feat(ai-gemini): upgrade @google/genai to v2 + migrate Interactions API#781
Open
jherr wants to merge 13 commits into
Open
feat(ai-gemini): upgrade @google/genai to v2 + migrate Interactions API#781jherr wants to merge 13 commits into
jherr wants to merge 13 commits into
Conversation
… API Bump @google/genai from ^1.43.0 to ^2.8.0. The v2 SDK replaces the legacy flat Interactions schema with the new step-based model, so migrate the experimental text-interactions adapter to the renamed SSE events (interaction.created / step.start / step.delta / step.stop / interaction.completed) and the Step data model, switch structured output to the polymorphic response_format, and drop the deprecated response_mime_type. Update the 25 text-interactions tests to the new event/step schema and set allowBuilds['@google/genai'] to false (its install scripts are no-ops for consumers). Co-authored-by: Cursor <cursoragent@cursor.com>
# Conflicts: # pnpm-workspace.yaml
… 2.x The @google/genai v2 migration moved geminiTextInteractions onto the SDK 2.x Interactions event protocol (step.start/step.delta/step.stop, interaction.created/completed). All published @copilotkit/aimock versions (≤1.31.0) still emit the SDK 1.x shapes (content.* / interaction.start / interaction.complete), so the e2e mock and adapter no longer agree on the wire format and the assistant message comes back empty. Mark the test test.fixme (skipped, not failing) and document the exact event shapes aimock must emit to re-enable it. SDK 2.x adapter behaviour stays covered by unit tests in packages/ai-gemini/tests/text-interactions-adapter.test.ts. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Contributor
📝 WalkthroughWalkthroughUpgrades ChangesGemini SDK 2.x Migration
Sequence Diagram(s)sequenceDiagram
participant GeminiSDK as Gemini SDK 2.x Stream
participant Adapter as GeminiTextInteractionsAdapter
participant AGUI as AG-UI Consumer
GeminiSDK->>Adapter: interaction.created {interactionId}
Adapter->>AGUI: RUN_STARTED
GeminiSDK->>Adapter: step.start {type: model_output}
Adapter->>AGUI: TEXT_MESSAGE_START
GeminiSDK->>Adapter: step.delta {text_delta}
Adapter->>AGUI: TEXT_MESSAGE_CONTENT
GeminiSDK->>Adapter: step.start {type: function_call, index, name}
Adapter->>AGUI: TOOL_CALL_START
loop arguments_delta fragments
GeminiSDK->>Adapter: step.delta {arguments_delta: string fragment}
Note over Adapter: Append to buffer, run parsePartialToolArguments
Adapter->>AGUI: TOOL_CALL_ARGS {args: accumulated buffer}
end
GeminiSDK->>Adapter: step.stop
Adapter->>AGUI: TOOL_CALL_END
GeminiSDK->>Adapter: interaction.completed
Adapter->>AGUI: RUN_FINISHED
Estimated code review effort🎯 4 (Complex) | ⏱️ ~60 minutes Possibly related issues
Possibly related PRs
Suggested reviewers
Poem
🚥 Pre-merge checks | ✅ 4 | ❌ 1❌ Failed checks (1 warning)
✅ Passed checks (4 passed)
✏️ Tip: You can configure your own custom pre-merge checks in the settings. ✨ Finishing Touches📝 Generate docstrings
🧪 Generate unit tests (beta)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
Contributor
🚀 Changeset Version Preview2 package(s) bumped directly, 0 bumped as dependents. 🟨 Minor bumps
🟩 Patch bumps
|
|
Review the following changes in direct dependencies. Learn more about Socket for GitHub.
|
|
View your CI Pipeline Execution ↗ for commit 9feeca7
☁️ Nx Cloud last updated this comment at |
@tanstack/ai
@tanstack/ai-angular
@tanstack/ai-anthropic
@tanstack/ai-client
@tanstack/ai-code-mode
@tanstack/ai-code-mode-skills
@tanstack/ai-devtools-core
@tanstack/ai-elevenlabs
@tanstack/ai-event-client
@tanstack/ai-fal
@tanstack/ai-gemini
@tanstack/ai-grok
@tanstack/ai-groq
@tanstack/ai-isolate-cloudflare
@tanstack/ai-isolate-node
@tanstack/ai-isolate-quickjs
@tanstack/ai-mcp
@tanstack/ai-ollama
@tanstack/ai-openai
@tanstack/ai-openrouter
@tanstack/ai-preact
@tanstack/ai-react
@tanstack/ai-react-ui
@tanstack/ai-solid
@tanstack/ai-solid-ui
@tanstack/ai-svelte
@tanstack/ai-utils
@tanstack/ai-vue
@tanstack/ai-vue-ui
@tanstack/openai-base
@tanstack/preact-ai-devtools
@tanstack/react-ai-devtools
@tanstack/solid-ai-devtools
commit: |
Resolve conflict in ai-gemini text-interactions adapter: the v2 Interactions migration moved function-call start handling into the `step.start` case and streams args via a new `arguments_delta` delta, so main's older `function_call` delta block is obsolete — kept the PR's architecture and dropped the conflicting block. Preserve main's #477 fix in the new architecture: emit `parentMessageId: messageId` on TOOL_CALL_START so a tool-first function call binds to the same assistant message id the eventual TEXT_MESSAGE_START uses. Ported the accompanying test from the old content.* wire shape to the v2 step.* shape. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
Contributor
There was a problem hiding this comment.
Actionable comments posted: 3
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Inline comments:
In @.agentsroom/agents.json:
- Line 8: The claudeSessionId property in agents.json (line 8) should not be
committed to version control, as it contradicts the project's established
pattern where agents-local.json and the sessions/ directory are explicitly
excluded in .gitignore. Remove the claudeSessionId from the committed
agents.json file and instead relocate it to agents-local.json or load it at
runtime from local/environment sources to align with the project's convention
for handling session and personal data.
In `@packages/ai-gemini/src/experimental/text-interactions/adapter.ts`:
- Around line 1306-1325: In the arguments_delta handling block where
safeParseToolArguments is called with the buffer parameter, the assignment
`state.args = parsed` unconditionally overwrites state.args with the result, but
safeParseToolArguments returns an empty object {} on parse failures for
incomplete JSON. To fix this and preserve the last-good args value for
incomplete buffers, only assign the parsed result to state.args when the buffer
actually represents a complete, valid JSON object. Add a condition to check
whether the parsed result is a successful parse before overwriting state.args,
ensuring incomplete fragments don't reset previously accumulated arguments to
{}.
In `@testing/e2e/tests/stateful-interactions.spec.ts`:
- Around line 38-52: The test 'two-turn conversation chained via
previous_interaction_id' is disabled using test.fixme due to a version mismatch
between the current gemini-upgrade adapter (which expects SDK 2.x events like
step.start, step.delta, interaction.created) and `@copilotkit/aimock` (which still
emits SDK 1.x events like content.delta, interaction.start). To fix this, update
the `@copilotkit/aimock` dependency in testing/e2e/package.json to a version that
emits SDK 2.x event shapes as documented in
stateful-interactions.AIMOCK-TODO.md, then change test.fixme back to test on the
disabled test case and verify it passes.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: defaults
Review profile: CHILL
Plan: Pro
Run ID: dcc3f577-6b3a-4a22-a393-f0471786dd1d
⛔ Files ignored due to path filters (1)
pnpm-lock.yamlis excluded by!**/pnpm-lock.yaml
📒 Files selected for processing (11)
.agentsroom/.gitignore.agentsroom/agents.json.agentsroom/prompts.jsonpackages/ai-gemini/package.jsonpackages/ai-gemini/src/experimental/text-interactions/adapter.tspackages/ai-gemini/src/experimental/text-interactions/events.tspackages/ai-gemini/src/experimental/text-interactions/provider-options.tspackages/ai-gemini/tests/text-interactions-adapter.test.tspnpm-workspace.yamltesting/e2e/tests/stateful-interactions.AIMOCK-TODO.mdtesting/e2e/tests/stateful-interactions.spec.ts
💤 Files with no reviewable changes (1)
- packages/ai-gemini/src/experimental/text-interactions/provider-options.ts
Comment on lines
+38
to
+52
| // DISABLED — blocked on aimock. The gemini-upgrade branch migrated | ||
| // `geminiTextInteractions` to the @google/genai v2 (SDK 2.x) Interactions | ||
| // event protocol, which streams `step.start` / `step.delta` / `step.stop` | ||
| // and `interaction.created` / `interaction.completed`. Every published | ||
| // `@copilotkit/aimock` (≤1.31.0) still emits the SDK 1.x shapes | ||
| // (`content.*` / `interaction.start` / `interaction.complete`), so the | ||
| // adapter consumes none of the mock's deltas and the assistant message | ||
| // comes back empty. This is a mock/adapter format mismatch, not an adapter | ||
| // bug — 2.x behaviour is covered by unit tests in | ||
| // packages/ai-gemini/tests/text-interactions-adapter.test.ts. | ||
| // | ||
| // Re-enable (test.fixme → test) once aimock emits SDK 2.x events. The exact | ||
| // required event shapes and migration steps are documented in | ||
| // ./stateful-interactions.AIMOCK-TODO.md | ||
| test.fixme('two-turn conversation chained via previous_interaction_id', async ({ |
Contributor
There was a problem hiding this comment.
