Skip to content
This repository has been archived by the owner on Feb 27, 2021. It is now read-only.

Archive

Jump to bottom
Stefan Beck edited this page Mar 10, 2020 · 26 revisions

Below is list of our meetings hold and what we discussed.

2020-03-10, 10 am CET

  • Technical scope of eduGAIN (SAML, Shibboleth etc.)
  • Design decisions for Shibboleth
    • We use one entityID for production and sandbox, since they run on hte same apache
  • OpenIng and its repositories
    • TUprints done
    • Braunschweig and Stuttgart develop on their repos to ingest the deposit from dissemin
  • Article in code4lib
    • Stefan will start a draft
    • Content should be: Historical part, how we merge papers, evaluation of deposits, description how deposits work
  • Remove OaiRecords with source ResearchGate
    • Should be discussed with all the team
  • Improve current design
    • Define a font instead of using standard font
    • Ask for precise changes, maybe Stefan can implement them

2019-12-13

  • Review about collaboration
  • Future of developer meetings
    • If there is need, we do them
    • Try with once a month
  • Localisation
    • Documentation on localisation cause of Vinaigrette
  • Speed of start page
    • Try with better index
    • Currently caching entire page
  • CrossRef / Citeproc
    • Add CrossRef batch for big ORCID profiles
  • BASE import has bug concerning Publication types
  • eduGAIN
    • No news

2019-11-26

  • Bootstrap4 on Sandbox
    • Finished, some testing would be nice, to fix some bugs
      • Some fixes
  • Production
    • Issue with UploadedPDF object in Admin interface on DepositRecord (HTTP 503)
      • Fix by not fetching all UploadedPDFs
  • Publicationtypes
    • CrossRef almost OK
    • dx.doi.org not really OK
    • Fix import by DOI
    • How long does doi-cache.dissem.in cache?
  • DataCite has API and can do bulks

2019-11-12

  • Update CrossRef (look for #777)
  • Update BASE (Extend PR #760)
  • TUPrints on production not before 2019-11-18 (also, new letter of declaration required by TUDA)
  • eduGAIN: No response since last time, ask them again

2019-10-29

  • Bootstrap 4
    • Centralize all urls in dissemin/urls.py
    • Rely entirely on JS for search as we require JS anyway for ORCID and the deposition
  • ReseachGate
    • Status is OA, since its available
    • Might get cleaned at some point in time, but no decision today

2019-10-15

  • Short review of OA Days at Hannover
    • Lot of interest in Dissemin (due to advertisement), roughly 50 persons
  • eduGAIN
    • No response yet, Antonin sends email
  • Migration to Bootstrap 4
    • Bootstrap 3 is rather old and the new version has a lot of advantages
    • Opportunity to make pages more "bootstrappy", i.e. rely more on container, rows and cols, use less "hacks" for mobile devices, would reduce maintenance
    • Design currently broken
    • Stefan Beck tries to migrate to Bootstrap 4
  • Adding new repository
    • Some standard steps are required
    • Organisation of process would be nice (checklist)
  • SHERPA/RoMEO API v2
  • Paper ingestion
    • Wasn't programmed for scale
    • PRs are in principial ready to accelerate import

2019-09-10

  • Feedback from Designer
    • We can more or less do what we want
    • Proposes one big graphic, including everything
    • Likes to tell a story
    • We followed some of his ideas
  • Implementation
    • Fonts
    • Partners
    • Graphics

2019-08-29 (extra meeting)

  • Decide for colours on new start page
    • Fine
  • Decide for Icons on new start page
    • We choose some ideas from our pool.

2019-08-20

  • Feedback on start page design
    • Less colours
    • No margin
    • Change padding
    • Last draft: https://framapic.org/S5aoewThJIiB/DQ2bA9nvfyFb.png
    • Done by Lucas - Discuss in next meeting colours and decide which icons to use
    • Try to get it done by 2019-09-29, so that it can be shown on OA Days Hannover
  • eduGAIN
    • Stefan writes letter for Antonin who applies on UK federation
    • Technical installation will be done by Stefan
    • Antonin moves sandbox.dissem.in to another server, so that it does not use same apache as production
  • German translations & OA Tage Hannover
    • Some translations are going on before Hannover, making deploying necessary more often
  • Dissemin blog text about OpenIng
    • Stefan writes a draft

2019-07-23

  • Library accounts
    • Intended for the usage of institutions, that upload on behalf of their researchers
    • Use djangos custom login, as these do not have ORCID, or get a shibboleth account from their it service departement
  • Todo list
    • Save a list of publications to upload them later
    • Considered as not so useful (most user upload right now, list of own publications exists + oa filter, motivates procrastination of upload, feature to be explained to the user, integration into UI)
    • Concept could be: Divide users into researchers and support staff (e.g. secretary), the last group has a todo list available
    • Discussion of concept is in OpenIng
  • Start page
    • Antonin ask some colleagues concerning colours
    • Ideas:
      • Use maybe different colour (including change of dissemin-blue)
      • Maybe less colours
      • First jumbotron with search field should be catchy
      • Put piechart with legend in light grey box with rounded corners for better contrast, as we currently use more or less every colour
      • Bigger piechart
    • Colour chooser tool: GPick. Has also some fivetone and fourtone models, not just sixtone

2019-07-02

  • Presenting ideas and progress on new DB-Layout
  • Start page redesign (PR #779)

2019-06-11

  • Design for start page, see PR #736
    • Will be delayed, because implementation of protocol got priority
    • Issues with Bootstrap 3: Vertical alignment in columns not supported and solutions break mobile support
    • TUDA has fotopool and contract with professional icon designer
    • Migrate to Bootstrap 4
  • Model for metadata -> we need a more flexible system to configure repository instances (determine which field is mandatory, for instance)
  • Change Codacy criteria -> anyone is free to tweak the settings to improve the Codacy feedback

2019-05-28

2019-05-14

  • Pin Issue #602

    • Pinned together with #730

  • Logging in RepositoryProtocol class - How does it work and relate to current logging? Change logging system here?

    • Current logging stores the log in the DepositRecord afterwards to make it easy to spot issues afterwards. Thus, we not print log messages to a separate file like /var/log/zenodo.log (for Zenodo deposits), because this requires access to server.
    • Failed deposits should go into Sentry because these are errors.
    • DepositRecord shall be shown an paper object.
    • PR PR #762

  • Licenses

    • Currently licences are bound to protocols, they should be related to repositories
    • Model for licenses?
    • Most licenses do not really need a translation
    • Move translation to Database and provide import django for initial database population from cvs
    • Maybe use django-modeltranslation for Repository-Descriptions as well #364
    • Issue #761

  • Test for Deposit - What to do? There are changes on the road.

    • Zenodo is not activated because it requires a Zenodo API key (not mocked). Goal would be to mock API code and test deposit interfaces with that. Probably better to have many small fixtures instead of this huge fixture loading a lot of data into database.
    • New test should use their own fixtures. This makes tests more clear and is more robust to changes

  • PR #694 - Status

    • Issue #506
    • Issue #730
    • Quite high priority, lot of complains about this. This would provide a fix for Papers objects showing with an ORCID badge but not appearing in a Researcher's profile. Antonin is on it :)

  • Import from BASE seems not to work properly?!

    • At the moment, the main issue is the speed at which we import papers (about 30-50 papers / seconds, quite slow). Ongoing work with cha on how to speed up things, by working on the database ingestion. We currently make a lot of round trips between Python code and the database. Hoping that the first big speed up will come from checking in batched papers in the database.
    • We use the dump from BASE because we are quite far behind. The idea afterwards is to use the API. The import from the API broke at some point and we are now massively behind.
    • We now have a sandbox at sandbox.dissem.in with similar hardware and setup as we have in production, so we can try things easily with it.

  • Status of the imports / update on ORCID dump import etc

    • BASE dump in progress (see above), ORCID dump, Crossref, Unpaywall (imported).
    • ORCID should be imported after Crossref is completely finished. ORCID dump only uses ORCID metadata which are often incomplete (authors in particular).
    • DOI is currently completely ignored in ORCID dump ingestion. Papers are only merged based on merging heuristics and available metadata.

2019-04-30

  • Feedback of OpenIng regarding hosting of Dissemin
  • Issue #577 - More information on / advertisement of Dissemin and Open Access
    • Items to add: list of supported open repositories, more details on green OA, how many papers have been deposited, success stories.
    • Let's write the content and see the design from there!
    • Deadline for content: 2019-05-11
  • PR #701 - What to do with OA Button
    • Invert "Data provided by SHERPA / ROMEO" line and "Contact authors" line to make two blocks.
  • For OpenIng we like to create additional metadata to make the moderation process on the institutional repositories easier.
    • Creating XSD, the target namespace should be https://dissem.in/terms/ or similar
    • XSD should be part of the documentation
    • Scheme to be presented in the documentation for repository administrators
    • Principle: METS allows to add extra metadata in a bundle. That extra metadata is normally ignored by the target repository but it is still stored and it would be useful to use that mechanism to add some Dissemin internal metadata there, that can be used later on for maintenance. A schema for that can be added in our docs, and we can use something like "https://dissem.in/terms/" as a namespace.
  • Programming for deposition in institutional repositories.
    • Directly merge changes into master branch
    • Layer-Structure, reuse existing structure
    • requests with mocking
    • Integration into documentation for developers and repo administrators
    • revamping the tests for the deposit protocols will require mocking HTTP requests and that could be a bit tricky for HAL since it does not use the "requests" library but "httplib", but we can refactor the depositing code to mock methods instead.
Clone this wiki locally