Add the possibility to set `edm4hep::utils::ParticleIDMeta` via the `IMetadataSvc` #273

tmadlener · 2024-12-20T15:38:02Z

BEGINRELEASENOTES

Make it possible to use edm4hep::utils::ParticleIDMeta with the MetadataSvc
- Add template specializations for get and put that defer to the corresponding utility calls in EDM4hep
- Add tests to ensure that metadata is indeed usable this way with the utilities in EDM4hep

ENDRELEASENOTES

This would be option 1 sketched out in #272

Small test that checks that metadata is also usable outside of framework
Needs Add documentation and a convenience overload to the ParticleID utilities EDM4hep#395

tmadlener · 2025-01-16T13:09:29Z

I have added a "small" test case that runs similar checks as in the functional algorithm as a standalone application (effectively hard-coding the expected parameter and algorithm names). I have done it in c++ because the PIDHandler is not yet really usable in python (see key4hep/EDM4hep#396).

Once the nightly builds have picked up key4hep/EDM4hep#395, CI should also start to pass again (tomorrow).

test/k4FWCoreTest/src/components/ExampleParticleIDConsumer.cpp

test/k4FWCoreTest/CMakeLists.txt

test/k4FWCoreTest/src/check_ParticleIDOutputs.cpp

Necessary because there is a new output collection in the functional producer that is used in several places

tmadlener · 2025-01-17T12:15:39Z

Since I touched the ExampleFunctionalProducerMultiple to also create ReconstructedParticles, I had to fix a few other tests to also make them pass. The alternative would be to introduce a dedicated producer for the ParticleID metadata tests. I am not sure if it's worth it, but it can obviously be done.

Technically already covered via Gaudi, but this makes it explicit

tmadlener · 2025-01-20T07:48:32Z

test/k4FWCoreTest/src/components/ExampleParticleIDConsumer.cpp

+  void operator()(const edm4hep::ParticleIDCollection& pidColl1, const edm4hep::ParticleIDCollection& pidColl2,
+                  const edm4hep::ReconstructedParticleCollection& recos) const {
+    auto pidHandler = edm4hep::utils::PIDHandler::from(pidColl1, pidColl2);
+    pidHandler.addMetaInfos(m_pidMeta1, m_pidMeta2);


This requires a new version of EDM4hep. Should I go for a version that works with 0.99? Otherwise we need to bump the EDM4hep version, I think. Even if this is only a test.

Up to you. But actually 0.99.1 in EDM4hep has the problem that key4hep/EDM4hep#379 is not merged in and this causes issues when trying to build on top of EDM4hep 0.99.1 since it won't find nlohmann_json, so having a tag with that is probably a good idea regardless.

jmcarcell · 2025-01-22T14:53:10Z

I'll give this another look and update key4hep/k4MarlinWrapper#216 to use this to see if anything is missing between today and tomorrow.

jmcarcell · 2025-01-22T18:01:14Z

test/k4FWCoreTest/src/components/ExampleParticleIDProducer.cpp

+
+  edm4hep::ParticleIDCollection operator()(const edm4hep::ReconstructedParticleCollection& recos) const {
+    auto pidColl = edm4hep::ParticleIDCollection{};
+    for (const auto r : recos) {


Suggested change

for (const auto r : recos) {

for (const auto& r : recos) {

jmcarcell · 2025-01-22T18:01:44Z

test/k4FWCoreTest/src/components/ExampleParticleIDProducer.cpp

+    return StatusCode::SUCCESS;
+  }
+
+  edm4hep::ParticleIDCollection operator()(const edm4hep::ReconstructedParticleCollection& recos) const {


Suggested change

edm4hep::ParticleIDCollection operator()(const edm4hep::ReconstructedParticleCollection& recos) const {

edm4hep::ParticleIDCollection operator()(const edm4hep::ReconstructedParticleCollection& recos) const override {

jmcarcell · 2025-01-22T18:02:52Z

test/k4FWCoreTest/src/components/ExampleParticleIDConsumer.cpp

+  }
+
+  void operator()(const edm4hep::ParticleIDCollection& pidColl1, const edm4hep::ParticleIDCollection& pidColl2,
+                  const edm4hep::ReconstructedParticleCollection& recos) const {


Suggested change

const edm4hep::ReconstructedParticleCollection& recos) const {

const edm4hep::ReconstructedParticleCollection& recos) const override {

jmcarcell · 2025-01-22T18:03:19Z

test/k4FWCoreTest/src/components/ExampleParticleIDConsumer.cpp

+#include <Gaudi/Property.h>
+
+#include "fmt/format.h"
+#include "fmt/ranges.h"


Suggested change

#include "fmt/ranges.h"

