feat(seer): collapse each response into one ThinkingBlock - #121972
Draft
natemoo-re wants to merge 3 commits into
Draft
feat(seer): collapse each response into one ThinkingBlock#121972natemoo-re wants to merge 3 commits into
natemoo-re wants to merge 3 commits into
Conversation
github-actions
Bot
added
the
Scope: Frontend
Contributor
Story previewsPreview the stories changed in this PR on the Vercel deployment:
Preview deployment: https://sentry-cmmvclial.sentry.dev |
Contributor
📊 Type Coverage Diff
🔍 6 new type safety issues introducedNon-null assertions (
This is informational only and does not block the PR. |
The Explorer rendered the flat block list 1:1, and thinking was drawn once per tool_use block, so a single assistant turn showed a wall of separate 'Thinking' and tool-call rows. Group the transcript into responses (a user block, then the run of assistant/tool_use blocks that follows) and render each response as one top-level ThinkingBlock: reasoning, intermediate narration, and every tool call interleaved in run order inside it, with the final answer hoisted out as a sibling below. Thinking prose stays gated on the showThinking toggle; a direct answer with no reasoning renders without a ThinkingBlock. Reuses the shared ToolCall wiring (now exported ToolCallList) for the nested tool rows.
Title the per-response ThinkingBlock with the most recent thing the agent did (the current tool while streaming) instead of a static "Thinking", so the block summarizes itself and ThinkingBlock's decode animation replays as each step lands. Prefers Code Mode call-record labels, falls back to a classic tool label (skipping Code Mode's own non-descriptive tool names), and falls back to "Thinking" before any tool runs. Never reads thinking_content, so the title never leaks reasoning when the toggle is off.
natemoo-re
force-pushed
the
nm/seer/tool-group
branch
from
August 13, 2026 19:57
620dcb1 to
05a1963
Compare
…mposed input, failure output
![github-advanced-security[bot]](https://avatars.githubusercontent.com/in/57789?s=60&v=4){kind=small}
| // Wrap a value that would otherwise re-tokenize wrong (spaces, quotes, parens) so the assembled | ||
| // query parses back to the same filter. | ||
| function quoteValue(value: string): string { | ||
| return /[\s"()]/.test(value) ? `"${value.replace(/"/g, '\\"')}"` : value; |
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
2 participants
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.
The Seer Explorer renders the flat block list 1:1, and thinking is drawn once per
tool_useblock — so a single assistant turn shows a wall of separate "Thinking" and tool-call rows (see the many stacked rows in the current UI).This groups the transcript into responses and renders each as a single top-level
ThinkingBlock:groupTranscript): auserblock is its own segment; every maximal run ofassistant/tool_useblocks after it is one response.ResponseGroup): reasoning, intermediate narration, and every tool call are interleaved in run order inside the block. The block is active/expanded while streaming and collapses (with elapsed time) once done — matchingThinkingBlock's existing behavior.assistantcontent renders as a sibling below the ThinkingBlock (its reasoning still lives inside). A direct answer with no reasoning renders with no ThinkingBlock at all.showThinkingtoggle; tool calls always render.Nested tool rows reuse the shared
ToolCallwiring from the stack (ToolCallListis now exported).