Appendix V: Notes taken at meeting of July 24 on this subject¶
From Mark F.:
Some SAM metadata values are known before the job begins. What distinguishes data tiers? * output of RECO, full list of products * skimmed * raw * raw filtered * route by group Per output module: * data tier * stream name * file format (e.g. ?artroot?) * file format era (as specified by art; generally not backwards compatible) * file format version (as specified by art; generally guarantees backwards compatibility) The "data tier" is to be specified on the art command line: parameter name is "--sam-data-tier=module1:something" module1 is an example module label; this can be repeated as many times an needed for however many modules you produce. The "stream name" can be specified as "--sam-stream-name=module1:something?. If the value is lacking in the configuration of the output module, then the module label is used as the stream name. The default value forbids two modules using the same stream name; this is only possible if the stream name is specified by the user. Per program execution: * application family: command line parameter "--sam-application-family=something" * application name: this is the art "process"; --process=something will be a new command line parameter. * application version: command line parameter "--sam-application-version=something" * file type: command line parameter "--sam-file-type=something" If any of these SAM metadata values are specified, all must be specified, either through default values or through being specified on the command line. Using SAM for File Delivery To use the SAM file-delivery service, both of the following command-line parameters must be specified. If only one is specified, it is an error. If both of these are supplied, the parameters for writing of SAM metadata (above) must be specified. * Web URI: command line parameter "--sam-web-uri=something" * SAM process id: command line parameter "--sam-process-id=something" For any program that used the SAM file delivery service, these items should be added to the SAM metadata. Experiment decides what is important (to record to user database). We also discussed having a means of allowing experiments to select what parameters from the input parameter sets they might save to their own database -- but this is NOT a SAM issue, this is a "user database" issue, and so we did not pursue it to a conclusion.
From Marc P.:
Present were: Adam L, Marc M, Mark F, Chris G, Marc P; on phone, Andrew N Topic: SAM use case discussion [Original March 13, 2012] Some SAM metadata values are known before the job begins. What distinguishes data tiers? • output of RECO, full list of products • skimmed • raw • raw filtered • route by group Per output module: • data tier • stream name • file format (e.g. “artroot”) • file format era (as specified by art; generally not backwards compatible) • file format version (as specified by art; generally guarantees backwards compatibility) The “data tier” is to be specified on the art command line: parameter name is “--sam-data-tier=module1:something” module1 is an example module label; this can be repeated as many times an needed for however many modules you produce. The “stream name” can be specified as “--sam-stream-name=module1:something”. If the value is lacking in the configuration of the output module, then the module label is used as the stream name. The default value forbids two modules using the same stream name; this is only possible if the stream name is specified by the user. Per program execution: • application family: command line parameter “--sam-application-family=something” • application name: this is the art “process”; --process=something will be a new command line parameter. • application version: command line parameter “--sam-application-version=something” • file type: command line parameter “--sam-file-type=something” If any of these SAM metadata values are specified, all must be specified, either through default values or through being specified on the command line. Using SAM for File Delivery To use the SAM file-delivery service, both of the following command-line parameters must be specified. If only one is specified, it is an error. If both of these are supplied, the parameters for writing of SAM metadata (above) must be specified. • Web URI: command line parameter “--sam-web-uri=something” • SAM process id: command line parameter “--sam-process-id=something” For any program that used the SAM file delivery service, these items should be added to the SAM metadata. Experiment decides what is important (to record to user database). We also discussed having a means of allowing experiments to select what parameters from the input parameter sets they might save to their own database -- but this is NOT a SAM issue, this is a “user database” issue, and so we did not pursue it to a conclusion. [Additions made March 28, 2012] Per output module calculated items: