Fix duplicate iteration execution tree ids

[vicksiyi]

Summary

• group iteration execution children by the concrete parent node instance instead of only parentNodeId
• generate virtual iteration tree ids from the parent uniqueNodeId to avoid duplicate RichTreeView ids when the same iteration node executes more than once
• apply the same fix to execution details and chat message execution cards

Tests

• pnpm --dir packages/ui build

Notes

• Fixes MUI X: The Tree View component requires all items to have a unique id property. #5248

[gemini-code-assist]

Code Review

This pull request refactors the iteration node grouping and tree-building logic in both ExecutionDetails.jsx and AgentExecutedDataCard.jsx to use unique parent node IDs (parentUniqueNodeId) instead of simple parent IDs. This allows for direct parent node lookups via a map rather than performing linear scans in the second pass. The feedback suggests further optimizing the first pass by replacing the backward linear scan for finding the parent's unique ID with an $O(1)$ map lookup of the most recently seen indices, which prevents potential UI performance issues on large execution trees.

Important

The consumer version of Gemini Code Assist on GitHub is being sunset. Starting June 18, 2026, new organization installations will be blocked, and all code review activity will officially cease on July 17, 2026.
For more details on the timeline and next steps, please review the Help Documentation.

[gemini-code-assist]

Instead of performing a backward linear scan ($O(N)$) for every iteration child node to find its parent node's unique ID, we can maintain a map of the most recently seen index for each nodeId as we iterate. This optimizes the lookup to $O(1)$ time complexity, preventing potential UI performance degradation when rendering large execution trees.

        const iterationGroups = new Map() // parent uniqueNodeId -> Map of iterationIndex -> nodes
        const lastSeenIndices = new Map()

        // Group iteration child nodes by their parent and iteration index
        nodes.forEach((node, index) => {
            lastSeenIndices.set(node.nodeId, index)
            if (node.data?.parentNodeId && node.data?.iterationIndex !== undefined) {
                const parentId = node.data.parentNodeId
                const iterationIndex = node.data.iterationIndex
                const parentIndex = lastSeenIndices.get(parentId)
                const parentUniqueNodeId = parentIndex !== undefined ? parentId + '_' + parentIndex : null

                if (!parentUniqueNodeId) return

[vicksiyi]

Updated in 9728aa1 to keep a last-seen parent index map during traversal, which removes the per-child backward scan while preserving the unique node id format.

[gemini-code-assist]

Instead of performing a backward linear scan ($O(N)$) for every iteration child node to find its parent node's unique ID, we can maintain a map of the most recently seen index for each nodeId as we iterate. This optimizes the lookup to $O(1)$ time complexity, preventing potential UI performance degradation when rendering large execution trees.

        const iterationGroups = new Map() // parent uniqueNodeId -> Map of iterationIndex -> nodes
        const lastSeenIndices = new Map()

        // Group iteration child nodes by their parent and iteration index
        nodes.forEach((node, index) => {
            lastSeenIndices.set(node.nodeId, index)
            if (node.data?.parentNodeId && node.data?.iterationIndex !== undefined) {
                const parentId = node.data.parentNodeId
                const iterationIndex = node.data.iterationIndex
                const parentIndex = lastSeenIndices.get(parentId)
                const parentUniqueNodeId = parentIndex !== undefined ? parentId + '_' + parentIndex : null

                if (!parentUniqueNodeId) return

[vicksiyi]

Updated in 9728aa1 to keep a last-seen parent index map during traversal, which removes the per-child backward scan while preserving the unique node id format.