Full metadata in output

[hansenjo]

Write all available analysis metadata (replay time/host/directory, user, database keys loaded, etc.) to the ROOT output file to make data provenance as transparent as possible.

[hansenjo]

Some ideas as to what specific metadata to include:

• [ ] Analyzer version (version & git commit)
• [ ] Actual ROOT version loaded at run time
• [ ] EVIO version loaded at run time
• [ ] Build time, host, compiler, user, directory
• [ ] Build flags (CXXFLAGS, others?)
• [ ] Replay time, host, user, work dir
• [ ] Analyzer flags configured (scalers, helicity)
• [ ] Analyzer class, decoder class
• [ ] Database type, URL (DB_DIR, SQL host, etc.)
• [ ] All database keys & values loaded, load times
• [ ] Replay script? Copy of source?
• [ ] Output & cut definition files (full copies)
• [ ] All output definitions
• [ ] Full paths of input & output file(s)

This is just a brainstormed list. Some of the items may not be realistic to include, and I'm sure I'm forgetting quite a bit. Feel free to add/modify.

Also, THaRun should be improved to include more or more accurate information about the run data.

[hansenjo]

Add to this list: A map of variable name prefixes (like "L.s0") to the name of the class using the prefix (e.g. THaScintillator). With this map, one can tell the meaning of each variable without working through the details of the replay script. Admittedly, one still doesn't always know which exact detector is referenced (e.g.: what exactly are "L" and "s0" in this example?), but at least one then knows what code was used to calculate the values.