Closed pl0xz0rz closed 1 month ago
Thanks. One test failed:
============================= FAILED test_ClientSideJoin.py::test_join - TypeError: unsupported operand type(s) for +: 'NoneType' and 'str' ========================================================================================================== 1 failed, 43 passed, 283 warnings in 86.91s (0:01:26)
We can make diff in respect to the varNorm - this is done per variable and we can compare different statistics with different weights
Combinations:
Old templated had only normVar - [Yes, No]
similar as above but with colorZ histXYWeight 0,0 - mean weight 0,1 - rms for weght 0,2 -entries for wight 1,0 - mean, median for reference wight 1,1. - -entries for reference wight 1,2 - -entries for reference wight 2,0 - diff mean, median weight. (diff - minus or ratio) 2,1 - diff rms for weight (diff - minus or ratio) 2,2 -diff entries for wight (diff - minus or ratio)
Thanks for the commit! The unit tests passed successfully.
Here are some comments on the implementation based on the screenshot:
histoXYweight
Entries Column: I believe it would be more appropriate to use a 2D representation for the entries column, specifically "entries:X". Currently, it's displayed in 3D as "entries vs Y vs color(X)". My preference is to maintain the same X-axis for both visualizations.There are many warnings - related to the unit test
After adding pytest.ini with unit test the amount was reduced to the 53
I will commit the pytest.ini after merging
(venv7) Singularity> cat $RootInteractive/pytest.ini
[pytest]
markers =
unittest: description for the unittest marker
I briefly reviewed the code and did not identify any obvious issues.
We will need to develop a new unit test in which we prepare specialized data and define invariants to facilitate easier testing of self-consistency.
I am merging the pull request, but the issue remains open for now.
This PR: -Fixes a major bug with generated javascript for using columns from different histograms/dataframes in one expression -Fixes a bug with counting the length in joins on the client -Adds a prototype implementation of #352