Some ideas on workflow using holding pen from DESY point of view.
Refers to
GUI mockup arXiv,
GUI mockup journals,
flow chart arXiv and
slides from January
The big picture (with priorities):
Some text along with the
GUI mockup arXiv
Sorry - I know this is a bit terse, I hope it is not too confusing. Don't hesitate to complain.
Title (1)
buttons to select / core (1)
pre-set depending on count of CORE stuff (2)
can be set and inactive for articles we always take (2)
algorithm more complex - ask florian - or let him do it
input fields for FC (3), TC (4), CNUM (4) pre-populated from metadata (in longer term via algorithm)
recid (6) (leads to merge via holdingpen)
search buttons (6) (for search query ask Annette)
number of references that are extracted in a useful way (2-3)
(would be great for arxiv, but currently we don't have it)
how many are CORE | in INSPIRE-HEP | total number
number of CORE Keywords (1-2) (if 2 show some dummy output) and list
depending on source of article / data available:
from fulltext (arXiv, maybe journals) (2)
from title + abstract + author keywords (otherwise) (3)
for the number of CORE KW, particles (to be given via KB - in the very long term from HEPont) don't count
this info will not be part of the ingested record (store separately or delete before ingestion)
number of CORE/total PACS (4) and list of clean text (info from KB)
number of Anti-KW (5) from output of a different ontology
number of CORE paper from these authors (3) (algorithm from florian)
list of keywords (BibClassify from title + abstract + author keywords) (3)
will be part of the ingested record if it is CORE
list of author keywords (3)
list of authors, truncate for long lists (2-3)
list of affiliations, ICN and/or plain text, truncate for long lists (2-3)
abstract, collapsible? (4)
publication-note / report number (1)
arXiv-note (1), DOI (3), number of pages(4), conference info (4)
Priorities:
- next sprint
- needed for serious arXiv processing
- needed for serious journal processing
- very useful for journal / arxiv / ticket processing
- useful for journal / ticket processing
- nice to have
Tasks for sprint:
- display:: title, report-number, pub-note, doi, arXiv-note
- buttons select, core:
some algorithm to determine 'state' for these buttons: not-set, pre-set, set-and-inactive
This is the 'enrich and propose' box in the flow-diagram.
For sprint: run BibClasify on fulltext. CORE-KW >= 2 -> pre-set CORE-button
optional: read list of particles from KB (to be created by Florian), count only CORE-KW if not in this list.
Long-term:
select-button set-and-inactive will be needed for journals
more information based on references, authors, affiliations, ... will be used to determine 'state' of buttons, suggest field-code
- Feel free to display more stuff from big picture
Question:
There are 2 different use-cases for essentially the same UI:
- selection (and CORE-setting) of records in holding pen (e.g. math not cross-listed)
- CORE-setting of records that are automatically harvested (e.g. math cross-listed to gr-qc) i.e. update of records in INSPIRE
Do you think you can implement both for the math test-run or shall we process case 2) at DESY?
Comments to DEV version on inspirevm11.cern.ch
Detailed view:
pull-downs/action on right side
input fields (field-code / type-code / ..) below action buttons
main area:
publisher info: normal font
INSPIRE enriched info: e.g. fixed width
If possible Positive numbers/info in green; Negative in red
(see
https://twiki.cern.ch/twiki/pub/Inspire/InputtingJournalsDESYmockup/HP-mockup-arXiv.pdf)
In general use layout/font/colour to structure information
Tabular view:
is it possible to sort/filter for date-created if it is not displayed in table?
If yes, we don't need it visible.
Title, Identifier , Core Info, (Created?), Actions
Examples:
Identifier:
Nucl.Phys. B123 (2014) 165-172
arXiv:1403.1234 (math.AG)
arXiv:1406.0481 (nuch-th, hep-th)
For journals we don't have categories in the moment, only for arXiv.
So it can be part of the Identifier. main category first, followed by cross-listings
Core Info means just numbers:
CORE-KW | CORE-PACS/all-PACS | CORE-Refs/all-Refs
could look like
5 | 1/3 | 7/19
--
KirstenSachs - 05 Jun 2014