Tuesday, 14th March, 15:00 UTC
SB, AB, HB, HG, MK, PM, PJ, RO, RB, WdN
Agreed period: end of June, 2023
- General requirements for changing definitions:
- modification in Base Classes or Application Definitions which affect reading and writing files: vote is needed
- fixing documentation or bugs: PR and review is needed
anyone can issue a PR, but it is suggested to
work in a small groups (with domain experts)
- AB: Topic suggested by Dectris - PR #1246
NX_CHAR_OR_NUMBER to be used for NXdata/AXISNAME
NX_CHAR_OR_NUMBER NX_CHAR, NX_NUMBER
- MK: make labels separate from NUMBER typed AXISNAME,
e.g. AXISNAME@index_labels(NX_CHAR): array of NX_CHAR labels
- RO: one shall demonstrate how to use it (e.g matplotlib)
- SB: nxpy, H5web? - others plotting applications
- AB/PJ: going with MK we do not break the standard
- Conclusion: AB will be taking it back to Dectris for further discussion, to see if indirection is acceptable.
is it generic enough?
discussion deferred until next time.
- WdN #1212
WdN: presented ambiguities around the interpretation of NXdata/@AXISNAME_indices
- Use Case 1:
axes=[[xp,xq],y,[z,time]] referreing to a 3D dataset
where the 1st axis can be labeled by both xp, and xq, the second by y, and the 3rd by both z and time.
- Use Case 2:
x_indices=0,1 (x is a 2D array to specify x for each column along z)
y_indices=0,1 (y is a 2D array…)
from this one can do a 2D plot for each slice along z where the z values at (i,j) are ploted at the position (x(i,j),y(i,j)).
- Note that documentation says that the size of axes “must must” be the same as the rank of the signal array, which must be a bug.
- MK: maybe axes could be an array of lists (as demonstrated above) to make it clear and then we do not even need AXISNAME_indices
- RO: these are around for a while, but is there anyone who are using them?
- AB: proposal for Wout to take the discussion and make a PR that is what he thinks is the best approach. We’ll use that as a discussion point for next time.
- PM #412
Documentation for NXdetector should be cleaned and updated
Please help to choose the date by responding to the poll by March 30. We are planning to hold the telco in the regular slot of UTC 15:00. Check your local time to avoid scheduling surprises!