From Grenouille to Polar Bears Managing metadata and data capture for the Astra-Gemini 0.5 PW laser

1
From Grenouille toPolar Bears Managing metadata
and data capture for the Astra-Gemini 0.5 PW laser

• All Hands Meeting, Edinburgh, 8-11 Sep 2008
• Laurent Lerusse, ESC
• Victoria Marshall, CICT
• Edwin Divall, CLF

2

• What is STFC?
• Our Vision of Data Management
• The Central Laser Facility
• CLF Architecture Overview
• Polar Bear System
• Elk the Electronic Log booK
• Whats next?

3
What is STFC?

• Science and Technology Facilities Council
• Funds researchers in universities.
• Provides scientific and technical expertise.
• Provides access to world-class facilities
  overseas CERN, ESA,ESO, ESRF, ILL, telescopes in
  Chile, Hawaii, La Palma and Australia.
• Manages world-class facilities in UK ISIS, the
  Central Laser Facility, HPCx and the Diamond
  Light Source.

4
Integrated e-Infrastructure
Data and Metadata Capture is automated.
Proposal
Publication
5
Integration via Metadata
Data
Data
Experiment
Metadata
Analysis
Metadata
Metadata
Simulation
Data
6
Central Laser Facility from dual beam Petawatt
to commercial size lasers

• Multiple lasers VULCAN, ASTRA, Laser For Science
• New laser Astra-Gemini - October 2007
• Targets area are rebuilt with new diagnostics
  for each experiment.

http//www.clf.rl.ac.uk/
7
Astra-Gemini

• Dual beam high intensity pulsed laser
• 0.5 PW or 15 J in 30 fs
• Repetition rate of 3 shots per minute
• Adaptive optics for high beam quality
• This upgrade represents an improvement of a
  factor
• 10 in intensity
• 100 in repetition rate

8
CLF Data Flow

• 15 types of data
• All interconnected

• At the moment, we are concentrating our effort on
  the proposals and laser data.

9
Top Level Architecture
eCAT Visualisation Analysis
PolarBear Metadata
ELk Manual
Laser representation
eCATDB Latest
DAQ Automatic
NeXus Writer
CLF SRB Archive
10
Polar Bear System

• Initial aim
• Gather information on the laser components and
  the data they produce in an XML format which
  could be integrated with the data.

11
Polar Bear System

• Instruments
• 33 different types in 2 categories
• Catalogueable (identifiable)
• Generic
• Need manual input to fill the details
• Serial no, focal length, model, etc
• Examples
• Grenouille (From Swamp Optics)
• Polarizer (aka Polar Bear)
• Telescope

12
Polar Bear System

• XML description of the laser.
• Laser defined using
• - Assemblages
• - Channels
• - DataStreams
• Meta-information within the structure
• Instruments
• Shot type
• Data collection type
• Next component

13
Polar Bear System

• Orthogonal classification/view of the laser
• Three-level hierarchy to locate each Channel
• - By room,
• - By table,
• - By the type of data.
• Simpler view for everyday use and the user
  interface.
• Includes the Channel and DataStream names.
• The data will contain both structure and
  description with all the metadata.

13
14
Electronic Lab booK (ELk)

• Initial aim
• Ruggedised WiFi PDA.
• Allows the operator to log non-automatic
  information from the laser shop floor.
• Safety goggles and reduced lighting made the
  interface challenging!
• Uses the PolarBear metadata to present a
  customised interface for manual readings and
  incidents.
• Integrates seamlessly with the rest of the DAQ
  system.

15
Electronic Lab booK (ELk)

• Parallel PDA desktop version, used to
• Present Daily instructions on how to switch
  the laser on.
• Record instrument configuration changes. (eg.
  Diode blows up.)
• Record environmental factors.
• Upload pictures, screen dumps and comments from
  dedicated h/ware.
• Stores the data in a database and uploads
  XML/NeXus files.

15
16
What did we learn?

• Defining a complex system such as Astra-Gemini in
  a machine readable format (XML) is not easy.
• Filling metadata is hard to set up, but pays
  dividends later.
• Culture shock between the laser
  scientists/operators and the computer scientists.
  (You want me to hand-edit XML??)
• Evolution, not revolution, for a new approach to
  data management/analysis.

16
17
Technical details

• DAQ LabView and Visual Basic
• neXusWriter C and some Python
• PolarBear ASP and JavaScript
• ELk Java, Ajax and JavaScript
• eCAT Java
• Data storage Oracle and SRB

18
Whats next?

• VULCAN the other powerful laser
• Target areas
• Need much more flexibility (new experiment every
  6 weeks),
• User friendliness (users change with each
  experiment),
• Robustness and (lots of new users ),
• Direct rewards (benefits for the user must be
  obvious and direct),
• Data Rights.