Deprecate google-generativeai in favor of google-genai #1673

jxnl · 2025-07-07T21:46:43Z

✅ Migrate to google-genai SDK

Summary

This PR completes the migration from the legacy google-generativeai package to the new unified google-genai SDK. The new SDK provides a unified interface for both Gemini API and Vertex AI.

Changes Made

Core Changes

✅ Added deprecation warnings to from_gemini and from_vertexai functions with clear migration examples
✅ Updated auto_client.py to use google-genai for all Google providers
✅ Removed google-generativeai from dependencies
✅ Resolved merge conflicts with main branch
✅ Version bumped to 1.10.0

Documentation Updates

✅ Updated docs/integrations/google.md with migration guide
✅ Updated docs/integrations/vertex.md with migration guide
✅ Added clear examples showing old vs new approaches
✅ Updated installation instructions

Test Migration

✅ Migrated all tests in tests/llm/test_gemini/ to use google-genai
✅ Updated imports from google.generativeai to from google import genai
✅ Updated client creation to use instructor.from_provider("google/model-name")
✅ All 25 Gemini tests passing
✅ Removed deprecated test_distil.py

Migration Examples

For Gemini Users

# Old (deprecated)
import google.generativeai as genai
from instructor import from_gemini
client = from_gemini(genai.GenerativeModel("gemini-1.5-flash"))

# New (recommended)
client = instructor.from_provider("google/gemini-1.5-flash")

For Vertex AI Users

# Old (deprecated)
from vertexai.generative_models import GenerativeModel
from instructor import from_vertexai
client = from_vertexai(GenerativeModel("gemini-1.5-flash"))

# New (recommended)
client = instructor.from_provider("vertexai/gemini-1.5-flash")

Why This Migration Matters

Single SDK: One SDK for both Gemini API and Vertex AI
Simplified Maintenance: Fewer dependencies and integration points
Better UX: Unified interface for all Google AI services
Future-proof: Google's recommended approach

Breaking Changes

from_gemini() and from_vertexai() now show deprecation warnings
These functions will be removed in a future version
Users should migrate to instructor.from_provider() for all Google AI models

Closes #1672

- Warns users to migrate to from_genai or from_provider - Provides clear migration examples - Part of deprecating google-generativeai in favor of google-genai 🤖 Generated with [Claude Code](https://claude.ai/code) Co-Authored-By: Claude <[email protected]>

- Migrate generative-ai provider to use google-genai - Migrate vertexai provider to use google-genai with vertexai=True - Add deprecation warning to from_vertexai function - All Google providers now use the unified google-genai SDK 🤖 Generated with [Claude Code](https://claude.ai/code) Co-Authored-By: Claude <[email protected]>

- Update Google integration docs to use google-genai instead of google-generativeai - Update Vertex AI docs with migration guide - Add clear migration instructions for both Gemini and Vertex AI users - Update all code examples to use from_provider or new SDK 🤖 Generated with [Claude Code](https://claude.ai/code) Co-Authored-By: Claude <[email protected]>

- Remove google-generativeai from pyproject.toml dependencies - Keep from_gemini import for backward compatibility with deprecation comment - Part of migrating to unified google-genai SDK 🤖 Generated with [Claude Code](https://claude.ai/code) Co-Authored-By: Claude <[email protected]>

ellipsis-dev

Important

Looks good to me! 👍

Reviewed everything up to d8142d2 in 1 minute and 56 seconds. Click for details.

Reviewed 515 lines of code in 8 files
Skipped 0 files when reviewing.
Skipped posting 15 draft comments. View those below.
Modify your settings and rules to customize what types of comments Ellipsis leaves. And don't forget to react with 👍 or 👎 to teach Ellipsis.

