feat(anthropic): improve prompt caching and type safety #317
Conversation
This file contains 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
- Add better type hints for Anthropic API types - Optimize prompt caching with ephemeral cache control - Extract message preparation into dedicated function - Improve documentation
![ellipsis-dev[bot]](https://avatars.githubusercontent.com/in/64358?s=60&v=4){kind=small}
There was a problem hiding this comment.
👍 Looks good to me! Reviewed everything up to 9c77d71 in 1 minute and 2 seconds
More details
- Looked at
220lines of code in1files - Skipped
0files when reviewing. - Skipped posting
6drafted comments based on config settings.
1. gptme/llm/llm_anthropic.py:19
- Draft comment:
The import 'import anthropic.types' is unused and can be removed to clean up the code. - Reason this comment was not posted:
Confidence changes required:10%
The import statement for 'anthropic.types' is not used in the file. It should be removed to clean up the code.
2. gptme/llm/llm_anthropic.py:20
- Draft comment:
The import 'import anthropic.types.beta.prompt_caching' is unused and can be removed to clean up the code. - Reason this comment was not posted:
Confidence changes required:10%
The import statement for 'anthropic.types.beta.prompt_caching' is not used in the file. It should be removed to clean up the code.
3. gptme/llm/llm_anthropic.py:21
- Draft comment:
The import 'from anthropic import Anthropic' is unused and can be removed to clean up the code. - Reason this comment was not posted:
Confidence changes required:10%
The import statement for 'anthropic' is not used in the file. It should be removed to clean up the code.
4. gptme/llm/llm_anthropic.py:202
- Draft comment:
The docstring mentions cache control logic that was removed. Update the docstring to reflect the current functionality. - Reason this comment was not posted:
Confidence changes required:20%
The function '_transform_system_messages' has a detailed docstring, but it doesn't mention the cache control logic that was removed. The docstring should be updated to reflect the current functionality.
5. gptme/llm/llm_anthropic.py:208
- Draft comment:
The docstring mentions cache control logic that was removed. Update the docstring to reflect the current functionality. - Reason this comment was not posted:
Confidence changes required:20%
The function '_transform_system_messages' has a detailed docstring, but it doesn't mention the cache control logic that was removed. The docstring should be updated to reflect the current functionality.
6. gptme/llm/llm_anthropic.py:212
- Draft comment:
The docstring mentions cache control logic that was removed. Update the docstring to reflect the current functionality. - Reason this comment was not posted:
Confidence changes required:20%
The function '_transform_system_messages' has a detailed docstring, but it doesn't mention the cache control logic that was removed. The docstring should be updated to reflect the current functionality.
Workflow ID: wflow_5z9GMKNg9BR8tfAH
You can customize Ellipsis with 👍 / 👎 feedback, review rules, user-specific overrides, quiet mode, and more.
Codecov ReportAttention: Patch coverage is
✅ All tests successful. No failed tests found.
Additional details and impacted files@@ Coverage Diff @@
## master #317 +/- ##
==========================================
- Coverage 73.56% 73.44% -0.13%
==========================================
Files 68 68
Lines 4975 4978 +3
==========================================
- Hits 3660 3656 -4
- Misses 1315 1322 +7
Flags with carried forward coverage won't be shown. Click here to find out more. ☔ View full report in Codecov by Sentry. |
ErikBjare
commented
Dec 10, 2024
|
|
||
| for part in raw_content: | ||
| if isinstance(part, dict): | ||
| if part.get("type") == "text" and i == len(messages_dicts) - 1: |
There was a problem hiding this comment.
Might want to review this later to cache at the optimal locations, especially considering RAG and "live" messages. Should probably also add a cache step at the system prompt.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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.
Costs were looking a bit expensive, realized that the prompt caching implementation is pretty sub-par and doesn't cache longer contexts when they appear in the conversation (only system messages).