Closed nsmith- closed 2 years ago
nsmith-
commented
2 years ago If we want to recover the old behavior of converting all --- tokens to 0 we could do
f = ["0" if tok == "-" * len(tok) else tok for tok in l.split()]or similar. But as far as I know, the only place we want this behavior is on nuisance effects
amarini
commented
2 years ago Could it be because the regexp were not compiled?
nsmith-
commented
2 years ago No, I tried pre-compiling them at first and it did not help much
hcombbot
commented
2 years ago Running options:
| comb_2019_hbb_boosted_standalone | comb_2019_hgg | comb_2019_hmm | comb_2019_htt | comb_2019_hww | comb_2019_tth_hbb | comb_2019_tth_hgg | comb_2019_tth_multilepton | comb_2019_vh_htt | comb_2019_vhbb | comb_2019_vhbb2017 |
|---|---|---|---|---|---|---|---|---|---|---|
| 1.00 | 0.96 | 0.99 | 1.30 | 1.00 | 1.00 | 0.67 | 1.00 | 1.01 | 1.00 | 1.00 |
You can find more detail at https://gitlab.cern.ch/cms-hcg/performances/ci/-/pipelines/4401589
hcombbot
commented
2 years ago Running options:
| comb_2019_hbb_boosted_standalone | comb_2019_hgg | comb_2019_hmm | comb_2019_htt | comb_2019_hww | comb_2019_tth_hbb | comb_2019_tth_hgg | comb_2019_tth_multilepton | comb_2019_vh_htt | comb_2019_vhbb | comb_2019_vhbb2017 |
|---|---|---|---|---|---|---|---|---|---|---|
| 1.00 | 0.96 | 0.99 | 1.30 | 1.00 | 1.00 | 0.67 | 1.00 | 1.01 | 1.00 | 1.00 |
You can find more detail at https://gitlab.cern.ch/cms-hcg/performances/ci/-/pipelines/4401595
hcombbot
commented
2 years ago Running options:
| comb_2019_hbb_boosted_standalone | comb_2019_hgg | comb_2019_hmm | comb_2019_htt | comb_2019_hww | comb_2019_tth_hbb | comb_2019_tth_hgg | comb_2019_tth_multilepton | comb_2019_vh_htt | comb_2019_vhbb | comb_2019_vhbb2017 |
|---|---|---|---|---|---|---|---|---|---|---|
| 1.00 | 0.96 | 0.99 | 1.00 | 1.00 | 1.00 | 0.67 | 1.00 | 1.01 | 1.00 | 1.00 |
You can find more detail at https://gitlab.cern.ch/cms-hcg/performances/ci/-/pipelines/4435886
hcombbot
commented
2 years ago Running options:
| comb_2019_hbb_boosted_standalone | comb_2019_hgg | comb_2019_hmm | comb_2019_htt | comb_2019_hww | comb_2019_tth_hbb | comb_2019_tth_hgg | comb_2019_tth_multilepton | comb_2019_vh_htt | comb_2019_vhbb | comb_2019_vhbb2017 |
|---|---|---|---|---|---|---|---|---|---|---|
| 1.00 | 0.96 | 0.99 | 1.00 | 1.00 | 1.00 | 0.67 | 1.00 | 1.01 | 1.00 | 1.00 |
You can find more detail at https://gitlab.cern.ch/cms-hcg/performances/ci/-/pipelines/4435889
Two regular expressions were applied to all lines in card, both of which were 10x slower than doing the equivalent transformation by hand in python.
\\s*#.*removes comments and whitespace at the end of the line(?<=\\s)-+(\\s|$)converts any-entries to0The latter was applied universally, whereas now it is only checked for parameter effect size arguments. This means that the interpretation of arguments for certain nuisances is now more restricted, e.g.:
used to be equivalent to
but now would raise an error on converting
-to int or float. (of course forgmNyou wouldn't really want 0 anyway) Using---to indicate no effect on the process column still works the same as before.