able to use two (or more generally an arbitrary list of) gain 8 dark images to correct the data and this functionality was broken in the last update.
from databroker import DataBroker as db
from csxtools.utils import get_fastccd_images
old_scan_no = 85528
slicerator = get_fastccd_images(db[old_scan_no], (db[old_scan_no-1, old_scan_no+1], None, None))
i.e. 85527 & 85529 are dark scan_ids with gain 8.
There was a problem hiding this comment.
Maybe we can do something like this:
images = []
if isinstance(header, (list, tuple)):
for i in range(len(header)):
images.append(header[i].data(tag))I'm not sure if there needs to be a corresponding update downstream.
There was a problem hiding this comment.
(h.data(tag) for h in header) is not working. It fails at get_images_to_3D function. We need a pims object.
|
@mpmdean What is the expected signature of this function? |
|
Below is the error from the latest CSX kernal on srv1. It's trying to access the _get_images method on the list of two headers. This works in older kernals. AttributeError Traceback (most recent call last) /opt/conda_envs/analysis-2018-1.0/lib/python3.6/site-packages/csxtools/utils.py in get_fastccd_images(light_header, dark_headers, flat, gain, tag, roi) /opt/conda_envs/analysis-2018-1.0/lib/python3.6/site-packages/csxtools/utils.py in _get_images(header, tag, roi) AttributeError: 'list' object has no attribute 'db' I believe this comes from this change: -def _get_images(header, tag, roi=None, handler_override=None): images = get_images(header, tag, handler_override=handler_override) |
|
I think this should work. I also did some local test. any more comment? |
|
Hi All, |
|
I am not sure what happened, but now this functionality seems to work. What is very scary is that the version of csxtools remains the same so I don't really understand why I had the issue in the first place. I tried all 3 23id1 servers before and found this issue. But just this week, I forgot and ran other notebooks and it seemed to be fixed. I am going to close the issue on csxtools |
This is to be compatible with previous api that _get_images can take list of headers as input.