Skip to content

Note where the enhanced query drops references (#760) - #764

Merged
adamjohnwright merged 3 commits into
mainfrom
enhanced-v1-note-3
Oct 30, 2026
Merged

adamjohnwright merged 3 commits into
mainfrom
enhanced-v1-note-3

Conversation

@adamjohnwright

Copy link
Copy Markdown
Contributor

This adds comments only, so the new site doesn't copy the fault behind #760. It replaces #761 and #763 after their reviews.

The fault is the default form of both endpoints. Asked without includeRef=true, /ContentService/data/query/enhanced/{id} and /enhanced/v2/{id} both write an object in full only once, and later as its bare dbId. An event's neighbours come before its literatureReference and the lists after it: its preceding and following events, and, since graph-core bb5e0ca0 (Aug 2025), its parent pathways. So a reference a neighbour also cites is a bare number in the event's list, and production's pathway browser drops those numbers (#760).

Measured on beta for R-HSA-3858489:

Request The event's literatureReference
plain v2 [351112, {…}, 350934]
v2 with includeRef=true [{@ref}, {…}, {@ref}]
v1 with includeRef=true [{@ref}, {…}, {@ref}]

The note goes on the node endpoint and the site's Event model. It says to ask with includeRef=true and read through the site's JSOGDeserializer, as the details pages do.

Changed from #761 and #763:

It is the includeRef parameter.

The two site readers still at risk (the IDG panel and the entity popup) are #762. Adding includeRef=true and the deserializer may be the smallest fix there.

No behaviour changes.

Refs #760, #762.

🤖 Generated with Claude Code

adamjohnwright and others added 3 commits October 30, 2026 18:26
/data/query/enhanced keeps Java's answer byte for byte, including a bare
dbId for an object written earlier; since graph-core bb5e0ca0 an event's
parent pathways come first, so references its parent also cites are bare
numbers in its own list. Production's browser loses them (#760). The
endpoint and the event model now say to read an event's lists from
/enhanced/v2, as the details pages do, so the new site does not copy the
fault.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…eferences too

v2 writes each object once as well, with @id, and points back with
@ref; it is safe only through the site's JSOGDeserializer, which the
note now says. Neighbouring events, not only parent pathways, come
before an event's lists in v1. The two v1 readers still at risk are
issue #762.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
… not the version

Without includeRef=true both /enhanced and /enhanced/v2 write bare
dbIds; with it either answers JSOG @id/@ref, which JSOGDeserializer
resolves (measured on beta: plain v2's literatureReference holds 351112
and 350934 as numbers). The neighbours come before literatureReference
and the lists after it, not every list.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
@adamjohnwright
adamjohnwright merged commit f3c0455 into main Oct 30, 2026
8 checks passed
@adamjohnwright
adamjohnwright deleted the enhanced-v1-note-3 branch October 30, 2026 18:48
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant