enrich CMS RAW data records with run number information

[tiborsimko]

Current behaviour

For CMS collision records, we have run_numbers field that lists all the data-taking run numbers for the given dataset.

Example: record 1

$ grep -A 10 '"recid": "1"' data/records/cms-primary-datasets.json
    "recid": "1",
    "relations": [
      {
        "title": "/BTau/Run2010B-v1/RAW",
        "type": "isChildOf"
      }
    ],
    "run_numbers": [
      "146240",
      "146242",
      "146248",

However, the CMS RAW data samples do not have this information.

Example: record 65

$ grep -A 10 '"recid": "65"' data/records/cms-raw-datasets-Run2012B.json
    "recid": "65",
    "run_period": [
      "Run2012B"
    ],
    "system_details": {
      "container_images": [
        {
          "name": "docker.io/cmsopendata/cmssw_5_3_32-slc6_amd64_gcc472:latest",
          "registry": "dockerhub"
        },
        {

The only information about run numbers is present in the abstract, see "Run range [194439,194691]" there.

Expected behaviour

It would be good to add the run number information to all CMS RAW sample records as well:

• data/records/cms-raw-datasets-Run2010B.json
• data/records/cms-raw-datasets-Run2011A.json
• data/records/cms-raw-datasets-Run2012B.json

CC @IssaAlBawwab @nasseralbess @katilp @tpmccauley

[katilp]

this query, on lxplus with the grid proxy available (voms-proxy-init -voms cms -valid 192:00), will return the run number(s) for a file:

$ dasgoclient -query "run file=/store/data/Run2010B/Jet/RAW/v1/000/146/807/04DC3275-DFCA-DF11-B54B-003048F024FA.root"
146807

• Get the file list for each 16 datasets
• Replace root://eospublic.cern.ch//eos/opendata/cms with /store/data for each file
• Get all runs for all files with the command above
• Note that one run can have events in several files so N_files > N_runs

The dasgoclientquery requires a grid certificate.