AMI A brief introduction to AMI Solveig Albrand

AMI A brief introduction to AMI Solveig Albrand Jerome Fulachier 3/6/2021 S. A. 1

AMI team • IN 2 P 3 : – Developers: Solveig Albrand, Jerome Fulachier, Fabian Lambert. (LPSC) – Physicists: Fairouz Malek (LPSC) Stephane Jezequel (LAPP) • Univ. Of Glasgow: – Tom Doherty, Paul Millar 3/6/2021 S. A. 2

AMI ATLAS review of AMI • April – July 2006. https: //twiki. cern. ch/twiki/bin/view/Atlas/Review. Of. AMI • “On the basis of the successful completion of this review, we recommend that AMI is deployed as the primary physicist interface to the metadata and Dataset Selection Catalog. ” 3/6/2021 S. A. 3

AMI ATLAS DDM Architecture AMI 3/6/2021 S. A. 4

AMI Where does AMI get information? prod. DB Other systems Push AMI Dataset Search Pull DDM Link to More Information Task Request 3/6/2021 S. A. 5

AMI Real Data? • The report of the Metadata Task Force, has just been released. It provides the basis for the treatment of real data. • We will take part in the various tests going on this year. (Stream Test, T 0 Test, FDR) • We think we have developed a framework which is sufficiently flexible to cope with whatever requirements the task force will produce. • Of course AMI is not the only actor in ATLAS Physics metadata. 3/6/2021 S. A. 6

AMI Metadata holders • ATLAS architecture is distributed. There are several, heterogeneous systems. – Conditions database (run, stream, interval of validity) – Event metadata – Production DB – Dataset / File physics metadata • One of our AMI ambitions is to help physicists find their way around. (for example setting up a Metadata dictionary) 3/6/2021 S. A. 7

AMI Atlas Metadata Interface (AMI) main points • Client-Server architecture. • Backend RDBMS : my. SQL, ORACLE, SQLite • Multi-catalogue support. – Examples: • CTB 04 (datasets, runs and files) • Tile csc (datasets and runs) • Csc Monte_Carlo (datasets and “event_ranges” (datablocks)) • Can search all catalogues in parallel, then refine the search – Web interface – CLI/API using py. AMI or java clients 3/6/2021 S. A. 8

AMI Learning about AMI • The AMI Portal: http: //ami. in 2 p 3. fr • We provide – A Fast Track Tutorial (about 1 hour) – A Full Tutorial (about 2. 5 hours) – The next collective tutorials is: • Monday 19 th March (During T&P week) 3/6/2021 S. A. 9

AMI Main Features • Catalogue Overview • Quick Search, Keyword search, Advanced Search. • Ordering of results. Choice of fields to display. Display of available values. Information about fields. • Getting details (Job Options, Tasks). Provenance. • Bookmarking. (Point to your favourite datasets) https: //twiki. cern. ch/twiki/bin/view/Atlas/BPhysics. Working. Group. Csc. Tasks • Powerful graphic interface for refining a search result. • Python client available for CLI use. 3/6/2021 S. A. 10

AMI Extra Slides • Client server model • Screen shots – in case of network problems… 3/6/2021 S. A. 11

AMI Web Service Same Commands, Same XML, different XLST Client Side AMI WS Client HTTP(S) (SOAP) Web Container : TOMCAT AMI Web Service Core AMI Web Interface HTTP(S) AMI framework About 100 WS commands are available 3/6/2021 DB S. A. 12

AMI Web Search 3/6/2021 S. A. 13

AMI Web Search 3/6/2021 S. A. 14

3/6/2021 S. A. 15

3/6/2021 S. A. 16

AMI Refine your search – with a generic query builder 3/6/2021 S. A. 17

AMI The most popular csc dataset is csc 11. 007003. singlepart_e_ Et 25. evgen. EVNT. v 110004 01 Dataset provenance tracking Just click on a dataset name! 3/6/2021 S. A. 18