Repository navigation
Redesign Query Insights - #1049
Draft
Tomasz Naumowicz (tnaum-ms) wants to merge 49 commits into
Draft
Tomasz Naumowicz (tnaum-ms) wants to merge 49 commits into
Tomasz Naumowicz (tnaum-ms) wants to merge 49 commits into
Conversation
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
…I-4) Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
…urce (WI-5) Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
…(WI-10) Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
…ta (WI-11) Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Rebuild Welcome/Configure/Analyze on the LocalQuickStart patterns: header with Feature preview, copyable explain snippets and privacy note, live non-throwing input inspection with line/column errors, sample plans, summary-row editors for optional context, and an Analyze summary bar with inline model change, Edit inputs and Re-run. The reused panel no longer nests its own scroll area inside the wizard page. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
The full sentence stays in the tooltip; the long value clipped in the right column at common widths. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
… the explorer Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Fluent tokens and chevron icon instead of raw SCSS, locale-formatted costs and token counts, compact identity line, visible edge in high contrast. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
…itle - Wizard model section uses the Optional context label/value table; no collapsible; missing data reads 'Not reported for this model'. - Default model: saved choice, else first Powerful (API category), else first. - Analyze step titled 'Query Insights analysis results', no transient subtitle. - New configure-models L2 scenario. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
…dges Same CounterBadge list as LocalQuickStart; Required/Optional badges sit next to each entry title instead of acting as bullets. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
…options (Phase F) Five options for the AI Performance Insights card's model selection (model pill, split button, settings strip, native quick pick, header chip popover) plus the current layout, rendered with mock models in the L2 scenario queryInsightsStandalone/tab-model-options. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
…e' label
- Compact ModelPicker becomes a VS Code chat-style model pill with a
grouped menu (cost badge per model).
- Card: primary action + pill, one muted meta line (prompt, Explore,
Learn more), existing cost footnote.
- The utility model never shows 'Free'; the careful footnote wording
('No additional cost for most GitHub Copilot subscribers.') is used
in the card and the wizard Cost row.
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
…over Replace the hand-rolled model pill (Menu + Tooltip with an overflowing table) with a Fluent Dropdown grouped by category, cost badges per option and an info Popover holding the model details. The listbox is height-capped and scrolls, and the model list is refetched when the dropdown opens. The wizard uses the same compact setup instead of the details table. Also bold Welcome list titles and narrow the JSON editor gutter. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
…d, matching Command type Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
…cards Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
… identities Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
…o the output channel Tab card (Phase K): - Actions row is model picker -> details -> Get AI Performance Insights. - One muted footer: cost note, accuracy disclaimer, Explore, and "Learn more about AI Performance Insights". - Prompt source removed from all UI and the getPromptSourceInfo RPC; PromptTemplateService traces it to the output channel on every request. Query Insights Explorer (Phase L): - Model picker moves into the Configure footer before Analyze query; the AI model section and "choice is saved" text are removed. - "Explain output" is the single term with one Required marker. - Explain editors auto-size and wrap long lines. - Namespace shown once (Optional context); prefilled explain opens on Configure. - "AI responses may be inaccurate" removed from every AI card header and stated once in the Powered-by byline (and where the user prompts). Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Contributor
❌ CI failed: fast tier (draft PR)Commit
Full run: add the Code quality
Build sizeNot measured: draft PR; runs once the PR is marked ready. Updated in place on every run. |
resources/prompts/local/query-insights.prompt.md is gitignored; when it exists (and no prompt-path setting is configured) it replaces the built-in index advisor prompt for find, aggregate and count. Prompt source: local-resource-file. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
…w plan reachable - Normalize shards/inputStages before @mongodb-js/explain-plan-helper, which fails with "x is not iterable" on non-array values; guard the same loops in StagePropertyExtractor. - Log the Stage 2 analysis error, stack and execution plan to the output channel. - Keep "View Raw Execution Stats" available after a Stage 2 failure. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Recommendations with action "none" (query change, verify, maintenance, data model, support) render as an advisory InsightCard: kind title and icon, optional copyable command, risks only when provided, no execute button. The kind comes from a new optional "kind" field, falling back to a "Kind:" prefix in the justification. Adds the analyze-insights L2 scenario. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
When an index only provides order (for example _id_ with a
streamingPrimaryKey cursor and every predicate in runtimeFilterSet),
DocumentDB reports `indexUsage: {}` instead of an array. The low-cardinality
index advisory iterated it and Stage 2 failed with "indexUsage is not
iterable". Both indexUsage readers now require an array, as do scanKeys.
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
- Shared CopyButton: the icon becomes a checkmark for two seconds and the result is announced, like the explorer's snippet buttons. - Index cards offer Copy command next to the primary action. - Rename the "Verify First" advice card to "Verify". Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
A "Copy as Markdown" action next to the Optimization Opportunities title exports every card in display order, then the model, prompt source, usage and the exact data messages sent to the model (user query and context, never the instruction prompt). The Stage 3 complete event now carries these modelInputs from the AI service. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
…lection scans A batched find on a non-indexed field runs as a streaming cursor that walks _id_ to resume each batch (cursorScanType: streamingPrimaryKey). The explain shows IXSCAN on _id_ with no indexFilterSet and every condition in runtimeFilterSet, so the analysis reported "Index used: _id_" for what is a full collection scan. Add a second, DocumentDB-only analysis pass (documentDbPlanRefinements.ts) that runs after the engine-neutral explain-plan-helper extraction and before the rating. It reports these plans as a collection scan and adds a neutral "Collection read in _id order" badge. It keys off DocumentDB-only fields, so MongoDB plans and real _id lookups are unchanged. Applied corrections are exposed as planRefinements and recorded in Stage 2 telemetry. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Replace "Copy as Markdown" with a Share menu: - Copy as Markdown: the full report, including the model inputs. - Copy as rich text: email-safe HTML (inline styles, tables, light palette, shape markers instead of color-only meaning) plus a plain Markdown alternative without the model inputs, written through the webview clipboard. Falls back to plain text through the host when rich text cannot be written. AI Markdown is rendered with micromark (already bundled through react-markdown; now a direct dependency at the locked 4.0.2), which escapes raw HTML and drops unsafe links; only http(s) links are kept. Copy feedback is shared through a useCopyFeedback hook, and the report uses the same visible badges as the panel (visibleDiagnostics). Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
…for both copy formats "Copy as rich text" now renders the same Markdown report with inline styles instead of a separate HTML layout. The header box, rating pill, progress bar and recommendations table did not paste well into Teams and Word; the report is now headings, paragraphs, lists and code blocks only. The Markdown gains a "Query Efficiency Analysis" section (rating as shape and word, metrics as a list, the visible badges), so both formats carry the same content, including the model inputs. Removes buildOptimizationShareHtml and the includeModelInputs option. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Tomasz Naumowicz (tnaum-ms)
added
ci:full-requested
This branch has not been deployed
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
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.
Redesigns the Query Insights experience with a streamlined analysis flow, shared model selection, richer model and cost information, and explorer entry points.
This draft is intended to share the updated Query Insights design and implementation for early feedback.