Skip to main content

The ATLAS Metadata Interface

Project description

For more information see the full pyAMI documentation

pyAMI 4 is a major upgrade of the AMI python client. The command line syntax has been rationalized, with improved help functions. pyAMI is available in the ATLAS software release and also installed centrally on lxplus. pyAMI can be installed standalone on a laptop. Instructions are given below for linux and windows.

All commands are simply subcommands of ami.

List AOD datasets matching a pattern:

ami list datasets --type AOD data11_7TeV

list files in a dataset:

ami list files -lhc

list latest NTUP_TAUMEDIUM data datasets (and number of events in each one) originating from AOD in periods L1 and L2 that are contained in a good runs list (GRL):

ami list data --project data11_7TeV --type NTUP_TAUMEDIUM --parent-type AOD \
--periods L1,L2 --stream physics_JetTauEtmiss --latest --fields events \
--grl mygrl.xml

list runs in a period or multiple periods:

ami list runs B M

display dataset metadata:

ami dataset info

and query projects, data types, dataset provenance, etc. pyAMI also provides an API allowing you to perform all of the same queries from within your own Python program.


pyAMI is installed centrally on LXPLUS at CERN. To begin using pyAMI simply:

source /afs/

then authenticate yourself if you haven’t already (see below).

pyAMI is also available through CVMFS:

export ATLAS_LOCAL_ROOT_BASE=/cvmfs/

Note : You may need to get a particular version of pyAMI - SLC6 requires:

source /afs/


pyAMI requires at least Python 2.4. lxml is optional but required for XSL transformations. libxml2 and libxslt (dev packages) must be installed to build lxml.

Automatic Installation

Automatically install the latest version of pyAMI with pip:

pip install --user pyAMI

or with easy_install:

easy_install --user pyAMI

Omit the --user for a system-wide installation (requires root privileges). Add ${HOME}/.local/bin to your ${PATH} if using --user and if it is not there already (put this in your .bashrc):

export PATH=${HOME}/.local/bin${PATH:+:$PATH}

To upgrade an existing installation use the -U option in the pip or easy_install commands above.

Manual Installation

Get the latest tarball on PyPI

Untar and install (replace X appropriately):

tar -zxvf pyAMI-X.tar.gz
cd pyAMI-X

pyAMI uses distribute to install but you may revert to a basic distutils install by setting the environment variable:


One advantage of using distribute is that all dependencies are automatically downloaded and installed for you. To install pyAMI into your home directory if using at least Python 2.6:

python install --user

or with older Python versions:

python install --prefix=~/.local

Add ${HOME}/.local/bin to your ${PATH} if it is not there already (put this in your .bashrc):

export PATH=${HOME}/.local/bin${PATH:+:$PATH}

If you are unable to satisfy the dependency on lxml (only used for XSL transformations) then you may disable this dependency before installation with:

export PYAMI_NO_LXML=1

Installation on MAC

Tested with System: OXS 10.7.4, Python 2.7.1, GCC 4.2.1

Just follow the instructions for Manual Installation, except that to change the PATH you should do:

export PATH=/Users/yourName/Library/Python/2.7/bin${PATH:+:$PATH}

Installation on Windows

  1. Install python, for example Python 2.7.3

  2. Install distribute (for easy installation of dependences). For 64 bit machines there may be problems, see

  3. Download pip. For example pip-1.1.tar.gz.

  4. Install pip. Open a “cmd” terminal in windows. Assuming that you installed pip below Python:

    cd C:\Python27\pip-1.1
    C:\Python27\python install
  5. Install pyAMI. pip.exe should be in C:\Python27\Scripts:

    cd C:\Python27\Scripts
    pip install pyAMI
  6. You might need to install lxml to enable XSLT in a different step - sometimes pip seems to prefer a version which will not build. If you don’t install lxml, pyAMI will still work, but you will only be able to obtain XML output on the command line. However if you only want to use the API of pyAMI you may skip this step.:

    easy_install,* lxml==2.2.8
  7. Now so that you can use pyAMI conveniently you must adjust your paths in the Windows environment. If you do not know how to do this follow the instructions here Add to the path:

    PATH    C:\PYTHON27;C:\PYTHON27\Scripts
  8. Lastly explain to Windows that a python script can be executed. Add to the PATHEXT variable:

  9. Then change the name of the file ami in C:\Python27\Scripts to so that script can be executed just by typing ami.


VOMS authentication is supported:

voms-proxy-init -voms atlas

The alternative option (if gLite is not available) is to send your username and password along with each AMI command. You can do this explicitly, or you can set up an encrypted configuration file by running:

ami auth

This will prompt you for your AMI username and password. You will only need to do this once since your credentials are stored in ~/.pyami/ami.cfg for later use. If your credentials change just run ami auth again. If you would like to keep your pyAMI configuration in a directory other than ~/.pyami then set the environment variable PYAMI_CONFIG_DIR.

