GitLab

Changes:

* add new seed handling: By default new scenario files will be created
  with a random seed, which is written in the attribute fixedSeed.
  If the Boolean switch useFixedSeed  is set, this defined seed will
  be used within one simulation. At the start of each simulation this
  seed, or if useFixedSeed is false a newly generated one, will be used
  and persisted in the attribute simulationSeed. All generated data
  of this simulation will be written to the defined output directory
  with a copy of the current scenario file. All data generated in this
  simulation run can be reproduced with specified simulationSeed.

* new features for VaderConsole(migrate): The migrate subcommand will
  now accept a list of files or Directories (or a mix of them) to
  perform a migration or a revert. If a path is specified the
  default behavior will only migrate the selected directory without
  traversing to the directory tree. If the --recursive or -r flag is
  set, the hole sub tree of all given directories will be converted.
  Directories or directory trees will be ignored if the directory
  contains one of the following marker files:
  - DO_NOT_MIGRATE_TREE or .DO_NOT_MIGRATE_TREE
  - DO_NOT_MIGRATE or .DO_NOT_MIGRATE

* remove hard coded sc  enario files encoded as strings to allow
  testResource migration with the VadereConsole

* ignore scenario.legacy files within testResources

* add test for MigrationAssistant v0.3 to v0.4

[BugFix] in NumberOverlapsProcessor. Before, the occurrences of overlaps were counted and divided by two (since in at least two pedestrians were part of each overlap). Now the number of overlaps is counted for each pedestrian and the result is divided by two. Also, the number displayed in the infoBox (if uncommented) was corrected.

Changes:

* Add UseFixedSeed (bool switch) and fixedSeed Attribute values
  This allows to rerun the simulation with different but fixed seeds
  for random numbers.
* Add simulationSeed to save currently used seed while simulation is
  running. This is useful to determine the seed used for processor
  outputs in the output directory.
* Add Test for new version and refactor code duplicates
* Change Hash calculation to exclude UseFixedSeed, fixedSeed and
  simulationSeed to highlight (green) output to selected scenario
  within the gui.

[Bugfix] NumberOverlapsProcessor: The number of overlaps needs to be divided by two since the overlaps occurs for both pedestrians that overlap.

Added a new outputprocessor, namely "NumberOverlapsProcessor" that is dependent of the PedestrianOverlapProcessor. It saves the number of overlaps that occurred during the simulation run.

Features (MigrationUtil):

* create new Transformation and Identity files: This will use the
current identity    Transformation as starting point for the new
version. The new version string is automatically included in the new
files. The files will be created based   on the defined naming
convention and saved in VadereSimulator/resources

* automate scenario migration without gui: The MigrationUtil will
migrate   all scenario files within the specified directory and all
child directories.

To exclude specific sub-trees or only specific
directories the igonoreDirs List can be expanded.

  To exclude sub-trees or specific directories without code change add
  one of the following marker-files to the directory.

  * DO_NOT_MIGRATE or .DO_NOT_MIGRATE: Ignore current directory but
    continue with existing child directories.

  * DO_NOT_MIGRATE_TREE or .DO_NOT_MIGRATE_TREE: Ignore the directories
    and the complete sub-tree.

  Use the marker files from the Tools directory. These files have a
  comment to explain for what they are used for.

  The content of the file is not used, only the file name is needed.
  These files must be added to the git cache to work.

 * automate migration revert: This will use the revert function of the
   MigrationAssistant to revert a migration if the <file-name>.legacy
   files still exist. For this to work the legacy file must be in the
   same directory as the scenario-file

* add helper methods to crate backup files with suffix
* add migration revert function to MigrationAssistant
* move migration logic from VadereConsole back to MigrationAssistant
* add Test for MigrationSubCommand

changes:

* move groupSizeDistribution from AttributesCGM to AttributesSources
* rename standardDerivation to standardDeviation in AttributesTimeCost
* update all existing scenario files to version v0.3

The VadereConsole command line tool is know able to migrate old
scenario files to a new version. All old capabilities were moved
to the subcommands project-run and suq. See usage below

usage: Vadere [-h] COMMAND ...

Runs the VADERE pedestrian simulator.

optional arguments:
  -h, --help             show this help message and exit

subcommands:
  valid subcommands

  COMMAND
    project-run          This command uses a Vadere Project and runs selected scenario.
    suq                  Run a single scenario file to specify to  fully controll folder structure for input and output.
    migrate              Run migration assistant on single sceanrio file

In "run_vadere_console_with_all_scenario_files.py", run ALL scenario files, catch exceptions and log exceptions.

Exit with return code 1 if at least one test failed. Otherwise, return 0.

This only affects scenarios which do not us group models. Due to
refactoring of the group model the group size distribution is
set by source. This allows different source to spawn different group
sizes.

In "run_vadere_console_with_all_scenario_files.py", print runtime (in seconds) for each scenario file.

Moved CI helper scripts to "Tools/ContinuousIntegration" and enabled Java assertions in POM files and helper scripts.

.gitlab-ci.yml:
- Refactored complete file: moved comment on top of file.
- Call helper scripts in "Tools/ContinuousIntegration" instead of just "Tools".

Tools/ContinuousIntegration/run_vadere_console_with_all_scenario_files.py:
- Execute scenario files alphabetically.
- Use "-enableassertions" when invoking Java.

POM files:
- Use "<enableAssertions>true</enableAssertions>" to be independent of SureFire's default settings.

Each JoltTransformation is create as a Singletone from the abstract
JoltTransformation class. These classes manage the transformation and
postHook processing if needed. All resource files were renamend to
automatically build resources names based on the current version alone.

This addition is needed because not all transformation can be done
with the Jolt Transformation. Each implementation of JoltTransformation
has a list of PostTransformHooks which will be executed in order on
the JsonNode representation _after_ the Jolt transformation is done.

This allows context specific transformation. If needed a postHook will
be added in the future.