You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
This is a hard one. Dataset names contain both slashes and dashes, both of which act as special characters at various places in the pMp pipeline. Commit 805743b moves query parameters into the URL's parameter string to avoid the issue of extra slashes within the path, and e269668 extends the suggestions API to allow some completion of dataset names, but this falls flat as soon as you type a slash or a dash.
Additionally, searching ElasticSearch by output dataset with a simple URL query returns really ambiguous results and so warrants a full query as JSON in the request body. However, dashes need to be escaped in ElasticSearch and \- is an invalid character escape sequence in JSON, causing it to throw an error before even getting to the search. A potential solution is to re-index requests in ElasticSearch so that the output_dataset field is not_analyzed - this should stop the ElasticSearch parser from processing the dashes in dataset names when requests are indexed.
I don't have nearly enough time left now to explore these options, so hopefully this will be useful to whoever takes this up next :)
emerged on discussion
2015/12/10
https://twiki.cern.ch/twiki/bin/view/CMS/PreMCCM
users would like to be able to:
Needs dataset name be added to the elastic indexing
The text was updated successfully, but these errors were encountered: