config-reproduction.properties

###
### GENERAL ###
###
# Should all repositories be cloned ahead of the evaluation? Warning: this may require several hundred GigaBytes of free disk space.
preload-repositories=false
# Delete each repository after it has been processed? This helps with freeing disk space, but makes it more difficult to rerun the evaluation later
# because the repositories have to be cloned again
clean-repositories=false
# The number of EXPERIMENT_TIMEOUT_UNIT to wait for a patcher to finish patching (long)
experiment.timeout.length=5
# The time unit for the timeout, e.g., SECONDS, MINUTES, ...
experiment.timeout.unit=MINUTES
# Number of threads for parallel execution. Note that the bottleneck might be the IO capacity of your disk.
experiment.thread-count=5
# Minimum number of cherries in a repository for it to be considered for the evaluation.
# -1 means all repositories are included.
experiment.dataset.min-size=-1
# Maximum number of commits in a repository for a dataset to be considered for the study. If a repository has
# more commits, it is simply ignored. Values of 0 or less are automatically converted to Integer.MAX_VALUE.
# -1 means all repositories are included.
experiment.dataset.max-size=-1
# Which patchers should be enabled?
experiment.patcher.gnu-patch=true
experiment.patcher.git-apply=true
experiment.patcher.git-cp=true
experiment.patcher.mpatch=true

###
### PATHS ###
# THESE SHOULD ONLY BE CHANGED IF YOU KNOW WHAT YOU ARE DOING
###
# The path to the list of dataset
experiment.datasets=/home/user/dataset/mined-cherries
# The path to the main working directory of the experiment
experiment.dir.main=/home/user/evaluation-workdir/main/reproduction
# The path to the results directory
experiment.dir.results=/home/user/evaluation-workdir/results/reproduction
# The path to the directory to which the SPL repos are cloned
experiment.dir.repos=/home/user/evaluation-workdir/REPOS
# The path to the file in which already processed evaluation runs are tracked
experiment.processed-file=/home/user/evaluation-workdir/results/processed-reproduction.txt

###
### SAMPLING ###
# ALLOWS FOR RUNNING THE EVALUATION ON A SMALL SUBSET OF THE DATASET
###
experiment.sample-file = /home/user/evaluation-workdir/reproduction-sample.ser
# Should the amount of data be reduced by sampling with a certain confidence?
experiment.enable-sampling=false
# (z) is the Z-score corresponding to your desired confidence level (for a 95% confidence level, (z = 1.96)),
# sampling.z=1.96
# Score for 99%
sampling.z=2.58
# (e) is the margin of error (in percent, such as 5% = 0.05).
# sampling.e=0.05
# smaller error
sampling.e=0.01
# (p) is the sample proportion (in percent, such as 50% = 0.5),
sampling.p=0.5
# A seed that is used during sampling to ensure repeatability of the experiments
sampling.seed=3
# The number of random experiment repetitions. In each run, random variants are generated and a random source variant
# is selected from which a diff is calculated. The higher the number of repeats the more random repeats for two specific
# SPL-commit pairs (inclusive start and end)
# This setting only makes sense when samping is enabled
experiment.repeats.start=1
experiment.repeats.end=1


###
### MISC ###
# THESE SHOULD ONLY BE CHANGED IF YOU KNOW WHAT YOU ARE DOING
###
# From which run id to start the experiments. Can be used to skip experimental runs. 
experiment.startid=0
# Which type of cherry-picks should be considered (Trivial, Complex, Both)? 
experiment.cherry-type = Complex
# Whether debug files should be written. If true, the evaluation will create DEBUG directories with various files in the working directories.
experiment.debug=false