1. docs/integrations/google.md:7

Draft comment:
Updated heading to 'Google GenAI SDK' clearly reflects the new migration target.
Reason this comment was not posted:
Confidence changes required: 0% <= threshold 85% None

2. docs/integrations/google.md:16

Draft comment:
Installation command updated to use [google-genai] extra; migration is well documented.
Reason this comment was not posted:
Confidence changes required: 0% <= threshold 85% None

3. docs/integrations/google.md:310

Draft comment:
Migration guide examples (old vs new) are clear and provide useful code snippets.
Reason this comment was not posted:
Confidence changes required: 0% <= threshold 85% None

4. docs/integrations/vertex.md:7

Draft comment:
The added warning block effectively notifies users about deprecation for direct vertexai usage.
Reason this comment was not posted:
Confidence changes required: 0% <= threshold 85% None

5. docs/integrations/vertex.md:37

Draft comment:
Setting environment variables for project and location is clearly demonstrated.
Reason this comment was not posted:
Confidence changes required: 0% <= threshold 85% None

6. docs/integrations/vertex.md:225

Draft comment:
The Vertex AI migration section comprehensively shows both 'from_provider' and 'from_genai' approaches.
Reason this comment was not posted:
Confidence changes required: 0% <= threshold 85% None

7. instructor/__init__.py:56

Draft comment:
Deprecated 'from_gemini' is noted; consider adding a similar inline comment for 'from_vertexai' if not already flagged.
Reason this comment was not posted:
Confidence changes required: 33% <= threshold 85% None

8. instructor/auto_client.py:205

Draft comment:
Properly pops 'vertexai' from kwargs to avoid duplicate parameter passing when initializing the client.
Reason this comment was not posted:
Confidence changes required: 0% <= threshold 85% None

9. instructor/auto_client.py:211

Draft comment:
Removal of explicit 'model' parameter in from_genai call is appropriate since the client is already configured with model.
Reason this comment was not posted:
Confidence changes required: 0% <= threshold 85% None

10. instructor/auto_client.py:367

Draft comment:
Retrieving 'project' and 'location' from kwargs or environment and raising a clear error if missing is robust.
Reason this comment was not posted:
Confidence changes required: 0% <= threshold 85% None

11. instructor/auto_client.py:380

Draft comment:
The 'generative-ai' branch consistently instantiates the client for non-vertex use with proper error messaging.
Reason this comment was not posted:
Confidence changes required: 0% <= threshold 85% None

12. instructor/client_gemini.py:32

Draft comment:
Deprecation warning in 'from_gemini' clearly instructs users on how to migrate to the new API.
Reason this comment was not posted:
Confidence changes required: 0% <= threshold 85% None

13. instructor/client_vertexai.py:142

Draft comment:
Deprecation warning in 'from_vertexai' is comprehensive and offers migration samples.
Reason this comment was not posted:
Confidence changes required: 0% <= threshold 85% None

14. pyproject.toml:85

Draft comment:
Removal of 'google-generativeai' dependency and update of extras is correct for the migration.
Reason this comment was not posted:
Confidence changes required: 0% <= threshold 85% None

15. uv.lock:1835

Draft comment:
Lock file updates reflect the removal of 'google-generativeai'; ensure consistency with dependency upgrades.
Reason this comment was not posted:
Confidence changes required: 0% <= threshold 85% None

Workflow ID: wflow_1wTwuzeWRKHbZ36j

^{You can customize}^{by changing your verbosity settings, reacting with 👍 or 👎, replying to comments, or adding code review rules.}

- Update test_simple_types.py to use from_provider - Update test_modes.py to use new SDK and modes - Update util.py with new modes (GENAI_TOOLS, GENAI_STRUCTURED_OUTPUTS) - Fix auto_client.py to properly pass API key from environment - Create new test_genai directory with basic tests - Update conftest.py to import google-genai Note: Tests require a valid GOOGLE_API_KEY to run 🤖 Generated with [Claude Code](https://claude.ai/code) Co-Authored-By: Claude <[email protected]>

ellipsis-dev

Caution

Changes requested ❌

Reviewed d5af543 in 1 minute and 49 seconds. Click for details.

Reviewed 231 lines of code in 6 files
Skipped 0 files when reviewing.
Skipped posting 1 draft comments. View those below.
Modify your settings and rules to customize what types of comments Ellipsis leaves. And don't forget to react with 👍 or 👎 to teach Ellipsis.

1. tests/llm/test_gemini/test_modes.py:23

Draft comment:
For consistency with migration examples, consider using instructor.from_provider instead of instructor.from_genai here.
Reason this comment was not posted:
Confidence changes required: 80% <= threshold 85% None

Workflow ID: wflow_KpbmCRNhgrpGwyWd

^{You can customize}^{by changing your verbosity settings, reacting with 👍 or 👎, replying to comments, or adding code review rules.}

ellipsis-dev · 2025-07-07T21:57:52Z

instructor/auto_client.py

+            vertexai_flag = kwargs.pop("vertexai", False)
+
+            # Get API key from kwargs or environment
+            api_key = kwargs.pop("api_key", os.environ.get("GOOGLE_API_KEY"))


API key extraction and model assignment are duplicated in multiple Google provider branches. Consider refactoring this logic into a helper function.

cloudflare-workers-and-pages · 2025-07-07T21:58:29Z

Deploying with Cloudflare Workers

The latest updates on your project. Learn more about integrating Git with Workers.

Status	Name	Latest Commit	Preview URL	Updated (UTC)
✅ Deployment successful! View logs	instructor	`24378ec`	Commit Preview URL	Jul 08 2025, 01:48 AM

…vider - Remove google-generativeai imports from all test files - Update tests to use instructor.from_provider() instead of from_gemini() - Update GitHub Actions to include GEMINI_API_KEY environment variable - Remove google-generativeai optional dependency from pyproject.toml - Update uv.lock after removing google-generativeai dependency This completes the migration of test files to the new google-genai SDK using the recommended from_provider approach for better consistency. 🤖 Generated with [Claude Code](https://claude.ai/code) Co-Authored-By: Claude <[email protected]>

ellipsis-dev

Important

Looks good to me! 👍

Reviewed 1e57441 in 1 minute and 28 seconds. Click for details.

Reviewed 1271 lines of code in 27 files
Skipped 0 files when reviewing.
Skipped posting 7 draft comments. View those below.
Modify your settings and rules to customize what types of comments Ellipsis leaves. And don't forget to react with 👍 or 👎 to teach Ellipsis.

1. .github/workflows/test.yml:38

Draft comment:
Ensure the GEMINI_API_KEY secret is properly configured in CI. This addition is correct but confirm that it's available in your secrets.
Reason this comment was not posted:
Confidence changes required: 80% <= threshold 85% None

2. pyproject.toml:155

Draft comment:
Removal of the google-generativeai dependency is correct. Verify that migration docs and tests are updated to reference the new google-genai package.
Reason this comment was not posted:
Comment was not on a location in the diff, so it can't be submitted as a review comment.

3. tests/llm/test_gemini/evals/test_classification_enums.py:29

Draft comment:
Switching from instructor.from_gemini to instructor.from_provider simplifies client instantiation. Note that parameterization via product() was removed—ensure all necessary model/mode variations are still covered.
Reason this comment was not posted:
Confidence changes required: 80% <= threshold 85% None

4. tests/llm/test_gemini/test_patch.py:14

Draft comment:
Inconsistent usage detected: some tests use instructor.from_genai while others use instructor.from_provider. Consider unifying these client instantiation methods for consistency.
Reason this comment was not posted:
Confidence changes required: 70% <= threshold 85% None

5. tests/llm/test_vertexai/test_deprecated_async.py:4

Draft comment:
Deprecation warning tests correctly capture the '_async' deprecation. Confirm that related documentation is updated to reflect the change to 'use_async'.
Reason this comment was not posted:
Confidence changes required: 70% <= threshold 85% None

6. uv.lock:1087

Draft comment:
Lockfile updates correctly remove google-generativeai entries in favor of google-genai. Ensure that future dependency changes trigger a regeneration of uv.lock.
Reason this comment was not posted:
Comment did not seem useful. Confidence is useful = 0% <= threshold 85% The comment is about a lockfile update, which is related to dependencies. The rules specify not to comment on dependency changes unless it's about documentation or tests. This comment doesn't mention either, so it should be removed.

7. tests/llm/test_gemini/test_retries.py:29

Draft comment:
It looks like there might be a typo in the test message on line 29: "Extract jason is 12". Did you mean to use "json" instead of "jason"?
Reason this comment was not posted:
Comment was not on a location in the diff, so it can't be submitted as a review comment.

Workflow ID: wflow_QC9Ev7tXjbInBRGl

^{You can customize}^{by changing your verbosity settings, reacting with 👍 or 👎, replying to comments, or adding code review rules.}

…rativeai - Update installation guides to recommend google-genai package - Update modes-comparison.md to use GENAI_TOOLS and GENAI_STRUCTURED_OUTPUTS - Add deprecation warning to gemini_schema property in function_calls.py - Update remaining test files to use google-genai imports - Add fallback imports for backward compatibility in utils.py This completes the documentation updates for the google-genai migration. 🤖 Generated with [Claude Code](https://claude.ai/code) Co-Authored-By: Claude <[email protected]>

ellipsis-dev

Important

Looks good to me! 👍

Reviewed 59d7393 in 1 minute and 42 seconds. Click for details.

Reviewed 235 lines of code in 7 files
Skipped 0 files when reviewing.
Skipped posting 7 draft comments. View those below.
Modify your settings and rules to customize what types of comments Ellipsis leaves. And don't forget to react with 👍 or 👎 to teach Ellipsis.

1. instructor/function_calls.py:159

Draft comment:
The gemini_schema property now emits a deprecation warning and falls back to importing from google.generativeai.types. Consider attempting to import from google.genai.types first (if available) to align closer with the new SDK and update the docstring to explain this fallback behavior.
Reason this comment was not posted:
Confidence changes required: 80% <= threshold 85% None

2. instructor/utils.py:870

Draft comment:
Good use of a try/except block to import HarmCategory and HarmBlockThreshold from google.genai.types with a fallback to the old package. Ensure that both SDK interfaces remain fully compatible.
Reason this comment was not posted:
Comment did not seem useful. Confidence is useful = 0% <= threshold 85% This comment is purely informative and asks the PR author to ensure compatibility, which violates the rules. It does not provide a specific suggestion or point out a specific issue.

3. tests/llm/test_gemini/test_multimodal_content.py:17

Draft comment:
Test has been updated to use instructor.from_provider with mode GENAI_STRUCTURED_OUTPUTS. Ensure additional tests cover the full functionality of the new google-genai SDK.
Reason this comment was not posted:
Decided after close inspection that this draft comment was likely wrong and/or not actionable: usefulness confidence = 10% vs. threshold = 85% The comment has two parts: 1) An observation about the code change (which is just stating what's visible), and 2) A suggestion to "ensure" more test coverage. This violates the rule about not asking authors to "ensure" things. Additionally, the comment about SDK coverage is speculative and would require knowledge of the full codebase to evaluate. The comment might be pointing to a legitimate concern about test coverage for a new SDK integration. Missing test coverage could lead to bugs. While test coverage is important, this comment doesn't provide specific, actionable feedback about what tests are missing. It's the type of speculative "ensure that..." comment that the rules explicitly say to avoid. The comment should be deleted as it combines a non-useful observation of visible changes with a speculative request to "ensure" more testing.

4. instructor/function_calls.py:327

Draft comment:
The assertion in parse_genai_tools enforcing a single non-thought part is strict. Confirm that this behavior aligns with expected use cases and update the error message if future support for multiple function calls is planned.
Reason this comment was not posted:
Confidence changes required: 70% <= threshold 85% None

5. docs/learning/getting_started/installation.md:57

Draft comment:
The installation instructions now reference 'instructor[google-genai]' for Google/Gemini. This correctly reflects the new dependency. Double-check that all related migration examples are updated consistently.
Reason this comment was not posted:
Comment did not seem useful. Confidence is useful = 0% <= threshold 85% This comment is purely informative and asks the PR author to double-check something, which violates the rules. It does not provide a specific code suggestion or request for a test.

6. tests/llm/test_gemini/test_multimodal_content.py:16

Draft comment:
Typo: The function name 'test_audio_compatability_list' contains a typo. Consider renaming it to 'test_audio_compatibility_list' to correct the spelling.
Reason this comment was not posted:
Comment was on unchanged code.

7. tests/llm/test_gemini/test_multimodal_content.py:45

Draft comment:
Typo: In the function name test_audio_compatability_multiple_messages, "compatability" should be spelled "compatibility".
Reason this comment was not posted:
Comment was on unchanged code.

Workflow ID: wflow_ThWeC0KfsNvkwlE9

^{You can customize}^{by changing your verbosity settings, reacting with 👍 or 👎, replying to comments, or adding code review rules.}

- Update test_modes.py to use from_provider and fix model parameter - Update test_stream.py to use from_provider - All tests now properly use the new google-genai approach 🤖 Generated with [Claude Code](https://claude.ai/code) Co-Authored-By: Claude <[email protected]>

These files still need to be migrated to use from_provider 🤖 Generated with [Claude Code](https://claude.ai/code) Co-Authored-By: Claude <[email protected]>

ellipsis-dev

Important

Looks good to me! 👍

Reviewed 3f3d124 in 54 seconds. Click for details.

