Loading func-difs-soda-vs-vlkb.tex 0 → 100644 +78 −0 Original line number Diff line number Diff line Conclusion to have one image for SKA and VLKB: * separate 1. & 2. to its own module * implement 3 & 4 also for pure-SODA * delete 5. 6. exposes endpoints */soda */availability */capabilities */cutouts [optional, only if 4. application/xml supported] Q: * Does 2 withoput 1 make sense ?? ________________________________________________ on top of soda functionality vlkb-cutout has 1. IA2 token validation -- env: (see auth.properties for params) ENABLE in config by: token-validation endpoint defined 2. authorization -- env: DB-table (AUTH_DB_*) columns: policy(PUBLIC|PRIVATE), groups ENABLE in config by: DB-table URI defined 3. */cutouts temp dir of cuts ** internal feature: if defined streaming is happening from the cache ** is pre-requisite for other features -- volume: /*/cutouts -- env: access-url to cache (and service type: download, bbcp,..) -- env: time-to-live ENABLE: 4. links instead of streaming by RESPONSEFORMAT=application/xml -- config in download_url=http://host:port/path/<file> -- special value: download_url=THISWEBAPP ** cutouts-dirs (and/or surveys-dirs) exported on the same host/tomcat instance ** no need for separate service ** this is tomcat config: **** context.xml for dir mappings and default webapp in web.xml & default-webapp config: listings-allowed **** only web.xml and context.xml are part of the app: src/main/webapp/WEB-INF/web.xml and META-INF/context.xml ENABLE in config: define URL download service (THISWEBAPP, http://, bbcp?) 5. add/modif keywords in cut-files given in metadata metadata: for search: adds info about surveys not present in FITS-headers for cutouts: may add keywords to cut FITS headers NOTE if keywords added at ingestion, metadata is needed only in search -- env: METADATA (/*/surveys/survey_populate.csv) TEMPORARILY instead of direct streaming: --> check: cfitsio Extended Syntax with output: fitsfile(outfile[modkey CUNIT3=km/s])[23:234 600:800 1:100] ** 1, create fitsfile in cache (cutouts e.g. use 4.) ** 2, add missing keywords to FITSheader ** 3, stream the modif'd file down ENABLE in config: define metadata file pathname 6. mcutout -- accept doPost with vlva-JDL in body -- produce tar.gz and reponse forms as for fits (stream tar.gz or put to cutouts/* cuts-cache) ENABLE: [no need; if suported, will read body of POST] 7. count of undef values in IMAGE_HDU -- runtime option webapi param: ?? Optional: 7. resolver by mapping -- env: DB-table (RESOLVER_DB_*) : obs_publisher_did, local pathname 8. AMQP connection API-engine (vlkbd) -- env: host, queue-name, amqp-user/password Loading
func-difs-soda-vs-vlkb.tex 0 → 100644 +78 −0 Original line number Diff line number Diff line Conclusion to have one image for SKA and VLKB: * separate 1. & 2. to its own module * implement 3 & 4 also for pure-SODA * delete 5. 6. exposes endpoints */soda */availability */capabilities */cutouts [optional, only if 4. application/xml supported] Q: * Does 2 withoput 1 make sense ?? ________________________________________________ on top of soda functionality vlkb-cutout has 1. IA2 token validation -- env: (see auth.properties for params) ENABLE in config by: token-validation endpoint defined 2. authorization -- env: DB-table (AUTH_DB_*) columns: policy(PUBLIC|PRIVATE), groups ENABLE in config by: DB-table URI defined 3. */cutouts temp dir of cuts ** internal feature: if defined streaming is happening from the cache ** is pre-requisite for other features -- volume: /*/cutouts -- env: access-url to cache (and service type: download, bbcp,..) -- env: time-to-live ENABLE: 4. links instead of streaming by RESPONSEFORMAT=application/xml -- config in download_url=http://host:port/path/<file> -- special value: download_url=THISWEBAPP ** cutouts-dirs (and/or surveys-dirs) exported on the same host/tomcat instance ** no need for separate service ** this is tomcat config: **** context.xml for dir mappings and default webapp in web.xml & default-webapp config: listings-allowed **** only web.xml and context.xml are part of the app: src/main/webapp/WEB-INF/web.xml and META-INF/context.xml ENABLE in config: define URL download service (THISWEBAPP, http://, bbcp?) 5. add/modif keywords in cut-files given in metadata metadata: for search: adds info about surveys not present in FITS-headers for cutouts: may add keywords to cut FITS headers NOTE if keywords added at ingestion, metadata is needed only in search -- env: METADATA (/*/surveys/survey_populate.csv) TEMPORARILY instead of direct streaming: --> check: cfitsio Extended Syntax with output: fitsfile(outfile[modkey CUNIT3=km/s])[23:234 600:800 1:100] ** 1, create fitsfile in cache (cutouts e.g. use 4.) ** 2, add missing keywords to FITSheader ** 3, stream the modif'd file down ENABLE in config: define metadata file pathname 6. mcutout -- accept doPost with vlva-JDL in body -- produce tar.gz and reponse forms as for fits (stream tar.gz or put to cutouts/* cuts-cache) ENABLE: [no need; if suported, will read body of POST] 7. count of undef values in IMAGE_HDU -- runtime option webapi param: ?? Optional: 7. resolver by mapping -- env: DB-table (RESOLVER_DB_*) : obs_publisher_did, local pathname 8. AMQP connection API-engine (vlkbd) -- env: host, queue-name, amqp-user/password