fix: further anthropic caching improvements

[ErikBjare]

Important

Improves caching by adjusting cache control settings for user messages in gptme/llm/llm_anthropic.py and simplifies an import statement.

• Behavior:
  • Adjusts cache control settings in _prepare_messages_for_api() to set cache points at the last two user messages, following Anthropic's documentation.
• Imports:
  • Simplifies import statement for PromptCachingBetaTextBlockParam by removing unnecessary line breaks.

^{This description was created by for c18a179. It will automatically update as commits are pushed.}

[ellipsis-dev]

👍 Looks good to me! Reviewed everything up to c18a179 in 19 seconds

More details

• Looked at 56 lines of code in 1 files
• Skipped 0 files when reviewing.
• Skipped posting 3 drafted comments based on config settings.

1. gptme/llm/llm_anthropic.py:22

• Draft comment:
  Unnecessary line breaks in import statement removed for better readability.
• Reason this comment was not posted:
  Confidence changes required: 0%
  The import statement was unnecessarily split across multiple lines, which is not a violation but a style preference. The change improves readability.

2. gptme/llm/llm_anthropic.py:339

• Draft comment:
  Removed unused loop variable i for cleaner code.
• Reason this comment was not posted:
  Confidence changes required: 0%
  The loop variable i was removed as it was not used in the loop body. This is a good change as it removes unnecessary code.

3. gptme/llm/llm_anthropic.py:355

• Draft comment:
  Cache control is now correctly applied to the last two user messages, aligning with Anthropic documentation.
• Reason this comment was not posted:
  Confidence changes required: 0%
  The previous logic incorrectly applied cache control only to the last message. The new logic correctly applies it to the last two user messages, as per the Anthropic documentation.

Workflow ID: wflow_VSwLwiAiVVAEOJcR

---

You can customize Ellipsis with 👍 / 👎 feedback, review rules, user-specific overrides, quiet mode, and more.

[codecov-commenter]

Codecov Report

Attention: Patch coverage is 85.71429% with 1 line in your changes missing coverage. Please review.

Project coverage is 73.73%. Comparing base (8757108) to head (c18a179).
Report is 3 commits behind head on master.

✅ All tests successful. No failed tests found.

Files with missing lines	Patch %	Lines
gptme/llm/llm_anthropic.py	85.71%	1 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #318      +/-   ##
==========================================
+ Coverage   72.51%   73.73%   +1.22%     
==========================================
  Files          67       67              
  Lines        4911     5106     +195     
==========================================
+ Hits         3561     3765     +204     
+ Misses       1350     1341       -9     

Flag	Coverage Δ
anthropic/claude-3-haiku-20240307	71.99% <85.71%> (+1.39%)	⬆️
openai/gpt-4o-mini	69.48% <0.00%> (+0.09%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.