Reviewed 426 lines of code in 7 files
Skipped 0 files when reviewing.
Skipped posting 2 draft comments. View those below.
Modify your settings and rules to customize what types of comments Ellipsis leaves. And don't forget to react with 👍 or 👎 to teach Ellipsis.

1. tests/llm/test_gemini/test_modes.py:22

Draft comment:
Migration looks good – using 'from_provider' with the formatted provider string. The change from the legacy SDK is correctly applied.
Reason this comment was not posted:
Confidence changes required: 0% <= threshold 85% None

2. tests/llm/test_gemini/test_stream.py:18

Draft comment:
The local variable 'model' shadows the test parameter 'model'. Consider renaming it (e.g., to 'response') for clarity.
Reason this comment was not posted:
Confidence changes required: 50% <= threshold 85% None

Workflow ID: wflow_eIQzuaeCyixEI3xQ

^{You can customize}^{by changing your verbosity settings, reacting with 👍 or 👎, replying to comments, or adding code review rules.}

- Changed all instances of instructor.from_genai to instructor.from_provider in test files. - Deleted test_long_prompt.py and test_retries.py as they are no longer needed. - Updated util.py to reflect the new model version. - Ensured all tests are aligned with the latest google-genai SDK approach.

ellipsis-dev

Important

Looks good to me! 👍

Reviewed 12c80b7 in 2 minutes and 36 seconds. Click for details.

Reviewed 342 lines of code in 8 files
Skipped 0 files when reviewing.
Skipped posting 6 draft comments. View those below.
Modify your settings and rules to customize what types of comments Ellipsis leaves. And don't forget to react with 👍 or 👎 to teach Ellipsis.

1. tests/llm/test_genai/test_format.py:6

Draft comment:
New imports from the google-genai SDK (genai and types) have been added. Confirm these meet migration requirements and don’t conflict with other usages.
Reason this comment was not posted:
Confidence changes required: 0% <= threshold 85% None

2. tests/llm/test_genai/test_invalid_schema.py:102

Draft comment:
A hard-coded provider string 'google/gemini-2.0-flash' is used here while the util file now defines models as 'gemini-2.5-flash'. Consider aligning these for consistency unless intentionally testing legacy behavior.
Reason this comment was not posted:
Confidence changes required: 80% <= threshold 85% None

3. tests/llm/test_genai/test_multimodal.py:49

Draft comment:
Multimodal tests have been updated to use 'instructor.from_provider' and correctly pass autodetect settings for images, PDF, and audio. Ensure that all new integrations (e.g., PDFWithGenaiFile) work as expected with the new SDK.
Reason this comment was not posted:
Confidence changes required: 0% <= threshold 85% None

4. tests/llm/test_genai/test_simple.py:19

Draft comment:
Simple extraction tests now use 'from_provider' with the proper 'async_client' flag for sync and async modes. This update is consistent with the migration and looks correct.
Reason this comment was not posted:
Confidence changes required: 0% <= threshold 85% None

5. tests/llm/test_genai/test_stream.py:19

Draft comment:
Stream-based tests have been updated to use 'from_provider' with the async_client flag. Verify that both streaming and iterable responses maintain expected behavior.
Reason this comment was not posted:
Confidence changes required: 0% <= threshold 85% None

6. tests/llm/test_genai/util.py:3

Draft comment:
The models constant has been updated to 'gemini-2.5-flash'. Ensure that tests using hard-coded provider strings (e.g., 'google/gemini-2.0-flash') are updated if needed for consistency.
Reason this comment was not posted:
Confidence changes required: 80% <= threshold 85% None

Workflow ID: wflow_34UPlliYO9M5s805

^{You can customize}^{by changing your verbosity settings, reacting with 👍 or 👎, replying to comments, or adding code review rules.}

…ner?

ellipsis-dev

Important

Looks good to me! 👍

Reviewed ea366ed in 44 seconds. Click for details.

Reviewed 10 lines of code in 1 files
Skipped 0 files when reviewing.
Skipped posting 1 draft comments. View those below.
Modify your settings and rules to customize what types of comments Ellipsis leaves. And don't forget to react with 👍 or 👎 to teach Ellipsis.

1. tests/llm/test_vertexai/util.py:3

Draft comment:
Updated model version to 'gemini-2.5-flash'. Ensure all tests and mocks expecting the old model 'gemini-2.0-flash' are updated accordingly.
Reason this comment was not posted:
Comment looked like it was already resolved.

