2022.2.11_6

[strogonoff]

Datatracker OAuth

This release introduces working Datatracker OAuth2 integration:

• Log in with Datatracker is now shown, if client ID and secret are configured.
• API now accepts not just the Datatracker development API token, but also recognizes OAuth credentials in user’s session. If access token is valid and not expired (this requires a trip to Datatracker), the function is called.
• XML & JSON download links are now displayed again for users logged in via Datatracker, making use of API recognizing OAuth session. (GUI download links for BibXML, JSON and potential other serialization formats in future #113)

Notable limitations

• The name of the logged-in user is not displayed correctly (Shape of user data returned from Datatracker’s OIDC endpoint #116)
• Access token refresh is not implemented (Datatracker OAuth2: refresh access tokens #117)
• If Datatracker OAuth is expired at the time of JSON or XML download, the downloaded file will contain an error message (If OAuth access token is expired at the time of XML or JSON download, the downloaded file contains an error message #118)

xml2rfc path handling

xml2rfc compatibility was extended to support manual mappings. Mappings dictate which docid.id a given xml2rfc-style path must correspond to. Mappings are stored in DB.

Note: existence of a mapping does not mean the path is always resolved; if a document with mapped docid.id does not exist in the index, the mapping will fail and “auto” or “fallback” method will be used.

Management GUI now includes an xml2rfc compatibility dashboard (https://dev.bibxml.org/management/xml2rfc-compat/). When a directory is selected, it shows a list of XML files inside. If there are more than 6000 files, user agent must have JavaScript turned on, and data windowing is used in that case (GUI becomes unusable with many thousands of DOM entries).

• Clicking a file allows to create/edit/remove its manual mapping, and show if path cannot resolve.
• For each file visible within the scrollable view, the effective path resolution method is dynamically loaded and displayed (this requires JavaScript). It can be one of the following, in order of preference:

  • method	behavior
    manual	if manual map was specified, try to find an item with mapped docid.id
    auto	if manual map does not exist or failed to resolve, use an automatic fetcher function that tries to construct a JSON query against Relaton bibliographic data based on XML filename in the path
    fallback	if neither of the above worked, proxy to data indexed from xml2rfc tools mirror Git repository
    N/A	if the fallback failed as well (e.g. the xml2rfc mirror source was not indexed or an error occurred during request)

• For a selected file it’s possible (with JavaScript enabled) to test resolution and view the resulting XML and resolution methods.
  • It allows to compare the XML against currently existing xml2rfc tools XML, but that JavaScript request currently fails due to cross-origin restrictions.

Other

GUI was updated in places. Bibliographic item detail view is now a bit more compact with some copy changes, and more properties (authors, publishers) now link to structured searches in addition to relations. Datatracker login block is more prominent.

Some bugs were fixed, including API for resolving DOI.

---

This discussion was created from the release 2022.2.11_6.