📐 Maintainability & Code Quality | 🟠 Major
🧩 Analysis chain
🏁 Script executed:
#!/bin/bash
# Check if there's an open issue or PR tracking the aimock SDK 2.x migration
gh issue list --repo CopilotKit/aimock --search "SDK 2.x OR Interactions 2.x OR Gemini 2.x" --limit 10 --json number,title,state,url
# Also check recent aimock releases to see if 2.x support might already be available
gh release list --repo CopilotKit/aimock --limit 5Repository: TanStack/ai
Length of output: 348
🏁 Script executed:
cat testing/e2e/package.json | grep -A 2 -B 2 aimockRepository: TanStack/ai
Length of output: 289
🏁 Script executed:
find . -name "text-interactions-adapter.test.ts" -type fRepository: TanStack/ai
Length of output: 210
🏁 Script executed:
rg "SDK 2\.x|2\.x.*event" --type ts --type tsx -iRepository: TanStack/ai
Length of output: 177
🏁 Script executed:
rg "SDK 2\.x|2\.x.*event|step\.start|step\.delta" --type ts -iRepository: TanStack/ai
Length of output: 149
🏁 Script executed:
head -50 ./packages/ai-gemini/tests/text-interactions-adapter.test.tsRepository: TanStack/ai
Length of output: 1659
🏁 Script executed:
grep -n "step\.start\|step\.delta\|interaction\.created\|interaction\.completed\|content\." ./packages/ai-gemini/tests/text-interactions-adapter.test.ts | head -20Repository: TanStack/ai
Length of output: 1093
🏁 Script executed:
cat ./testing/e2e/tests/stateful-interactions.AIMOCK-TODO.md 2>/dev/null || echo "TODO file not found"Repository: TanStack/ai
Length of output: 7388
E2E test coverage is disabled, violating mandatory testing guidelines.
This test violates the requirement that "Every feature, bug fix, or behavior change MUST include E2E test coverage using Playwright + aimock." While the comments, comprehensive stateful-interactions.AIMOCK-TODO.md file, and unit tests (25 tests in packages/ai-gemini/tests/text-interactions-adapter.test.ts) provide excellent coverage of the 2.x adapter behavior, the E2E gap remains a blocker for merging.
The root cause is legitimate: @copilotkit/aimock (currently v1.33.0) still emits SDK 1.x event shapes (interaction.start, content.delta, etc.) while the adapter expects SDK 2.x shapes (interaction.created, step.delta, etc.), creating a complete format mismatch that prevents the mock from functioning with the current adapter code.
The migration path for aimock is documented in detail in ./stateful-interactions.AIMOCK-TODO.md (including exact event shape changes for buildInteractionsTextSSEEvents and buildInteractionsToolCallSSEEvents), but the work has not yet shipped. Once aimock releases SDK 2.x event support:
- Update
@copilotkit/aimockintesting/e2e/package.json - Change
test.fixmeback toteston line 52 - Verify the test passes before merging
🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
In `@testing/e2e/tests/stateful-interactions.spec.ts` around lines 38 - 52, The
test 'two-turn conversation chained via previous_interaction_id' is disabled
using test.fixme due to a version mismatch between the current gemini-upgrade
adapter (which expects SDK 2.x events like step.start, step.delta,
interaction.created) and `@copilotkit/aimock` (which still emits SDK 1.x events
like content.delta, interaction.start). To fix this, update the
`@copilotkit/aimock` dependency in testing/e2e/package.json to a version that
emits SDK 2.x event shapes as documented in
stateful-interactions.AIMOCK-TODO.md, then change test.fixme back to test on the
disabled test case and verify it passes.
Source: Coding guidelines
| // be an empty `{}` placeholder when streaming, where the | ||
| // real args arrive as `arguments_delta` events. Treat both | ||
| // uniformly: stash whatever we got, stringify once. | ||
| const initialArgs = step.arguments as Record<string, unknown> |
| @@ -0,0 +1,10 @@ | |||
| [ | |||
Contributor
There was a problem hiding this comment.
This whole folder should be git ingnored, right?
| // parse produced. The wire-level args fragments don't have | ||
| // to be individually-parseable, but the *full* string at | ||
| // step.stop will be. | ||
| const parsed = safeParseToolArguments(buffer, logger) |
Contributor
There was a problem hiding this comment.