Workflow ID: wflow_FESE9Uzjbwc9rt2C

^{You can customize}^{by changing your verbosity settings, reacting with 👍 or 👎, replying to comments, or adding code review rules.}

- Use dynamic environment variable approach for provider tests - Remove duplicate GOOGLE_API_KEY/GEMINI_API_KEY entries - Keep XAI_API_KEY from main branch - Fix merge conflicts in auto_client.py for genai provider - Use from_genai with model parameter consistently 🤖 Generated with [Claude Code](https://claude.ai/code) Co-Authored-By: Claude <[email protected]>

🤖 Generated with [Claude Code](https://claude.ai/code) Co-Authored-By: Claude <[email protected]>

ellipsis-dev

Important

Looks good to me! 👍

Reviewed dd816e0 in 1 minute and 0 seconds. Click for details.

Reviewed 26 lines of code in 2 files
Skipped 0 files when reviewing.
Skipped posting 2 draft comments. View those below.
Modify your settings and rules to customize what types of comments Ellipsis leaves. And don't forget to react with 👍 or 👎 to teach Ellipsis.

1. pyproject.toml:5

Draft comment:
Version bump updated to 1.10.0. Ensure accompanying changelog and version tags are updated.
Reason this comment was not posted:
Confidence changes required: 0% <= threshold 85% None

2. uv.lock:5

Draft comment:
Lock file version updated to 1.10.0. Verify consistency with pyproject.toml.
Reason this comment was not posted:
Confidence changes required: 0% <= threshold 85% None

Workflow ID: wflow_wt5dqh1fUtrORq8z

^{You can customize}^{by changing your verbosity settings, reacting with 👍 or 👎, replying to comments, or adding code review rules.}

- Changed instances of instructor.from_gemini to instructor.from_provider in test_multimodal_content.py. - Updated file handling logic to accommodate changes in the google-genai API. - Added new evals directory and initialized __init__.py. - Updated test_extract_users.py to reflect the new model parameter format. These changes align the tests with the latest google-genai SDK approach.

- The file test_distil.py has been deleted to streamline the test suite and remove outdated tests. This change helps maintain focus on relevant and current testing practices.

ellipsis-dev

Important

Looks good to me! 👍

Reviewed 24378ec in 49 seconds. Click for details.

Reviewed 107 lines of code in 1 files
Skipped 0 files when reviewing.
Skipped posting 1 draft comments. View those below.
Modify your settings and rules to customize what types of comments Ellipsis leaves. And don't forget to react with 👍 or 👎 to teach Ellipsis.

1. tests/test_distil.py:1

Draft comment:
The entire tests file was removed. Ensure that the distil functionality is still tested (via migrated tests or new tests using google-genai) to avoid losing coverage.
Reason this comment was not posted:
Comment was not on a location in the diff, so it can't be submitted as a review comment.

Workflow ID: wflow_7LlmfRPFcYCaZVUM

^{You can customize}^{by changing your verbosity settings, reacting with 👍 or 👎, replying to comments, or adding code review rules.}

ellipsis-dev

Important

Looks good to me! 👍

Reviewed 375e639 in 1 minute and 41 seconds. Click for details.

Reviewed 92 lines of code in 2 files
Skipped 0 files when reviewing.
Skipped posting 5 draft comments. View those below.
Modify your settings and rules to customize what types of comments Ellipsis leaves. And don't forget to react with 👍 or 👎 to teach Ellipsis.

1. tests/llm/test_gemini/evals/test_extract_users.py:25

Draft comment:
Replaced deprecated from_gemini with from_provider. Ensure the provider prefix ('google/') and mode mapping align with the new unified SDK.
Reason this comment was not posted:
Comment looked like it was already resolved.

2. tests/llm/test_gemini/test_multimodal_content.py:16

Draft comment:
Migrated to use from_provider with model string 'google/gemini-1.5-flash-latest'. Confirm that the mode (GEMINI_JSON) is correctly supported by the new SDK.
Reason this comment was not posted:
Comment did not seem useful. Confidence is useful = 0% <= threshold 85% The comment is asking the PR author to confirm if a specific mode is supported by the new SDK. This falls under asking the author to confirm their intention or to double-check things, which is against the rules. Therefore, this comment should not be approved.

