Develop - #665
Merged
Merged
Conversation
Register a 'catalogue' presentation variation on the AI Chatbot block
(@eeacms/volto-eea-chatbot), following the cross-addon variation push
pattern used by the TabsBlock 'spotlight' variation.
Rough first iteration of the presentation:
- reuses the chatbot core (useChatController, markdown components,
AutoResizeTextarea, EmptyState)
- streamed answer rendered inline, no Sources tab
- cited documents rendered below the answer as document cards in the
style of the Navigator catalogue cards (Onyx-provided metadata only
for now; Plone REST enrichment with full catalogue metadata comes
in a follow-up)
Block fields are read from top-level props (ChatBlockView renders
<Presentation persona=... {...blockData} />), matching the classic
ChatWindow contract.
…nt cards
Fixes:
- Block fields (onyxVersion, qgenAsistantId, enableQgen, placeholder,
starter prompts, ...) are read from top-level props, matching the
<Presentation persona=... {...blockData} /> contract of the classic
ChatWindow. Previously a non-existent 'data' prop was read, silently
dropping onyxVersion and falling back to the v2 API.
- Assistant messages now render through the chatbot core's
RendererComponent (as AIMessage does), so the answer streams with the
typewriter reveal, blinking cursor and citations; a BlinkingDot shows
while the assistant is working (search-tool phase). Message keys
switched to the stable nodeId.
Features (document cards):
- Inline ![[doc: Title]] markers in the answer text are converted by a
remark plugin (docCards.js) into cca-doc-card elements, rendered as
document cards via extraMarkdownComponents (new pass-through props on
the chatbot's MessageTextRenderer/RendererComponent).
- DocumentCard extracted to its own module; the 'Relevant documents'
grid and the inline cards both use EnhancedDocCard, which upgrades
from the basic Onyx-fields card to the full Navigator catalogue card
once the document is fetched.
- useCatalogueDoc: fetches a document from the globalsearch ES index by
URL (_id) via searchlib's fetchResult (GET /_es/globalsearch/_doc/{id}
passthrough, whitelisted in the volto-searchlib middleware), appConfig
from the searchlib registry (settings.searchlib.searchui), with a
module-level cache per URL.
- Extract Navigator catalogue card CSS into a shared mixin and apply it in the chat (.catalogue-chat .catalogue-chat-navigator-card), not only in the searchlib results page — the card was rendering unstyled in chat. - docCards: match ![[doc: ...]] markers in any text node (own line, list items, mid-sentence), not only whole paragraphs. - Hide the "Relevant documents" grid (inline cards replace it; kept for easy restore) and drop the temporary RAW MESSAGE debug log.
Restore the full classic chatbot functionality (Thinking/tools accordion, styled New-chat button, streaming placeholder, quality check, related questions, feedback) by reusing the classic ChatWindow instead of the hand-rolled view. The catalogue variation now only: - passes hideSourcesTab (classic Sources tab/sidebar/inline list hidden; the inline catalogue cards are the document UI), and - injects the ![[doc: ...]] remark plugin + a stable cca-doc-card component (CcaDocCard) that reads the owning message from ChatMessageContext and renders the inline Navigator-style catalogue cards. CatalogueChatView is now a thin wrapper; removed the now-unused view subcomponents and their CSS. Doc-card CSS in navigator.less no longer needs the .catalogue-chat parent.
The core MessageTextRenderer re-renders the whole markdown tree on every streaming tick, which was re-rendering every (potentially heavy) catalogue card tens of times a second. - Memoize DocumentCard/EnhancedDocCard with a field-based comparator (the inline card rebuilds its `source` object each render, so plain `===` would never memoize). Cards now only re-render when their own data changes; the one-time basic->full upgrade is still driven by internal state. - useCatalogueDoc: keep a resolved-value cache and initialize/short-circuit from it, so an already-known doc renders the full card immediately instead of flashing the basic card first (repeat doc, remount, same URL twice).
The `.catalogue-compare` label (checkbox + "Compare" text) is only made into a centered inline-flex inside `.searchapp-navigatorCatalogueSearch`. The chat catalogue card lives under `.catalogue-chat-navigator-card`, not that ancestor, so the label fell back to default inline flow and the Semantic UI checkbox's default spacing pushed the "Compare" text out of line with the box. Extend the selector to cover the chat card wrapper.
…talogue-variation
Import the chat module namespace instead of the ChatMessageContext named export so the app still compiles when an environment resolves a pre-4.1.0 @eeacms/volto-eea-chatbot; CcaDocCard then falls back to the basic metadata card (no document matching).
Covers the remark doc-marker plugin, the document cards (basic, enhanced, inline, context-based CcaDocCard), the catalogue ES lookup hook (fetch, cache, error paths) and the variation registration / ChatWindow wrapper, restoring the SonarQube coverage gate after the new code landed.
- Use Navigator catalogue card structure and components in base DocumentCard - Add cleanDocumentTitle and resilient matchesDocumentTitle for pipeline titles - Add normalizeCatalogueUrl in useCatalogueDoc for Elasticsearch URL matching - Compact inline card bottom margin in catalogue-chat.less - Add unit tests for title cleaning, view button, and URL normalization
feat: Chatbot 'catalogue' presentation variation (cards & suggested next steps)
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.
No description provided.