This is a very large PR, I got carried away. The changes and reasoning are documented in the following description, organized by type of change:

Bug Fixes

Although this PR is supposed to be a refactor, the following bugs were fixed:

• timelines RF -> "rftime 12 diff*" are now being filled with data; the RF period from CCDB was not actually being used in monitor2p2GeV.java
• reliance on external coatjava and having a groovy version that matches the POM version are no longer needed; the Maven build is now effectively standalone

Renames

• "monitoring" Java classes are now referred to as "histogramming" classes, and
  are all part of the package org.jlab.clas.timeline.histograms
  • they are now run by run_histograms.java (renamed from ana_2p2)
• "detectors" Groovy classes, for detector timelines, are now referred to as
  "analysis" classes, and are all part of the package org.jlab.clas.timeline.analysis
  • they are now run by run_analysis.groovy (renamed from run_detectors.groovy)
• updated documentation accordingly
• renamed the following histogramming classes:
  • cndCheckPlots.java -> CND.java
  • central.java -> CTOF.java
  • monitor2p2GeV.java -> GeneralMon.java
  • tof_monitor.java -> DCandFTOF.java
• rename analysis class methods for consistency with histogramming class methods:
  • processDirectory -> processRun (cf. histogramming classes' processEvent)
  • close -> write (cf. histogramming classes' write)

Removals

Histogramming Class Cleanup

• remove unused variables and includes
  • variables that are obviously unused have been deleted
  • variables that may have been used for debugging have been commented out
• remove retrieval of unused banks in histogramming's processEvent methods
  • NOTE: this should hopefully speed things up a bit
• run_histograms.java (renamed from ana_2p2.java) should only have to call
  processEvent and write for each histogramming class
  • this makes calling the monitor classes a bit more general: for each event,
    call processEvent, at the end, call write
  • in some cases, two methods were called after event processing; these
    additional methods are now just called by write instead
  • there are still variations in the methods' parameters, as well as their
    constructors, but that's fine for now
• remove main and plot methods
  • they just mostly contained stuff to make plot PNGs
  • plot viewing should use the web front-end, not custom canvases and PNG
    output; we can allow webserver access to those who need it.
• remove any other unused methods

Histogramming Class Removals

We have removed the following histogramming classes, none of which have been
used in quite a while:

• occupancies.java, for BST and BMT occupancy
  • BST and BMT occupancies appear to be handled by monitor2p2
• deuterontarget.java
  • many plots are handled elsewhere, namely by monitor2p2
• dst_mon.java
  • The produced histograms are not being plotted; they have been disabled for
    some time, since Daniel Carman said they are not needed.
  • There are several other histograms being filled in dst_mon, beyond the
    FTOF ones; however, most of them are either unused or produced by something
    else, namely monitor2p2.

Analysis Class Removals

• remove unused CTOF and FTOF analysis groovy classes
  • these all have been disabled for some time, since Daniel Carman said they
    are not needed.

Build Changes (Maven)

• Use the same compiler plugin for all modules: groovy-eclipse-compiler
  • we are already using it for the Groovy code
  • it can handle mixed Java+Groovy projects pretty well
  • required adding a new Maven repo, so we can use the latest version
    • told dependabot to track it too
• combine detectors and monitoring modules to a single module
• organize the source tree; we now have the following packages:
  • org.jlab.clas.timeline: main package for all timeline code
  • org.jlab.clas.timeline.histograms: reads HIPO data files and makes histograms
  • org.jlab.clas.timeline.analysis: analyzes histograms
  • org.jlab.clas.timeline.fitter: fits histograms; analysis classes call fitter classes
  • org.jlab.clas.timeline.util: miscellaneous common stuff
• do not require groovy JARs to be in caller classpath
  • The current kluge to find Groovy JAR files at runtime fails in Arch
    Linux (and therefore in our container images), but has been working fine
    in production on ifarm. We should not need to add Groovy JARs to the
    class path; instead we can simply copy dependencies to the installation
    directory. I added the script install.sh, since the mvn command is
    slightly complicated.
  • now we no longer have to have coatjava's JAR file and Groovy JAR files
    in the class path (in environ.sh); we just need target and
    target/dependency.
    • we have also removed any need for coatjava environment variables ($COATJAVA, $PATH, etc.)
  • strangely, this required us to switch from YamlSlurper to snakeyaml, because
    of some dependency conflict
• force all groovy calls to use the Maven build's version of groovy
  • to make sure there are no version conflicts with a user's system groovy
  • replaces Coatjava's run-groovy with our run-groovy-timelines.sh
  • cleaned up associated environment variables etc.
• correct the version number in pom.xml
  • we'll call this 2.99.0, since we have more work to do before tagging 3.0.0

Miscellaneous

• make sure the URI for RCDB is consistent everywhere, using environment variable $RCDB_CONNECTION

Style Changes

• fix indentation in Java files; replace tabs with spaces
• resisted the urge to run an auto-formatter, to avoid making the git diff impossible to understand