fix: migrated to latest anthropic version v0.42 (prompt caching now stable)

gptme/llm/llm_anthropic.py

@@ -17,9 +17,7 @@
 if TYPE_CHECKING:
     # noreorder
     import anthropic.types  # fmt: skip
-    import anthropic.types.beta.prompt_caching  # fmt: skip
     from anthropic import Anthropic  # fmt: skip
-    from anthropic.types.beta.prompt_caching import PromptCachingBetaTextBlockParam
 
 logger = logging.getLogger(__name__)
 
@@ -55,7 +53,7 @@ def chat(messages: list[Message], model: str, tools: list[ToolSpec] | None) -> s
         messages, tools
     )
 
-    response = _anthropic.beta.prompt_caching.messages.create(
+    response = _anthropic.messages.create(
         model=model,
         messages=messages_dicts,
         system=system_messages,
@@ -75,15 +73,14 @@ def stream(
     messages: list[Message], model: str, tools: list[ToolSpec] | None
 ) -> Generator[str, None, None]:
     import anthropic.types  # fmt: skip
-    import anthropic.types.beta.prompt_caching  # fmt: skip
     from anthropic import NOT_GIVEN  # fmt: skip
 
     assert _anthropic, "LLM not initialized"
     messages_dicts, system_messages, tools_dict = _prepare_messages_for_api(
         messages, tools
     )
 
-    with _anthropic.beta.prompt_caching.messages.stream(
+    with _anthropic.messages.stream(
         model=model,
         messages=messages_dicts,
         system=system_messages,
@@ -121,7 +118,7 @@ def stream(
                     pass
                 case "message_start":
                     chunk = cast(
-                        anthropic.types.beta.prompt_caching.RawPromptCachingBetaMessageStartEvent,
+                        anthropic.types.MessageStartEvent,
                         chunk,
                     )
                     logger.debug(chunk.message.usage)
@@ -200,7 +197,7 @@ def _process_file(message_dict: dict) -> dict:
 
 def _transform_system_messages(
     messages: list[Message],
-) -> tuple[list[Message], list["PromptCachingBetaTextBlockParam"]]:
+) -> tuple[list[Message], list["anthropic.types.TextBlockParam"]]:
     """Transform system messages into Anthropic's expected format.
 
     This function:
@@ -215,7 +212,7 @@ def _transform_system_messages(
     2. Earlier messages in multi-turn conversations
 
     Returns:
-        tuple[list[Message], list[PromptCachingBetaTextBlockParam]]: Transformed messages and system messages
+        tuple[list[Message], list[TextBlockParam]]: Transformed messages and system messages
     """
     assert messages[0].role == "system"
     system_prompt = messages[0].content
@@ -244,7 +241,7 @@ def _transform_system_messages(
         else:
             messages_new.append(message)
     messages = messages_new
-    system_messages: list[PromptCachingBetaTextBlockParam] = [
+    system_messages: list[anthropic.types.TextBlockParam] = [
         {
             "type": "text",
             "text": system_prompt,
@@ -273,7 +270,7 @@ def parameters2dict(parameters: list[Parameter]) -> dict[str, object]:
 
 def _spec2tool(
     spec: ToolSpec,
-) -> "anthropic.types.beta.prompt_caching.PromptCachingBetaToolParam":
+) -> "anthropic.types.ToolParam":
     name = spec.name
     if spec.block_types:
         name = spec.block_types[0]
@@ -289,9 +286,9 @@ def _spec2tool(
 def _prepare_messages_for_api(
     messages: list[Message], tools: list[ToolSpec] | None
 ) -> tuple[
-    list["anthropic.types.beta.prompt_caching.PromptCachingBetaMessageParam"],
-    list["anthropic.types.beta.prompt_caching.PromptCachingBetaTextBlockParam"],
-    list["anthropic.types.beta.prompt_caching.PromptCachingBetaToolParam"] | None,
+    list["anthropic.types.MessageParam"],
+    list["anthropic.types.TextBlockParam"],
+    list["anthropic.types.ToolParam"] | None,
 ]:
     """Prepare messages for the Anthropic API.
 
@@ -312,13 +309,7 @@ def _prepare_messages_for_api(
         - Tool dictionaries (if tools provided)
     """
     # noreorder
-    from anthropic.types.beta.prompt_caching import (  # fmt: skip
-        PromptCachingBetaImageBlockParam,
-        PromptCachingBetaMessageParam,
-        PromptCachingBetaTextBlockParam,
-        PromptCachingBetaToolResultBlockParam,
-        PromptCachingBetaToolUseBlockParam,
-    )
+    import anthropic.types  # fmt: skip
 
     # Transform system messages
     messages, system_messages = _transform_system_messages(messages)
@@ -327,13 +318,13 @@ def _prepare_messages_for_api(
     messages_dicts = _handle_files(msgs2dicts(messages))
 
     # Apply cache control to optimize performance
-    messages_dicts_new: list[PromptCachingBetaMessageParam] = []
+    messages_dicts_new: list[anthropic.types.MessageParam] = []
     for msg in messages_dicts:
         content_parts: list[
-            PromptCachingBetaTextBlockParam
-            | PromptCachingBetaImageBlockParam
-            | PromptCachingBetaToolUseBlockParam
-            | PromptCachingBetaToolResultBlockParam
+            anthropic.types.TextBlockParam
+            | anthropic.types.ImageBlockParam
+            | anthropic.types.ToolUseBlockParam
+            | anthropic.types.ToolResultBlockParam
         ] = []
         raw_content = (
             msg["content"]
@@ -359,7 +350,7 @@ def _prepare_messages_for_api(
     # https://docs.anthropic.com/en/docs/build-with-claude/prompt-caching#continuing-a-multi-turn-conversation
     for msgp in [msg for msg in messages_dicts_new if msg["role"] == "user"][-2:]:
         assert isinstance(msgp["content"], list)
-        msgp["content"][-1]["cache_control"] = {"type": "ephemeral"}
+        msgp["content"][-1]["cache_control"] = {"type": "ephemeral"}  # type: ignore
 
     # Prepare tools
     tools_dict = [_spec2tool(tool) for tool in tools] if tools else None

pyproject.toml

@@ -59,7 +59,7 @@ json-repair = "^0.30.2"
 
 # providers
 openai = "^1.0"
-anthropic = "^0.40"
+anthropic = "^0.42"
 
 # tools
 ipython = "^8.17.2"