Skip to content

Latest commit

 

History

History
191 lines (133 loc) · 7.63 KB

README.manual.md

File metadata and controls

191 lines (133 loc) · 7.63 KB
Raw

Notes for manual Codeface installations

NOTE: The information in this file is likely not entirely accurate. The preferred way to install Codeface is to use vagrant; the corresponding scripts in integration-scripts/ give the definitive steps for installing the system.

Note that only the sections /Git Clone/ and /Analysis Setup/ are relevant if your machine is already set up for Codeface. If you want to use your own database instance, you can search/replace 'codeface' with 'my_database_name' in the step "Database Setup", and modify codeface.conf accordingly.

Required Programs

  • Install GNU R, mysql, mysql-workbench 6 (see below), node.js and npm from the distribution repository

  • Install mysql-workbench 6 Version 5.x is not sufficient because there are subtle differences when it comes to handling the binary ER models that make life fairly hard. Since the latest revision is not included in Ubuntu 12.04, you need to download the package directly from Oracle (http://dev.mysql.com/downloads/tools/workbench/) and install it via the usual distribution mechanisms.

  • Graphviz often comes in ancient versions with distributions. For Ubuntu 12.04, use the recent packages fro AT&T: http://www.graphviz.org/Download_linux_ubuntu.php (>= 2.30 is fine; download and install the main package and libgraphviz4{,-dev}. There may be some additional prerequisites for the packages that can be satisfied from the distro repo)

  • Make sure that GNU R is available in a sufficiently new release. Not sure which one is exactly the oldest possible one, but Ubuntu tends to have fairly vintage stuff available. To install the packages from CRAN, use

      sudo -E apt-key adv --keyserver keyserver.ubuntu.com --recv-keys E084DAB9

    Add deb http://cran.r-project.org/bin/linux/ubuntu precise/ to /etc/apt/sources.list, and execute

      sudo -E apt-get update 
  sudo -E apt-get install r-base r-base-dev

    Codeface has been tested to work with R 2.15.x and 3.0.x.

  • Make sure that the following distribution packages are installed (the list is for sufficient for a pristine default installation of Ubuntu 12.04 Desktop):

       # Generic packages 
   sudo apt-get install python-mysqldb sinntp texlive default-jdk \
                        mysql-common mysql-client mysql-server python-dev \
                        exuberant-ctags nodejs npm git subversion \
                        libgles2-mesa python-pip sloccount graphviz 

   # Devel packages required to build the R packages below from source 
   sudo apt-get install libxml2-dev libcurl4-openssl-dev xorg-dev \
                        libx11-dev libgles2-mesa-dev libglu1-mesa-dev \
                        libmysqlclient-dev libcairo2-dev libxt-dev \
                        libcairo2-dev libmysqlclient-dev 

   # Devel packages required for Ubuntu 14.04 
   sudo apt-get install libpoppler-dev libpoppler-glib-dev 

   # Devel packages required for python packages 
   sudo apt-get install libyaml-dev

  • When using the feature or feature_file analysis you need to have a working "cppstats" in your path. One way to get it is:

      cd ~
  git clone https://github.com/clhunsen/cppstats.git

    Now create a file like ~/scripts/cppstats and add ~/scripts to your PATH. It should have something along the lines of:

      #!/bin/bash 

  cd ~/cppstats 
  PYTHONPATH="~/cppstats/lib" ~/cppstats/cppstats.py "$@"

    Note that the script has to be executable:

      chmod +x ~/scripts/cppstats

    and then add ~/scripts to your PATH. (maybe you have to replace ~ with the full path (/home/$user) if it doesn't work).

    You can test this script by running "~/scripts/cppstats --help" and validate that you get an help message

Preparing the R installation

  • Run sudo R CMD javareconf; make sure that the tool reports success in finding a java version and compiling programs with the native interface.

  • Install RGraphviz from bioconductor. In an R shell, execute

      source("http://bioconductor.org/biocLite.R") 
  biocLite(c("BiRewire", "graph", "Rgraphviz"))

  • Install the required R packages in an R shell with

      install.packages(c("statnet", "ggplot2", "tm", "tm.plugin.mail", "optparse",
                     "igraph", "zoo", "xts", "lubridate", "xtable",
                     "reshape", "wordnet", "stringr", "yaml", "plyr",
                     "scales", "gridExtra", "scales", "RMySQL",
                     "RCurl", "mgcv", "shiny", "dtw", "httpuv", "devtools",
                     "corrgram", "logging", "png", "rjson", "lsa", "ineq",
                     "arules", "data.table", "RJSONIO"), dependencies=T)

    If necessary, make sure before the installation that /usr/local/lib/R/site-library/ is writeable by the current user so that the packages are made available system-wide.

  • Some packages for R need to be installed from github resp. r-forge:

      devtools::install_github("shiny-gridster", "wch")

  • Currently, the development versions of tm.plugin.mail and snatm need to be installed. In a R session, use

      install.packages(c("snatm", "tm-plugin-mail"),
                   repos="http://R-Forge.R-project.org")

    Should an installable package be unavailable on R-Forge (which can happen from time to time), clone the source manually

      svn checkout svn://r-forge.r-project.org/svnroot/tm-plugin-mail/
  svn checkout svn://r-forge.r-project.org/svnroot/snatm

    and install each package with cd pkg; R CMD INSTALL ..

Installing Python packages

  • Install the required python packages using pip:

      sudo -E pip install pyyaml progressbar python-ctags

Clone the git repository

  • Create a base directory $BASEDIR for the software.

  • Clone the Codeface repository into $BASEDIR with

      git clone https://github.com/siemens/codeface

    which results in the directory $BASEDIR/codeface (referred to as $CFDIR in the following)

Database Setup

NOTE: Updating the database schema after analyses have been performed will naturally delete all existing data stored in the schema.

  • Create a database user codeface with sufficient privileges to create and modify tables: Start mysql-workbench and connect to the database.

    • Select Management->Users and Privileges
    • Click "Add Account", and create a new user (you may want to limit the connectivity to localhost). Click apply.
    • Select tab "Schema Privileges", click "Add Entry", and click "Select ALL". Click Apply.
    • Select tab "Administrative Roles", and select DBDesigner and DBManager. Click Apply.

  • For a fresh setup, install the database schema from $CFDIR/datamodel/codeface.mwb respectively $CFDIR/datamodel/codeface_schema.sql:

      mysql -ucodeface -pcodeface < codeface_schema.sql

Build the Bug extractor

See bugextractor/INSTALL for all java-related details.

Prerequisites for the ID service

  • Ubuntu 12.04 comes with a fairly outdated node.js release. Make sure to install a more recent one (binary packages are provided at http://nodejs.org/dist/), and pick a stable release (even version numbers, 0.10.x at the time of writing).

  • Debian-based distros like Ubuntu name the node.js binary differently than upstream. To fix this, run

      sudo update-alternatives --install /usr/bin/node node /usr/bin/nodejs 99

  • The ID service requires a few node.js packages. Install them by running

      npm install addressparser express js-yaml mysql body-parser

    in the id_service directory.