Version 4.1.2 (26/2/2013)

  • Upgrade to distribute 0.6.34

  • Fix installation on Windows

  • Docs now build properly with buildout

  • Install setup_pyAMI with the pyAMI bundle for backward compatibility

Version 4.1.1 (13/2/2013)

  • Upgrade to buildout 2.0.0

  • Fix a few import issues in pyAMI.extern.ZSI

Version 4.1.0 (4/2/2013)

  • A patched version of ZSI is now included in pyAMI.extern that removes dependency on PyXML.

  • argparse is now included in pyAMI.extern.

  • A minimal pyAMI installation (no lxml) now does not depend on any external non-standard library packages.

  • Added the -F and –from-file to the “ami list datasets” command where the pattern is interpreted as a file name which will be read for patterns (one per line, ignoring blank lines or lines beginning with #).

  • Added the option to perform literal matches with “ami list datasets” using the new optional flags -L or –literal-match. In this case the pattern is interpreted as a literal string and the results must match exactly. A literal match query is much faster than a patterned match and can greatly speed up commands where you want to query information for known elements. This option pairs nicely with the new –from-file feature mentioned above for files listing many complete dataset names that should be interpreted literally and not as patterns.

  • AMI errors are now printed correctly.

Version 4.0.4 (6/11/2012)

  • Fix globbing issue. Only surround pattern with % if the pattern didn’t have % on both ends originally. (it needed more fixing!)

Version 4.0.3 (11/10/2012)

  • Skip version 4.0.2 to sync with ATLAS SVN tag version

  • Fix for py2.4 compatibility

  • Make ami-status=VALID default

  • Fix globbing issue. Only surround pattern with % if the pattern didn’t have % on both ends originally.

Version 4.0.1 (20/9/2012)

  • This is pyAMI-04-00-01 in the ATLAS release

  • Minor syntax changes in the API

  • Added Windows installation to the doc

  • Expanded doc section on API

Version 4.0.0a3 (17/4/2012)

  • Make distribute the default install method

  • Include patched urllib2 module in pyAMI.backports

Version 4.0.0a2 (10/4/2012)

  • Second prerelease of the new pyAMI

  • Support for connection through a proxy (BNL)

  • Other minor improvements

Version 4.0.0a1 (28/3/2012)

  • Prerelease of the new pyAMI

  • Package renamed to pyAMI

  • Speed improvements

  • VOMS authentication

  • Fixed permissions of ~/.pyami/ami.cfg

  • Removed intermediate ami package Everything is now under pyAMI

atlasmeta versions below:

Version 0.4.0 (4/2/2012)

  • Now compatible with Python 2.4

  • Install with distutils by default and make installation with distribute optional

  • Improved parent-type feature in dataset query

  • Ability to specify a good runs list (GRL) in ‘ami list data’ and only the runs contained in the GRL are shown

  • Installed on LXPLUS at CERN

  • XSL transformations reimplemented

Version 0.3.1 (10/1/2012)

  • Fix bug in file listing where size/events is NULL in database

Version 0.3 (20/12/2011)

  • Add ability to query datasets by parent type with –parent-type option

  • list files in dataset with ‘ami list files’ and optionally display total size and events

  • new schema module replicating AMI schema

  • query files, projects, subprojects, datatypes, subdatatypes, tags, nomenclatures, production steps

Version 0.2 (17/12/2011)

  • Query and display additional dataset fields (i.e. datatype, number of events)

Version 0.1 (1/12/2011)

  • Forked pyAMI

  • Major code rewrite

  • better command-line interface

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

pyAMI-4.1.3.tar.gz (235.5 kB view hashes)

Uploaded source

Supported by

AWS AWS Cloud computing Datadog Datadog Monitoring Facebook / Instagram Facebook / Instagram PSF Sponsor Fastly Fastly CDN Google Google Object Storage and Download Analytics Huawei Huawei PSF Sponsor Microsoft Microsoft PSF Sponsor NVIDIA NVIDIA PSF Sponsor Pingdom Pingdom Monitoring Salesforce Salesforce PSF Sponsor Sentry Sentry Error logging StatusPage StatusPage Status page