Added filter cuts on top of new Main#95
Conversation
comane
left a comment
comane
left a comment
There was a problem hiding this comment.
Hi @FrancescoMerlotti, couple of comments
| from validphys.filters import AddedFilterRule | ||
| return tuple(AddedFilterRule(**rule) for rule in rules) if rules else None |
There was a problem hiding this comment.
Is there a circular import error if you remove it from within the function?
|
|
||
| # test for empty datasets | ||
| while len(level0_commondata_wc) != len(level1_commondata_instances_wc): | ||
| to_append = [] | ||
| lv1_names = [lv1_cd.setname for lv1_cd in level1_commondata_instances_wc] | ||
| for lv0_cd in level0_commondata_wc: | ||
| if lv0_cd.setname not in lv1_names: | ||
| to_append.append(lv0_cd) | ||
| for cd in to_append: | ||
| level1_commondata_instances_wc.append(cd) | ||
|
|
|
Hi Francesco, |
|
I have found a bug @FrancescoMerlotti , 
|
|
On the physics side, I have run a contaminated closure test with both the branches and the cuts seem to not be applied on the new branch. Here is a luminosity plot which shows it: |
|
After investigation, I find that the high energy points have been cut in the
In a closure test the fit is performed on the filter files, is that correct? |
|
The filter file is full of 0s where there are the cuts, hence it should be alright. I am trying to understand where the |
|
You are right about the filter file. Now I am very confused as to how my new branch fit replicates the no cut fit... |
|
I am a bit confused as well. |
|
I suggest that you try with a more stringent cut, something like 180 GeV |
|
I am currently running a fit without any datapoint above Q = 150 GeV |
|
By the way does the fit still error out if all the datapoints of an observables are cut? It is definitely the case on the previous branch |
|
now it should be ok, for example, in with a 150 GeV cut all the data in HLLHC datasets are cut |
|
I run the kinematic coverage on two fits generated with this and the old branch, and the cuts seem to be applied the same way |
|
Hi @FrancescoMerlotti , |
|
Hi @FrancescoMerlotti , I am not sure what happened last time, I must have messed up something in my fit of failed to pull the latest code updates, my bad... I am happy to merge it now. |
Glad to here this, lovely, we can merge after we finish the runs, or do you want to do that before? |
3 participants
Added validphys.config.parse_added_filter_rules to parse the added_filter_rules in the vp-setupfit runcard, and updated the validphys.config.produce_rules function.