Skip to content

PDL Tools is a library of reusable tools used and developed by the Pivotal Data Science and Data Engineering teams.

Notifications You must be signed in to change notification settings

seethb/PDLTools

 
 

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

PDL Tools

PDL Tools is a library of reusable tools used and developed by the Pivotal Data Science and Data Engineering teams.

Build Status

Usage docs

http://pivotalsoftware.github.io/PDLTools/

Binaries (Pivotal internal)

PDLTools binaries

Pre-requisites

The following are the pre-requisites for building PDLTools:

Required:

  • Pivotal Greenplum or Apache HAWQ (GPDB sandbox, HAWQ sandbox)
  • Apache MADlib (Download)
  • cmake (3.5 recommended)
  • GNU C and C++ compilers (gcc, g++)
  • Flex (>= 2.5.33)
  • Bison (>= 2.4)
  • rpmbuild

Optional:

  • Doxygen (1.8.7 recommended, if generating HTML docs),
  • LaTeX (if generating PDF docs)

Building

For CentOS or Red Hat Enterprise Linux, install the pre-requisite tools:

 sudo yum install cmake gcc gcc-c++ flex bison rpm-build

From either the Greenplum or HAWQ master node, follow these steps as the gpadmin user:

 curl -L -o pdltools-1.7.zip https://github.com/pivotalsoftware/PDLTools/archive/v1.7.zip
 unzip  pdltools-1.7.zip 
 cd PDLTools-1.7
 source /usr/local/hawq/greenplum_path.sh
 mkdir build
 cd build
 cmake .. -DRPM_INSTALL_PREFIX=$GPHOME
 curl -L -o third_party/downloads/uriparser-0.7.9.tar.bz2 https://sourceforge.net/projects/uriparser/files/uriparser-0.7.9.tar.bz2
 curl -L -o third_party/downloads/cpptest-1.1.2.tar.gz https://sourceforge.net/projects/cpptest/files/cpptest-1.1.2.tar.gz
 make 2> /dev/null

Generating Doxygen User Docs

You can generate Doxygen docs for the project as follows:

 make doc

This will create the user docs under $BUILD/doc/user/html. You can also generate a PDF of the user doc by running

 cd build/doc/user/latex && make pdf

This will generate a PDF titled refman.pdf in $PDLTOOLS/build/doc/user/latex

Packaging

To create an rpm package which you can ship for installation into other machines, run the following (from the build directory):

 make package

When installing on a cluster, it is best to create a gppkg installer. Run the following (from the build directory):

 make gppkg

Installation

Installation is a two-step process. First, you will have to install MADlib and PDL Tools on either the Greenplum or HAWQ master node. To do this, you will run the following:

 gppkg -i <pdltools gppkg file>

For example, run the following (from the build directory):

 gppkg -i deploy/gppkg/2.0/pdltools-1.7-hawq2.0-rhel5-x86_64.gppkg

This will place all the relevant binaries & SQL files at the appropriate location (usually $GPHOME/pdltools). Next, you will have to install the SQL UDFs in the target database.

To install pdltools into a database of your choice, run the following (consider adding pdlpack in your PATH):

 $GPHOME/pdltools/bin/pdlpack install [-s <schema name>] [-S <SUgAR schema name>] [-M <MADlib schema name>] -c <username>@<hostname>:<port>/<database name>

For example:

 $GPHOME/pdltools/bin/pdlpack install -s pdltools -c gpadmin@mdw:5432/testdb

The default schemas are pdltools for the main schema, sugarlib for SUgAR and madlib to search for MADlib objects.

Running Install Check Tests

Post installation, you can run the unit tests in PDL Tools with the install-check command like so:

 $GPHOME/pdltools/bin/pdlpack install-check -s pdltools -c gpadmin@mdw:5432/testdb

Parameters for install-check are the same as parameters for install. If any of the tests fail, you will see an error message displayed on your console.

Contributing to PDLTools

If you're interested in contributing to PDLTools, please refer to the instructions at Guidelines for contributing to PDLTools

Legal

Copyright (c) 2013-2016 Pivotal Software, Inc. All rights reserved.

Unauthorized use, copying or distribution of this source code via any medium is strictly prohibited without the express written consent of Pivotal Software, Inc.

THE SOFTWARE IS PROVIDED “AS IS”, WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.

About

PDL Tools is a library of reusable tools used and developed by the Pivotal Data Science and Data Engineering teams.

Resources

Stars

Watchers

Forks

Packages

No packages published

Languages

  • Python 54.2%
  • C 19.8%
  • CMake 14.0%
  • Jupyter Notebook 10.9%
  • Shell 0.8%
  • C++ 0.2%
  • HTML 0.1%