This is going to throw for incomplete buffers and cause errors to be logged. As the new api streams function params you'll get error noise here. Maybe switch to partial json?
tombeckenham
requested changes
Jun 23, 2026
tombeckenham
left a comment
tombeckenham
left a comment
Contributor
There was a problem hiding this comment.
Pretty good. Nicely commented. Couple of things to resolve, but I'd say please raise an upstream issue on aimock to support the new gemini API. Alem or one of the team will add support and we can add e2e tests. Add a follow up issue to add e2e tests back for this. Also there's no unit tests for streamed arguments_delta (multi-fragment) and thought/thought_summary reasoning
- Stream function-call args via partial-json instead of strict JSON.parse:
incomplete `arguments_delta` fragments no longer log a parse error each,
and a truncated buffer keeps the last good args instead of resetting to {}
(tombeckenham, CodeRabbit).
- Drop the unnecessary `as Record<string, unknown>` cast on step.arguments;
FunctionCallStep.arguments is already an index-signature object
(tombeckenham).
- Remove accidentally-committed `.agentsroom/` tooling (incl. a personal
claudeSessionId), unrelated to this change (CodeRabbit).
- Add unit tests: multi-fragment arguments_delta accumulation, args-stream
truncation preserving last good args, and thought_summary -> REASONING_*.
- Add changeset for the @google/genai v2 upgrade.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Contributor
Author
|
Thanks for the review @tombeckenham 🙏 — addressed in 9feeca7: Code
Tests (your note on missing coverage)
Changeset: added ( Still open (tracked in the PR's TODO, not in this push): the |
jherr
changed the title
jherr
changed the title
Contributor
There was a problem hiding this comment.
Actionable comments posted: 1
🧹 Nitpick comments (1)
packages/ai-gemini/tests/text-interactions-adapter.test.ts (1)
506-506: 📐 Maintainability & Code Quality | 🔵 Trivial | 💤 Low valueConsider a more precise assertion for the number of TOOL_CALL_ARGS events.
The test emits 3
arguments_deltafragments but only assertstoBeGreaterThan(1). If the adapter emits oneTOOL_CALL_ARGSevent per delta fragment, you could strengthen this totoBe(3)to catch unexpected buffering/coalescing behavior.✨ Optional stronger assertion
- expect(argsEvents.length).toBeGreaterThan(1) + expect(argsEvents.length).toBe(3)🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@packages/ai-gemini/tests/text-interactions-adapter.test.ts` at line 506, The assertion for argsEvents.length in the test is too loose and does not precisely validate the expected behavior. Change the expect statement from toBeGreaterThan(1) to toBe(3) to match the exact number of arguments_delta fragments that are emitted in the test, which will help catch unexpected buffering or coalescing behavior in the adapter where TOOL_CALL_ARGS events might be unexpectedly combined instead of emitted individually per delta fragment.
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Inline comments:
In `@packages/ai-gemini/tests/text-interactions-adapter.test.ts`:
- Around line 599-602: The test fixture is missing a step.start event for index
1 before the step.delta event that emits text content. Add a step.start event
with index 1 and appropriate start event properties before the existing
step.delta event block (which has event_type: 'step.delta' with index: 1) to
establish the SDK 2.x event model sequence of step.start → step.delta →
step.stop as followed by other fixtures in this test file.
---
Nitpick comments:
In `@packages/ai-gemini/tests/text-interactions-adapter.test.ts`:
- Line 506: The assertion for argsEvents.length in the test is too loose and
does not precisely validate the expected behavior. Change the expect statement
from toBeGreaterThan(1) to toBe(3) to match the exact number of arguments_delta
fragments that are emitted in the test, which will help catch unexpected
buffering or coalescing behavior in the adapter where TOOL_CALL_ARGS events
might be unexpectedly combined instead of emitted individually per delta
fragment.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: defaults
Review profile: CHILL
Plan: Pro
Run ID: 28cd2ae9-b604-4d83-aced-432817a342ae
⛔ Files ignored due to path filters (1)
pnpm-lock.yamlis excluded by!**/pnpm-lock.yaml
📒 Files selected for processing (4)
.changeset/gemini-genai-v2-upgrade.mdpackages/ai-gemini/package.jsonpackages/ai-gemini/src/experimental/text-interactions/adapter.tspackages/ai-gemini/tests/text-interactions-adapter.test.ts
✅ Files skipped from review due to trivial changes (1)
- .changeset/gemini-genai-v2-upgrade.md
🚧 Files skipped from review as they are similar to previous changes (1)
- packages/ai-gemini/src/experimental/text-interactions/adapter.ts
Comment on lines
+599
to
+602
| event_type: 'step.delta', | ||
| index: 1, | ||
| delta: { type: 'text', text: 'It is sunny.' }, | ||
| }, |
Contributor
There was a problem hiding this comment.