Only fmt::format is being used

jmcarcell

All good related to key4hep/k4MarlinWrapper#216

m-fila · 2025-01-23T11:16:13Z

CMakeLists.txt

@@ -30,6 +30,7 @@ find_package(ROOT COMPONENTS RIO Tree REQUIRED)
 find_package(Gaudi REQUIRED)
 find_package(podio 1.0.1 REQUIRED)
 find_package(EDM4HEP 0.99 REQUIRED)
+find_package(fmt REQUIRED)


Should fmt be also added to k4FWCoreConfig.cmake.in (and mentioned in README as a dependency)?

tmadlener mentioned this pull request Jan 16, 2025

Add support for running with IOSvc key4hep/k4MarlinWrapper#216

Open

2 tasks

tmadlener force-pushed the add-pid-metadata branch from ff46b52 to baecff1 Compare January 16, 2025 13:08

tmadlener changed the title ~~[WIP] Add the possibility to set edm4hep::utils::ParticleIDMeta via the IMetadataSvc~~ Add the possibility to set edm4hep::utils::ParticleIDMeta via the IMetadataSvc Jan 16, 2025

jmcarcell reviewed Jan 17, 2025

View reviewed changes

tmadlener added 6 commits January 17, 2025 10:48

Add overloads for handling ParticleIDMeta

74a1456

Add a producer and a consumer for PID with metadata

255c581

Add a test sequence for PID with metadata

0f660eb

Fix a few minor slips

8b45d86

Add a check for checking the standalone output

c3cb331

Move configuration into global variables

60c885a

jmcarcell force-pushed the add-pid-metadata branch from b2dc6c8 to 60c885a Compare January 17, 2025 09:48

tmadlener added 3 commits January 17, 2025 10:59

Make test functions consistent and check specific values

3803077

Add license header

e4a1ff4

Fix unrelated test failures

cc8a2af

Necessary because there is a new output collection in the functional producer that is used in several places

tmadlener added 2 commits January 17, 2025 13:16

Find fmtlib in cmake if we build the tests

50d288a

Technically already covered via Gaudi, but this makes it explicit

Find fmtlib unconditionally

842a5d5

tmadlener commented Jan 20, 2025

View reviewed changes

jmcarcell reviewed Jan 22, 2025

View reviewed changes

jmcarcell approved these changes Jan 22, 2025

View reviewed changes

m-fila reviewed Jan 24, 2025

View reviewed changes

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Add the possibility to set `edm4hep::utils::ParticleIDMeta` via the `IMetadataSvc` #273

Add the possibility to set `edm4hep::utils::ParticleIDMeta` via the `IMetadataSvc` #273

tmadlener commented Dec 20, 2024 •

edited

Loading

tmadlener commented Jan 16, 2025 •

edited

Loading

tmadlener commented Jan 17, 2025

tmadlener Jan 20, 2025

jmcarcell Jan 22, 2025 •

edited

Loading

jmcarcell commented Jan 22, 2025

jmcarcell Jan 22, 2025

jmcarcell Jan 22, 2025

jmcarcell Jan 22, 2025

jmcarcell Jan 22, 2025

jmcarcell left a comment

m-fila Jan 23, 2025

	edm4hep::ParticleIDCollection operator()(const edm4hep::ReconstructedParticleCollection& recos) const {
	edm4hep::ParticleIDCollection operator()(const edm4hep::ReconstructedParticleCollection& recos) const override {

	const edm4hep::ReconstructedParticleCollection& recos) const {
	const edm4hep::ReconstructedParticleCollection& recos) const override {

Add the possibility to set edm4hep::utils::ParticleIDMeta via the IMetadataSvc #273

Are you sure you want to change the base?

Add the possibility to set edm4hep::utils::ParticleIDMeta via the IMetadataSvc #273

Conversation

tmadlener commented Dec 20, 2024 • edited Loading

tmadlener commented Jan 16, 2025 • edited Loading

tmadlener commented Jan 17, 2025

tmadlener Jan 20, 2025

Choose a reason for hiding this comment

jmcarcell Jan 22, 2025 • edited Loading

Choose a reason for hiding this comment

jmcarcell commented Jan 22, 2025

jmcarcell Jan 22, 2025

Choose a reason for hiding this comment

jmcarcell Jan 22, 2025

Choose a reason for hiding this comment

jmcarcell Jan 22, 2025

Choose a reason for hiding this comment

jmcarcell Jan 22, 2025

Choose a reason for hiding this comment

jmcarcell left a comment

Choose a reason for hiding this comment

m-fila Jan 23, 2025

Choose a reason for hiding this comment

Add the possibility to set `edm4hep::utils::ParticleIDMeta` via the `IMetadataSvc` #273

Add the possibility to set `edm4hep::utils::ParticleIDMeta` via the `IMetadataSvc` #273

tmadlener commented Dec 20, 2024 •

edited

Loading

tmadlener commented Jan 16, 2025 •

edited

Loading

jmcarcell Jan 22, 2025 •

edited

Loading