Invalid branch content for mc truth branches in DV tutorial5

I noticed that the DV tutorial tutorial5_mctruth.py generate a tuple with invalid branch contents. In particular the branches with the mc truth information X_TRUEP and X_TRUEID are not filled correctly since their value are always nan and 0, respectively. The corresponding test pass, since it checks only if the tuple is not empty.

The idea is to fix the tutorial in order to have a working example and also improve the DV tests checking not only if the final tuple is filled but also the checking the validity of the values in the .root file.

assigned to @dfazzini

changed the description

Outch! Well spotted! Indeed the contents are just as important. CC @amathad to ensure that the various tutorials are "as robust as possible".

It is strange because this was not the case before i.e. we did a successful starter-kit lesson with these samples. So something has changed since that has broken this. My suggestions was to indeed look into values of observables for DV testing.

So also assigning to you? (Well I just did).

added testing label

assigned to @amathad

If I understand correctly, truth matching needs PP2MCP relations. Log file for this test says they are not in the input file.

PP2MCPRelationUnpacker                 INFO Number of counters : 1
 |    Counter                                      |     #     |    sum     | mean/eff^* | rms/err^*  |     min     |     max     |
 | "# UnknownBuffer"                               |        12 |          0 |      0.0000 |
PP2MCPRelationUnpacker#1               INFO Number of counters : 1
 |    Counter                                      |     #     |    sum     | mean/eff^* | rms/err^*  |     min     |     max     |
 | "# UnknownBuffer"                               |        12 |          0 |      0.0000 |

what is the input to this test? Can you point me to it?

https://gitlab.cern.ch/lhcb/DaVinci/-/blob/master/DaVinciTutorials/options.yaml

@dfazzini @amathad Can you test with !767 (closed) to see if the problem is resolved?

I checked that this gets fixed in !731 (merged). Perhaps we give precedence to that.

,root [6] TTreeName->Scan("Bs_TRUEID:Jpsi_TRUEID:Phi_TRUEID:Kp_TRUEID:Km_TRUEID:Mum_TRUEID:Mup_TRUEID")
************************************************************************************************
*    Row   * Bs_TRUEID * Jpsi_TRUE * Phi_TRUEI * Kp_TRUEID * Km_TRUEID * Mum_TRUEI * Mup_TRUEI *
************************************************************************************************
*        0 *         0 *         0 *         0 *      2212 *      -321 *      -321 *       321 *
*        1 *         0 *         0 *         0 *      2212 *         0 *      -321 *       321 *
*        2 *         0 *         0 *         0 *      2212 *      -321 *      -321 *       321 *

Nice! I did not check the long MR - does the to-do above "The idea is to fix the tutorial in order to have a working example and also improve the DV tests checking not only if the final tuple is filled but also the checking the validity of the values in the .root file." get fixed? Then great.

So that issue can be closed?

also improve the DV tests checking not only if the final tuple is filled but also the checking the validity of the values in the .root file

This is not in either of the MR Sevda's (!767 (closed)) or mine (!731 (merged)). I can add that in mine and suggest a snippet to Sevda.

Apologies, this seems to be my fault again. configured_MCTruthAndBkgCatAlgneeds to know the options.process. Default is Spruce so it looks for the relations at the wrong place. I will commit a fix ASAP.

mentioned in merge request !767 (closed)

Does this get fixed with !731 (merged) ?

Yes.I have edited that MR by mentiong the magic work "closes", so once merged it should automatically close this.

mentioned in merge request !731 (merged)

mentioned in commit 87085f94

closed with merge request !731 (merged)