PodioLegacyDataSvc vs PodioDataSvc?

[ebrondol]

Running locally I noticed that the code is broken since PR#42.

In file included from /afs/cern.ch/user/e/ebrondol/work/clue/k4Clue/src/ClueGaudiAlgorithmWrapper.cpp:1:
/afs/cern.ch/user/e/ebrondol/work/clue/k4Clue/src/ClueGaudiAlgorithmWrapper.h:52:3: error: 'PodioLegacyDataSvc' does not name a type; did you mean 'PodioDataSvc'?
   52 |   PodioLegacyDataSvc* m_podioDataSvc;
      |   ^~~~~~~~~~~~~~~~~~
      |   PodioDataSvc

Unfortunately, the CI was not running for the past 4 weeks and thus, the problem was not spotted before.

[tmadlener]

Is the CI running against the nightlies or against the release? Because for the nightlies the PodioLegacyDataSvc should be available.

Having had a brief look at the code, it also looks like the DataSvc is mainly needed to get at the CellIDEncoding, which could also be done via the newly introduced MetaDataHandle, so no need to explicitly have any DataSvc in the code any longer.

[ebrondol]

Hi @tmadlener , currently both are broken (for different but related reasons?). I have started a fake PR to see the status of both CIs: #44

Yes, I don't have anything against switching to MetaDataHandle. Is there an example of the code that I can have a look at? Thanks.

[tmadlener]

Yes, the "related reason" is key4hep/k4FWCore#100 ;)

There is a minimal example in the tests of k4FWCore (in both cases highlighting where the MetaDataHandle is used for getting data from or or putting data in it, there is a bit of setup necessary, but it should be somewhat straight forward to get that from the examples):

• reading: https://github.com/key4hep/k4FWCore/blob/master/test/k4FWCoreTest/src/components/k4FWCoreTest_cellID_reader.cpp#L25
• writing: https://github.com/key4hep/k4FWCore/blob/master/test/k4FWCoreTest/src/components/k4FWCoreTest_cellID_writer.cpp#L17

[ebrondol]

Update:

• MetaDataHandle is currently available only for the nightlies
• I have tested locally switching to MetaDataHandle, but the input file used to test it does not contain podio_metadata
• I have followed the official recipe, using the nightlies, to produce the new sample, but the reconstruction with input edm4hep files is not working due to

ToolSvc             ERROR Error initializing tool 'ToolSvc.Lcio2EDM4hep'
Output_REC          ERROR Failing initializing ToolHandles : Output_REC      Failed to retrieve tool Lcio2EDM4hepTool/Lcio2EDM4hep     StatusCode=FAILURE
EventLoopMgr        ERROR Unable to initialize Algorithm: Output_REC

@tmadlener suggested to wait for key4hep/k4MarlinWrapper#120 to be finalized and it should solve the issue.