🎯 Functional Correctness | 🟠 Major | ⚡ Quick win
Missing step.start event for index 1.
The fixture emits step.delta for index 1 (text) without a prior step.start event for that step. All other test fixtures in this PR follow the step.start → step.delta → step.stop protocol described in the PR objectives.
The sequence should include a step.start for index 1 before line 599 to match the SDK 2.x event model.
🔧 Proposed fix to add missing step.start
{ event_type: 'step.stop', index: 0 },
+ {
+ event_type: 'step.start',
+ index: 1,
+ step: { type: 'model_output', id: 'output_1' },
+ },
{
event_type: 'step.delta',
index: 1,
delta: { type: 'text', text: 'It is sunny.' },
},
+ { event_type: 'step.stop', index: 1 },
{
event_type: 'interaction.completed',
interaction: { id: 'int_think', status: 'completed' },📝 Committable suggestion
‼️ IMPORTANT
Carefully review the code before committing. Ensure that it accurately replaces the highlighted code, contains no missing lines, and has no issues with indentation. Thoroughly test & benchmark the code to ensure it meets the requirements.
Suggested change
| event_type: 'step.delta', | |
| index: 1, | |
| delta: { type: 'text', text: 'It is sunny.' }, | |
| }, | |
| { event_type: 'step.stop', index: 0 }, | |
| { | |
| event_type: 'step.start', | |
| index: 1, | |
| step: { type: 'model_output', id: 'output_1' }, | |
| }, | |
| { | |
| event_type: 'step.delta', | |
| index: 1, | |
| delta: { type: 'text', text: 'It is sunny.' }, | |
| }, | |
| { event_type: 'step.stop', index: 1 }, | |
| { | |
| event_type: 'interaction.completed', | |
| interaction: { id: 'int_think', status: 'completed' }, | |
| }, |
🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
In `@packages/ai-gemini/tests/text-interactions-adapter.test.ts` around lines 599
- 602, The test fixture is missing a step.start event for index 1 before the
step.delta event that emits text content. Add a step.start event with index 1
and appropriate start event properties before the existing step.delta event
block (which has event_type: 'step.delta' with index: 1) to establish the SDK
2.x event model sequence of step.start → step.delta → step.stop as followed by
other fixtures in this test file.
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.
Summary
Upgrades
@tanstack/ai-geminito the@google/genaiv2 SDK and migrates thestateful
geminiTextInteractionsadapter to the SDK 2.x Interactions eventprotocol (the May-2026 Interactions breaking changes):
response_mime_type+ flatresponse_format→ polymorphicresponse_format: { type, mime_type, schema }.indexand consumestep.start/step.delta/step.stopandinteraction.created/interaction.completed, including streamedfunction-call
arguments_deltafragments.The existing
geminiTextInteractionssupport is carried forward, notremoved — this PR builds on PR #502 and adapts it to the new SDK.
Verification
packages/ai-gemini— 206 passed, incl. all 29text-interactions-adaptertests (rewritten to emit SDK 2.xstep.*events).pnpm test:pr: sherif, knip, docs, eslint, lib, types, build allgreen across affected packages.
SDK against aimock.
Review feedback addressed
partial-jsoninstead of strictJSON.parse: incompletearguments_deltafragments no longer log a parseerror each, and a truncated buffer keeps the last good args instead of
resetting to
{}.as Record<string, unknown>cast onstep.arguments..agentsroom/tooling (including a personalclaudeSessionId).arguments_deltaaccumulation, argsstream truncation, and
thought_summary→REASONING_*.Known open item — aimock lags SDK 2.x
The
stateful-interactionse2e test is markedtest.fixme(skipped, notfailing). Every published
@copilotkit/aimock(≤ 1.31.0, latest as of2026-06-17) still emits the SDK 1.x Interactions events
(
content.*/interaction.start/interaction.complete), which the migrated2.x adapter doesn't consume — so the mock can't exercise this path yet. This is a
mock/adapter format mismatch, not an adapter bug; 2.x behaviour is covered by
unit tests.
The exact event shapes aimock must emit to re-enable the test are documented in
testing/e2e/tests/stateful-interactions.AIMOCK-TODO.md.TODO before un-drafting
un-
fixmethe e2e test (+ open a follow-up issue to track re-enabling it).pnpm test:prpass.🤖 Generated with Claude Code
Summary by CodeRabbit
New Features
Bug Fixes
Tests
Chores