Answers draw more diagram types - #191
Merged
Merged
Conversation
Owner
Author
|
Flow-corpus eval, BA audience, two runs of ten (same setup as docs/measurements/2026-09-17-ba-diagrams.md):
Reference on #188: 7/10 and 6/9, rubric 27/30 and 23/26. All 16 fences pass the renderer's parser (9 flowchart, 7 sequence). Four of them are in the corpus now. |
An answer's diagram is a mermaid fence chosen by the shape of the content: a flowchart for branches, a sequence for exchanges, a state machine, an entity model, or a trigger to target mapping drawn as two subgraphs. The two hand-written shapes forced a mapping into a flowchart with an invented order and a target drawn twice. The picture no longer cites per node; the sentence that introduces it carries the markers. That removes the reason for a hand-written layout, so the renderer, the chip placement and the src renumbering go. The renumberer treats the fence as code. Stored answers with the older JSON fence still draw through the existing conversion. The renderer runs with securityLevel strict, no bindFunctions, the theme read from the app's tokens, the classic look without shadows. Corpus tests now assert the fence leaves the backend byte for byte and passes the renderer's own parser.
The renderer's viewBox starts above and left of the origin, so a ground rect at 0,0 left a bare band along two edges of the downloaded file. A render that failed is no longer cached: the first diagram of a type loads its chunk, and a chunk missing after a redeploy would otherwise follow the source for the rest of the session.
trick77
force-pushed
the
feat/more-diagram-types
branch
from
September 17, 2026 18:36
9551bad to
66aefef
Compare
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
An answer's diagram is now chosen by the shape of the content: a flowchart for branches, a sequence for exchanges, a state machine, an entity model, or a trigger to target mapping drawn as two subgraphs. With only two hand-drawn shapes on offer, a mapping question (which event produces to which Kafka topic) came back as a flowchart with an invented order between the topics and one topic drawn twice.
The picture no longer cites per node; the sentence that introduces it carries the markers. That removes the reason for the hand-written layout, so the renderer, chip placement and src renumbering go. The renumberer treats the fence as code. Stored answers with the older JSON fence still draw through the existing conversion.
Rendering runs with strict sanitising, no click bindings, colours from the app's tokens, flat look. Corpus tests assert the fence leaves the backend byte for byte and passes the renderer's parser; a Playwright pass covered a stored legacy thread, a mapping answer, the share page, the full view, the SVG download and the project wiring graph.
Flow-corpus eval (BA audience) is running; its numbers follow in a comment.