GitLab

That was introduced by two recent merge requests crossing.

The goal here is to introduce a single python configurable in
AnalysisBase, instead of the current scheme in which there is a
separate configurable for AnaAlgorithm, AsgTool, etc. and essentially
duplicating a lot of code.

For now this duplicates even more code.  Once this is all tested and
validated, the old code will eitehr be completely removed or made to
inherit from the new classes.

I presumed it was fine to pass a `nullptr` into the algorithm setter
function, but apparently it is not.

The goal of this structure is to provide an easy way to pass around
all the information an algorithm needs about the context it is running
in.  This allows `AnaAlgorithmConfig` to provide an initialized
algorithm to the caller, but more importantly it should allow further
restructuring in the future.

That way it can hopefully have an Athena-implementation as well.

I completely overlooked initializing public tools.

Before it was only used for algorithms, which have no nesting, but
tools can have nesting.

Essentially I split up AnaAlgorithmConfig into two classes, one
containing all the parts that work for any AsgComponent and then a
very stripped down AnaAlgorithmConfig that contains only the
AnaAlgorithm specific parts.

This is mostly based on AnaAlgorithmConfig minus the parts that were
specific to AnaAlgorithm.  I also derived AnaAlgorithmConfig from this
class, allowing me to avoid

Former-commit-id: 27da0b0e623d3ab7f41731a9e6d5aadff525b5b2

Attila Krasznahorkay authored May 28, 2018 and Attila Krasznahorkay committed May 29, 2018

This addition helped me make the jet analysis job work correctly
with EventLoop. Since the public tool doesn't have a parent that
we would need to set a property on.


Former-commit-id: 84e81806551b6998cc89e92db01afa75136f42a8

In order to allow GCC 5.4 to build the code, removed an implicit conversion from it.
Which modern compilers can deal with correctly, but GCC 5.4 can not.


Former-commit-id: 3f91a3feb254c5763aa7eb2fa1aaa2a645dae7c3

Maybe done a little silly and completely over-organized, but it seems
to sort of work.


Former-commit-id: 3108500e80a5158a47040bf5cf3786a7193f8ffb

So far it was the user, but some people took issue with that, and
admittedly this is a cleaner design.


Former-commit-id: fe686d9f65b2350139794365aa9dd3c7534fef63

Changes made according to MR !8280.


Former-commit-id: ff196c505eb8fce891023b2772675962f87f5084

This should allow creating tools and nested sub-tools to arbitrary
depths, including configuring them and having them cleaned up at the
end.

Note that so far I haven't tested the nested sub-tools at all.


Former-commit-id: bdb943e0bcea3977c4163085dd144f71c36ba99e

Also fixed some function signatures.


Former-commit-id: fb3eb4b3abc045036d7b738dc52496762d1f622f

That should make configuration easier in some situations.


Former-commit-id: 6d72e3d16d5cf5b322a3733af3deeb71b7b8c25c

non-templated means. Mostly to allow setting properties from Python.


Former-commit-id: 34812d407f11f5db083974d2328de1c1c74067aa

Apparently the dictionaries don't get auto-loaded and I need to load
them manually instead.


Former-commit-id: 154071b45187fe9801d5f2fb518fc813da9c2107

Seems a fair rename, as Attila didn't like the name EventLoopComps.


Former-commit-id: ab6a6ccbce6199312212dce4cd4144a18dc881d0

There were quite a number of things broken/missing which are fixed
now.


Former-commit-id: cbe6acc5e7bf1d05b0d3607ac33fc231fe226cd2

Mostly to double check that I can really create an AnaAlgorithm and
set a property.  Unfortunately, I have some serious problems right now
in running the unit tests locally, as it seems to be picking up the
libraries from the wrong location. :(


Former-commit-id: 1a99656b8ef385b69f84b8c5f391988076fce269

This should address a number of issues with the existing algorithm
base class.


Former-commit-id: b3ef0a61e9a5f21d1114350f562e381609b1279d