Commit graph

31 commits

Author SHA1 Message Date
Chris Estreich
a9ca17717c
OpenRouter Gemini caching (#2847)
* OpenRouter Gemini caching

* Fix tests

* Remove unsupported models

* Clean up the task header a bit

* Update src/api/providers/openrouter.ts

Co-authored-by: ellipsis-dev[bot] <65095814+ellipsis-dev[bot]@users.noreply.github.com>

* Remove model that doesn't seem to work

---------

Co-authored-by: ellipsis-dev[bot] <65095814+ellipsis-dev[bot]@users.noreply.github.com>
2025-04-23 09:45:57 -04:00
Chris Estreich
5abea50cf1
Support Gemini 2.5 Flash thinking (#2752) 2025-04-18 12:26:17 -07:00
Smartsheet-JB-Brown
919bb121d0
Jbbrown/bedrock caching (#2071)
* Bedrock prompt caching

* remove whitespace only change to src/api/providers/base-provider.ts

* remove integration test that calls bedrock

* PR prep cleanup

* test file and redundant code cleanup

* updated multi-cache-point

* optimized cache strategy

* cache logic updated. commit before trying to simplify newConversationPlacements

* simplified initial cachePoint logic

* fix the missing caching params for claude 3.7 on bedrock

* undo unintentional changes to README.md files

* fix merge bug that hid use prompt cache checkbox

* prep for PR

* remove unintended changes

* revert .gitignore change

* undo unintended changes

* Issue 1998 - Unable to use ARN. https://github.com/RooVetGit/Roo-Code/issues/1998 - Simplify region extraction from Arns. Change example of ARN use from the foundational model ARN to an inference profile ARN which is what is needed.

* small updates to error message

* more flexible support for ARNs and hardening of ARN handling

* update comment on regex use for platform indipendent cosideration

* small code comment update

* remove duplicative logic to add a region to the arn when cross-region is selected

* Update README files from main branch

* improve error handling when on-demand throughput is not avialable for a foundational model

* fix prompt router bug

* minor tweaks for simplicity

* error messgae formatting

* set new version of bedrock runtime package

* fixes: 2229 - complete prompt doesn't work and appears like a no-op

* remove commented out logger.debug blocks per PR review request

* localize new prompt cache UI inputs

* single source of truth for AWS Region information

---------

Co-authored-by: Matt Rubens <mrubens@users.noreply.github.com>
2025-04-03 13:05:29 -04:00
ashktn
4caa49228f feat: Add support for Gemini models on Vertex AI 2025-02-28 18:59:26 -05:00
cte
52030b12e1 Remove unused code 2025-02-24 22:13:45 -08:00
cte
849f8bf8e8 Add tests 2025-02-24 21:13:23 -08:00
cte
42108193c7 Add tests, remove unused code 2025-02-24 21:08:08 -08:00
cte
543b68f785 Fix tests 2025-02-24 16:09:31 -08:00
cte
22391e5011 Fix type errors 2025-02-24 15:55:04 -08:00
Matt Rubens
816847c5c0 Add Knip check for unused files 2025-02-23 01:01:51 -05:00
Roo Code
6225994a7c Rename 'write_to_file' tool to 'create_file' across codebase and fixed tests 2025-02-22 20:02:17 -07:00
Roo Code
2ddf54de5d Support the Ark provider when used through OpenAI compatible 2025-02-08 23:53:19 -05:00
Matt Rubens
18c7f57afb Add test 2025-01-28 00:15:14 -05:00
Piotr Rogowski
cb23be6346 Extend deepseek-r1 support 2025-01-27 21:27:24 +01:00
Piotr Rogowski
c6607065b9 Add support for displaying reasoning for openrouter models 2025-01-26 02:11:56 +01:00
Matt Rubens
b8e0aa0cde Custom modes 2025-01-21 09:39:54 -05:00
Matt Rubens
60a0a824b9 Prettier backfill 2025-01-17 14:11:28 -05:00
Saoud Rizwan
077fa84374 Add Mistral API provider 2025-01-17 01:11:16 -05:00
Matt Rubens
c673edd526 Merge remote-tracking branch 'origin/main' into vscode-lm-provider 2025-01-15 01:37:37 -05:00
Matt Rubens
d1c39c17e9 Add tests 2025-01-08 17:59:22 -05:00
Matt Rubens
7137f8c528 Spacing adjustments 2025-01-08 17:28:32 -05:00
Matt Rubens
5fbfe9b775 Add test coverage 2025-01-07 16:20:11 -05:00
RaySinner
9d62a7bb77 feat(vscode-lm): implement VS Code Language Models provider 2025-01-07 01:23:22 +03:00
Cline
ca41c54cb5 test(bedrock): add comprehensive test coverage for Bedrock integration
- Add tests for AWS Bedrock handler (stream handling, config, errors)
- Add tests for message format conversion (text, images, tools)
- Add tests for stream event parsing and transformation
- Add tests for cross-region inference and prompt cache
- Add tests for metadata and message lifecycle events
2024-12-11 22:48:49 +02:00
Cline
51a57d5bbf fix(bedrock): improve stream handling and type safety
- Fix TypeScript error in ConverseStreamCommand payload
- Add proper JSON parsing for test stream events
- Improve error handling with proper Error objects
- Add test-specific model info with required fields
- Fix cross-region inference and prompt cache config
2024-12-11 22:47:59 +02:00
Cline
140318cecd feat(api): unify Bedrock provider using Runtime API
Problem:
The current Bedrock implementation uses the Bedrock SDK, which requires separate handling for different model types and doesn't provide a unified streaming interface.

Solution:
Integrate the Bedrock Runtime API to provide a single, unified interface for all Bedrock models (Claude and Nova) using the ConverseStream API. This eliminates the need for separate handlers while maintaining all existing functionality.

Key Changes:
- Refactored AwsBedrockHandler to use @aws-sdk/client-bedrock-runtime
- Enhanced bedrock-converse-format.ts to handle all content types and properly transform between Anthropic and Bedrock formats
- Maintained cross-region inference support with proper region prefixing
- Added support for prompt caching configuration
- Improved AWS credentials handling to better support default providers
- Added proper error handling and token tracking for all response types

Dependencies:
- Added @aws-sdk/client-bedrock-runtime for unified API access
- Removed @anthropic-ai/bedrock-sdk dependency

Testing:
- Verified message format conversion for all content types
- Tested cross-region inference functionality
- Validated streaming responses for both Claude and Nova models

This change simplifies the codebase by providing a single, consistent interface for all Bedrock models while maintaining full compatibility with existing features.
2024-12-11 22:47:56 +02:00
Mark Percival
93e70c62f1
Chore: Prettier for consistant formatting (#794)
* Chore: Pretier for consistant formatting

- TODO: This PR needs to be updated by Saoud after he runs `npm install` & `npm format:fix` and commits the results of the prettier changes.

* Revert prettier config

* Run npm install

* Fix prettier config and ignore package lock

* Run format

---------

Co-authored-by: Saoud Rizwan <7799382+saoudrizwan@users.noreply.github.com>
2024-11-21 13:13:54 -08:00
Saoud Rizwan
8f4a65bc79 Fixes 2024-10-09 01:49:59 -04:00
Saoud Rizwan
06ccaf6f67 Implement streaming for all providers 2024-10-09 01:49:55 -04:00
Saoud Rizwan
7271152f62 Add ApiStream generator to interface with providers 2024-10-09 01:49:55 -04:00
Saoud Rizwan
a009c84597 Refactor API 2024-09-24 10:43:31 -04:00