Downloading AO2D from LEGO train conversion

If you need a single input files from a conversion train, please take them from the train test, by:

  • Going to the train Run3_Conversion

  • Click on the train number of the train you are interested in then open the test results by clicking the + button

  • Scroll to "full train" and click on "output"

  • Download the AO2D.root

  • You can also click on processing progress and then follow the output directories to download a larger AO2D.root. All this is only suitable for manual downloading of a few files. For everything else see the next item:

Downloading AO2D from the Grid

You can use the either JAliEn Cache Manager (Hyperloop downloading tool) or just alien.py cp (or alien_cp) commands of xjalienfs package. To get this up on your local box, you need a few steps:

  • Install xjalienfs with aliBuild:
aliBuild install xjalienfs

Use the JAliEn Cache Manager

  • Checkout the cache manager from git:
git clone ssh://git@gitlab.cern.ch:7999/jalien/jalien-cache-manager.git
  • Add the target directory jalien-cache-manager to your path or give the absolute path below.
  • Install python dependencies:
pip install click humanfriendly cachier
  • Create a directory for the files:
export CACHE_DIR=$HOME/jalien-cache
mkdir $CACHE_DIR
  • Enter the environment:
alienv enter xjalienfs/latest-release,ROOT/latest

The download itself happens with the following command:

jcachemgr --cache-dir $CACHE_DIR pull --out wn.xml --parallel=8 --limit=50 /alice/data/2021/OCT/505637/apass1/AOD AO2D.root

which would download at maximum 50 files of type AO2D.root from /alice/data/2021/OCT/505637/apass1/AOD

When the command terminates it creates a wn.xml which contains the list of files. This is the input to a Grid job. If you want to get a text file list from this to use as input for O2, the following line does it for you:

sed -rn 's/.*turl="([^"]*)".*/\1/p' wn.xml > input.txt

Use the xjalienfs package commands (alienpy in pypi repository)

General description of file transfer capabilities of alien.py can be found in JAliEn documentation pages

General form of command: alien.py cp args source destination where local targets MUST have the designator file: and the lack of any designation imply remote (GRID) target

Brief list of options

Destination will be automatically taken as directory (and automatically created) if the source is directory or a file selection. The source allow inside globbing so the source can have the form of /alice/data/2021/OCT/505637/apass1/AOD*AO2D.root The other selection options are:

  • -glob STRING : form like -glob AO2D.root /alice/data/2021/OCT/505637/apass1/AOD file:local_destination
  • -select PATTERN : regex based selection of files in source directory (applied to full path)
  • -name PATTERN : regex based selection of files in source directory (applied only to file name)

Force checking of destination: -f When destination exists the file is skipped from re-downloading (it will change to always check the md5sum) With -f the md5 will be checked (it will change to become default)

Nr Parallel downloads (defaults to 8 for downloads) : -T Nr

Retry Nr times the replica list for a failing LFN : -retry Nr

Keep Nr components of the full path for the destination file : -parent Nr e.g. alien.py cp -parent 99 /alice/data/2021/OCT/505637/apass1/AOD/001/AO2D.root file:/my_cache_dir will download the file to /my_cache_dir/alice/data/2021/OCT/505637/apass1/AOD/001/

{min, max}{Size, depth, time} selection is possible (see the linked manual)

For writing xml collections see toXml command (and the help content)