3. tests/llm/test_gemini/test_multimodal_content.py:15

Draft comment:
Function name 'test_audio_compatability_list' contains a typo. Use 'compatibility' instead of 'compatability'.
Reason this comment was not posted:
Comment was on unchanged code.

4. tests/llm/test_gemini/test_multimodal_content.py:36

Draft comment:
Function name 'test_audio_compatability_multiple_messages' contains a typo. Correct to 'compatibility'.
Reason this comment was not posted:
Comment was on unchanged code.

5. tests/llm/test_gemini/test_multimodal_content.py:36

Draft comment:
Typo: 'compatability' seems to be misspelled. Consider using 'compatibility'.
Reason this comment was not posted:
Comment was on unchanged code.

Workflow ID: wflow_9VNCEODH69UWVfHi

^{You can customize}^{by changing your verbosity settings, reacting with 👍 or 👎, replying to comments, or adding code review rules.}

jxnl and others added 4 commits July 7, 2025 17:40

ellipsis-dev bot reviewed Jul 7, 2025

View reviewed changes

jxnl and others added 2 commits July 7, 2025 18:58

restore: re-add eval test files that were accidentally deleted

2a7f27e

These files still need to be migrated to use from_provider 🤖 Generated with [Claude Code](https://claude.ai/code) Co-Authored-By: Claude <[email protected]>

ellipsis-dev bot reviewed Jul 7, 2025

View reviewed changes

jxnl added 2 commits July 7, 2025 19:17

tests

cc75577

ellipsis-dev bot reviewed Jul 7, 2025

View reviewed changes

inspect the tests and the github workflows, can the workflows be clea…

ea366ed

…ner?

ellipsis-dev bot reviewed Jul 7, 2025

View reviewed changes

jxnl and others added 2 commits July 7, 2025 20:26

chore: bump version to 1.10.0

dd816e0

🤖 Generated with [Claude Code](https://claude.ai/code) Co-Authored-By: Claude <[email protected]>

ellipsis-dev bot reviewed Jul 8, 2025

View reviewed changes

jxnl added 2 commits July 7, 2025 20:36

delete: remove test_distil.py as it is no longer needed

24378ec

- The file test_distil.py has been deleted to streamline the test suite and remove outdated tests. This change helps maintain focus on relevant and current testing practices.

ellipsis-dev bot reviewed Jul 8, 2025

View reviewed changes

jxnl merged commit eb2228a into main Jul 8, 2025
11 of 14 checks passed

jxnl deleted the deprecate-google-generativeai branch July 8, 2025 00:45

jxnl mentioned this pull request Jul 8, 2025

Google GenAI stopped working #1658

Closed

Uh oh!

Deprecate google-generativeai in favor of google-genai #1673

Deprecate google-generativeai in favor of google-genai #1673

Uh oh!

Conversation

jxnl commented Jul 7, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

✅ Migrate to google-genai SDK

Summary

Changes Made

Core Changes

Documentation Updates

Test Migration

Migration Examples

For Gemini Users

For Vertex AI Users

Why This Migration Matters

Breaking Changes

Uh oh!

ellipsis-dev bot left a comment

Choose a reason for hiding this comment

Uh oh!

ellipsis-dev bot left a comment

Choose a reason for hiding this comment

Uh oh!

ellipsis-dev bot Jul 7, 2025

Choose a reason for hiding this comment

Uh oh!

cloudflare-workers-and-pages bot commented Jul 7, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Deploying with Cloudflare Workers

Uh oh!

ellipsis-dev bot left a comment

Choose a reason for hiding this comment

Uh oh!

ellipsis-dev bot left a comment

Choose a reason for hiding this comment

Uh oh!

ellipsis-dev bot left a comment

Choose a reason for hiding this comment

Uh oh!

ellipsis-dev bot left a comment

Choose a reason for hiding this comment

Uh oh!

ellipsis-dev bot left a comment

Choose a reason for hiding this comment

Uh oh!

ellipsis-dev bot left a comment

Choose a reason for hiding this comment

Uh oh!

ellipsis-dev bot left a comment

Choose a reason for hiding this comment

Uh oh!

ellipsis-dev bot left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

1 participant

jxnl commented Jul 7, 2025 •

edited

Loading

cloudflare-workers-and-pages bot commented Jul 7, 2025 •

edited

Loading