Architecture-Guided Test Resource Allocation Via Logic

(TAP2021 Publication Supporting Data)

• Find the experimental data in the evaluation subfolder, including the trigger scripts.
• The Python API and helper classes are located in pyASTRAHL.
• Evaluation scripts and Jupyter notebooks are in notebooks
• src contains the main tool source code (and the Makefile to build it)
• Below you can find some help for building and using it.

Installing

• Install cJSON
• Clone this repository
• Go to the src directory
• Launch make

Install procedure cJSON:

git clone https://github.com/DaveGamble/cJSON.git
cd cJSON
mkdir -p build
cmake ..
make
make install

Install procedure gc

apt-get install -y autoconf automake libtool &> /dev/null

# this is the stuff for gc
git clone git://github.com/ivmai/bdwgc.git ; 
cd bdwgc ; 
git clone git://github.com/ivmai/libatomic_ops.git ;

# Build gc
cd bdwgc 
autoreconf -vif 
automake --add-missing 
./configure 
make
make install 

For Python usage (convenience API)

sudo apt-get install jq graphviz
pip install graphviz libgraphviz-dev joblib tqdm numpy pandas

Using

Typical usage (not implemented yet):

./main -f input.json

(See "JSON format" section for more information.)

Options

./main --test xxx

Launches a bunch of unit tests for a given file. Unexpected results are printed in red

./main --test [all]

Launches a bunch of unit tests on all files. Unexpected results are printed in red.

JSON format

This tool uses JSON format to describe:

• the input fault tree,
• the confidence formulas for all components,
• the resources already spent on all components,
• the whole amount of resources to spend on the problem.

Fault trees

Connectives:

• wire: describes a base wire (component), needs:

  • index: an integer for identification,

• and, or, pand: gates, need:

  • subtree1, subtree2: JSON objects representing the sub-fault trees.

For an example, see fault_tree.json.

Expressions

Connectives:

• const: describes a constant, needs:

  • constant: a double,

• var: describsed a variable, needs:

  • index: an integer for identification,

  • name: a name for the variable (for printing),

  • WARNING: no check is made to verify the naming scheme is coherent,

• opp: unary operations, need:

  • expression1: JSON object representing the sub-expression.

• add, sub, mul, div: binary operations, need:

  • expression1, expression2: JSON objects representing the sub-expressions.

For an example, see expression.json.