Retirement of Analysis project

added cleanup label

changed the description

Without thinking too much, I would consider moving the C++ to Rec and the python to DaVinci.

If Funtuple goes to DaVinci, HltEfficiencyChecker in MooreAnalysis (which is dependent on it) I guess has to go there too. Then MooreAnalysis can probably also be retired. It was discussed elsewhere that this was probably the thing to do anyway, but no1 has got around to it yet. Just to let you know that it will be a requirement.

mentioned in issue #38 (closed)

For clarity - this suggestion was made at the past DPA WP3 meeting where all outstanding tasks and MRs were reviewed. It was felt that one could simplify matters by retiring Analysis the same way we did with Phys, given that most package in Analysis should in fact be retired by now and we believe most of them (listed in the header) aren't used anywhere. By retiring Analysis we also minimise the amount of time preparing and releasing the "DaVinci stack", simplify documentation, etc.

@amathad, 2 things we should do for clarity - email PAC and probably the OPG just so that everyone with maintenance/operations responsibilities is made aware, and set a DaVinci release version by which "point" the move will be made. The latter is standard for deprecation warnings and alike. Of course it should be a major release number.

One should be careful in getting everything to Rec as then also DaVinci can go as pure config/helpers and we loose any flexibility offline, having it all under Moore + dependencies.

If Funtuple goes to DaVinci, HltEfficiencyChecker in MooreAnalysis (which is dependent on it) I guess has to go there too. Then MooreAnalysis can probably also be retired. It was discussed elsewhere that this was probably the thing to do anyway, but no1 has got around to it yet. Just to let you know that it will be a requirement.

It would be about time that MooreAnalysis gets retired ;-). I never understood why it even got created in the first place! We have been asking for a loooooonnnnnng time for HltEfficiencyChecker to be modernised to FunTuple and to see what special there remained in MooreAnalysis, to retire it. I know certain moves got finalised over the past year. I hope this task here will give that extra push towards having things better organised/modernised/simplified overall. And next year is not the time to do these matters. We ought to get them done this year.

We (@clemenci , @rmatev, @cmarinbe and @amathad) also had a discussion about this in the PAC. I'm trying to summarize it here and then link this in the minutes of the PAC.

The general agreement was that retiring Analysis is good as Analysis has become very small.

Then the discussion was about where to move it or parts of it.

An advantage of moving some things from Analysis to Rec is that this would make the retirement of MooreAnalysis possible. For example, access to the MC tools would allow to test more easily Moore within Moore instead of requiring a workaround like MooreAnalysis or running Moore in DaVinci. Being able to run FunTuple in Moore could also be useful for testing.

The disadvantage, as also mentioned in other comments, would be that changes in FunTuple (or MC matching code) would require tagging and releasing Rec. This was not seen as a major show stopper as DaVinci anyhow already depends on Moore and a lot of the code required for DaVinci is already in Rec.

As not everyone interested in the discussion was presented at the meeting, the idea would be to discuss a bit further here and then follow up with a discussion in an RTA+DPA meeting to come to a conclusion.

Please comment if my summary is incorrect or missing some things.

mentioned in merge request !1021 (merged)

added DPA label

changed the description

changed due date to December 20, 2023

I have put a hard deadline for retiring Analysis by Dec 20th. Ofcourse some of the work will be done before that. Please take note of the deadline for the MooreAnalysis (Tag @rjhunter ). The conclusion of the discussion is here.

mentioned in merge request !1031 (merged)

mentioned in merge request !1032 (merged)

Morning @pkoppenb, I know that you have taken care of moving all packages around across projects, where seen relevant, as far as DPA is concerned (clever moving keeping the history). Would you be available to do it also for the package Phys/DaVinciMCTools, which we agreed to move to Rec? That would get us rather close to getting this task done. Thanks a lot.

assigned to @amathad

OK. Can do that. Fits nicely with my removal campaign.

mentioned in merge request !1033 (merged)

mentioned in merge request !1034 (merged)

changed the description

marked this issue as blocking Rec#358

mentioned in merge request Rec!3676 (merged)

mentioned in merge request Rec!3673 (merged)

mentioned in merge request MooreAnalysis!130 (closed)

mentioned in merge request MooreAnalysis!131 (merged)

mentioned in merge request !1036 (closed)

mentioned in merge request !1011 (closed)

Good morning @amathad, seems we're pretty close to being done here. This being said, there are several MRs open, which we need to deal with. Some are very old. I suggest we ping the authors one last time and then close them, requesting that they be done in DaVinci. Take the case of !964 (closed), which has been effectively dead since 9 months.

Thanks @erodrigu . Yes will do the rounds...

mentioned in issue MooreAnalysis#45 (closed)

marked this issue as blocked by MooreAnalysis#45 (closed)

I see @erodrigu has already noticed it, but I just mention here that MooreAnalysis#45 (closed) details the remaining steps for retirement there. I'd say 90% of the work is done, but we just require timely merging to make the retirement happen quickly/by the deadline.

mentioned in merge request MooreAnalysis!134 (merged)

changed due date to February 02, 2024

mentioned in merge request !1049 (merged)

mentioned in commit 9901239e

closed with merge request !1049 (merged)

Retirement of Analysis project

Designs

Child items ...

Activity

Retirement of Analysis project

Blocks

Is blocked by

Activity