Skip to main content

Static analysis framework for ROS.

Project description

HAROS is a framework for static analysis of ROS-based code. It has been published in the IROS 2016 conference. If you want to cite HAROS in your publications, please cite the original paper.

Static analysis consists on extracting information from the source code without executing it (and, sometimes, even without compiling it). The kind of properties that can be verified include simple conformity checks, such as checking whether a variable is initialised, to more complex properties, such as functional behavior of the program. This allows early detection of problems in the software development life cycle, which would otherwise go unnoticed into later stages or even into production.

Needless to say, sometimes in robotics it is very hard (or very expensive) to properly test software, not to talk about possible risks. Hence the appeal of static analysis.

Current Status

HAROS is still being developed, as of October 2017, even though it may be evolving at a relatively slow pace.

Installation

Here are some instructions to help you get HAROS running in your machine. This assumes that you already have a working installation of ROS. HAROS has been tested with ROS Indigo and ROS Kinetic, on Linux Mint and Linux Ubuntu. These setups should provide you with most of the basic dependencies of HAROS, namely Python 2.7 and a Web browser (if you want to use the visualiser).

NOTE This tool depends on other analysis tools. If you would rather install these dependencies first, then Ctrl+F :math:`dependencies`. Otherwise, just keep reading.

Method 1: Running Without Installation

There is an executable script in the root of this repository to help you get started. It allows you to run haros without installing it. Make sure that your terminal is at the root of the repository.

cd haros
python haros-runner.py <args>

You can also run it with the executable package syntax.

python -m haros <args>

Method 2: Installing HAROS on Your Machine

You can install HAROS from source or from a wheel. Either of the following commands will install HAROS for you.

[sudo] pip install haros
python setup.py install

After installation, you should be able to run the command haros in your terminal from anywhere.

Requirements

Before you can actually run analyses with HAROS, you need to perform some initialisation operations. These operations include downloading a basic set of analysis plugins. Do so with:

haros init

Note: if you opted for running HAROS without installing it, replace haros with your preferred method.

After initialisation, you still need to install some analysis tools that HAROS uses behind the curtains. Install these :math:`dependencies` with the following commands.

sudo apt-get install cppcheck

HAROS is now installed and ready to use.

Usage

Here is a basic example to help you get started with HAROS. Additional examples should be added in a future update.

HAROS works with the concept of index files. These files are more or less an equivalent to a project description, and they tell HAROS which packages you want to analyse. For this basic example, you should have the packages installed, and with available source code. If you run rospack find my_package and it returns the location of your package’s source code, you’re good to go.

HAROS will only use one index file at a time, but you can create as many as you want (e.g. one for each of your robots). The default index file (empty) lies in ~/.haros/index.yaml, but feel free to create your own, like so.

touch my_index.yaml
nano my_index.yaml

And my_index.yaml’s contents:

%YAML 1.1
---
packages:
    - package1
    - package2
    - package3

Now, you are ready to run analysis and visualisation on the given list of packages.

haros full -p my_index.yaml

The full command tells HAROS to run analysis and then visualisation. If you just want to run analysis, use the analyse command instead.

The -p option lets you specify an index file of your own, instead of using the default one.

When the analysis finishes, HAROS should start a visualisation server and your web browser on the appropriate page. To exit, just close your browser and press Enter on the terminal.

If you want to analyse several projects, or groups of packages, it is recommended to create an index file for each project, and define a project name as well. This way, HAROS will store analysis results separately. Example:

%YAML 1.1
---
project: my_robot
packages:
    - package1
    - package2

Below you can find the basic commands that HAROS provides.

haros init

This command runs initialisation and setup operations. This command needs to be run before the first analysis takes place. You can also run this command later on when you update HAROS.

haros analyse

This command runs analysis on a given list of packages.

haros analyse (no options)

Runs analysis with the list of packages found within the default index file (~/.haros/index.yaml). You are free to edit this file.

haros analyse -p INDEX_FILE

Uses the given index file to run the analysis, instead of the default one.

haros analyse -r

Uses repository information when available. If HAROS cannot find one of the packages you specified, it will look for it in the official ROS distribution and download it.

If your package is not in the official distribution, you can modify your index file to tell HAROS in which repository to look for the source (e.g. you can specify private repositories this way). Here is an example:

%YAML 1.1
---
packages:
    - my_package
repositories:
    repository_name:
        type:       git
        url:        https://github.com/git-user/repository_name.git
        version:    master
        packages:
            - my_package
            - another_package

The only supported repository type, for now, is git. There is partial support for hg and svn, but these have not been fully tested.

haros analyse -w PLUGIN [-w PLUGIN, …]

Whitelist the given plugins. The analysis will only run these plugins. This option does not work with -b.

haros analyse -b PLUGIN [-b PLUGIN, …]

Blacklist the given plugins. The analysis will not run these plugins. This option does not work with -w.

haros analyse -d DATA_DIR

Export analysis results to the given directory, instead of the default one. This option will also install the visualisation files. If DATA_DIR contains a previous analysis database for the current project within its tree, it will be loaded and new results will be added to that database.

Note: it is advised to use an empty/dedicated directory for this purpose. Previous versions deleted any existing files within DATA_DIR.

haros export

This command exports the analysis results (e.g. JSON files) to a location of your choosing. It assumes that some analyses were run previously.

haros export DATA_DIR

Exports analysis data to the given directory. This command will create files and directories within the given directory.

haros export -v

Export visualisation files along with analysis data.

Note: it is advised to use an empty/dedicated directory for this purpose. Previous versions deleted any existing files within DATA_DIR.

haros export -p PROJECT_NAME

Export a specific project’s data, instead of the default one. A special project name, all, can be used to export all available projects.

haros viz

This command runs the visualisation only. It assumes that some analyses were run previously.

haros viz (no options)

Launches the web visualiser and the visualisation server at localhost:8080.

haros viz -s HOST:PORT

Launches the web visusaliser and the visualisation server at the given host.

haros viz -d DATA_DIR

Serve the given directory, instead of the default one.

haros viz –headless

Start the viz server without launching a web browser.

haros full

Runs analysis and visualisation. This command accepts the same options as haros analyse and haros viz.

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

haros-2.1.2.tar.gz (714.2 kB view hashes)

Uploaded Source

Built Distribution

haros-2.1.2-py2-none-any.whl (1.5 MB view hashes)

Uploaded Python 2

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page