langchain-core 1.0.0a6__py3-none-any.whl → 1.0.4__py3-none-any.whl

langchain_core/messages/block_translators/groq.py

@@ -1,39 +1,135 @@
 """Derivations of standard content blocks from Groq content."""
 
-import warnings
+import json
+import re
+from typing import Any
 
 from langchain_core.messages import AIMessage, AIMessageChunk
 from langchain_core.messages import content as types
+from langchain_core.messages.base import _extract_reasoning_from_additional_kwargs
 
-WARNED = False
 
+def _populate_extras(
+    standard_block: types.ContentBlock, block: dict[str, Any], known_fields: set[str]
+) -> types.ContentBlock:
+    """Mutate a block, populating extras."""
+    if standard_block.get("type") == "non_standard":
+        return standard_block
 
-def translate_content(message: AIMessage) -> list[types.ContentBlock]:  # noqa: ARG001
-    """Derive standard content blocks from a message with Groq content."""
-    global WARNED  # noqa: PLW0603
-    if not WARNED:
-        warning_message = (
-            "Content block standardization is not yet fully supported for Groq."
+    for key, value in block.items():
+        if key not in known_fields:
+            if "extras" not in standard_block:
+                # Below type-ignores are because mypy thinks a non-standard block can
+                # get here, although we exclude them above.
+                standard_block["extras"] = {}  # type: ignore[typeddict-unknown-key]
+            standard_block["extras"][key] = value  # type: ignore[typeddict-item]
+
+    return standard_block
+
+
+def _parse_code_json(s: str) -> dict:
+    """Extract Python code from Groq built-in tool content.
+
+    Extracts the value of the 'code' field from a string of the form:
+    {"code": some_arbitrary_text_with_unescaped_quotes}
+
+    As Groq may not escape quotes in the executed tools, e.g.:
+    ```
+    '{"code": "import math; print("The square root of 101 is: "); print(math.sqrt(101))"}'
+    ```
+    """  # noqa: E501
+    m = re.fullmatch(r'\s*\{\s*"code"\s*:\s*"(.*)"\s*\}\s*', s, flags=re.DOTALL)
+    if not m:
+        msg = (
+            "Could not extract Python code from Groq tool arguments. "
+            "Expected a JSON object with a 'code' field."
         )
-        warnings.warn(warning_message, stacklevel=2)
-        WARNED = True
-    raise NotImplementedError
+        raise ValueError(msg)
+    return {"code": m.group(1)}
+
+
+def _convert_to_v1_from_groq(message: AIMessage) -> list[types.ContentBlock]:
+    """Convert groq message content to v1 format."""
+    content_blocks: list[types.ContentBlock] = []
 
+    if reasoning_block := _extract_reasoning_from_additional_kwargs(message):
+        content_blocks.append(reasoning_block)
 
-def translate_content_chunk(message: AIMessageChunk) -> list[types.ContentBlock]:  # noqa: ARG001
-    """Derive standard content blocks from a message chunk with Groq content."""
-    global WARNED  # noqa: PLW0603
-    if not WARNED:
-        warning_message = (
-            "Content block standardization is not yet fully supported for Groq."
+    if executed_tools := message.additional_kwargs.get("executed_tools"):
+        for idx, executed_tool in enumerate(executed_tools):
+            args: dict[str, Any] | None = None
+            if arguments := executed_tool.get("arguments"):
+                try:
+                    args = json.loads(arguments)
+                except json.JSONDecodeError:
+                    if executed_tool.get("type") == "python":
+                        try:
+                            args = _parse_code_json(arguments)
+                        except ValueError:
+                            continue
+                    elif (
+                        executed_tool.get("type") == "function"
+                        and executed_tool.get("name") == "python"
+                    ):
+                        # GPT-OSS
+                        args = {"code": arguments}
+                    else:
+                        continue
+            if isinstance(args, dict):
+                name = ""
+                if executed_tool.get("type") == "search":
+                    name = "web_search"
+                elif executed_tool.get("type") == "python" or (
+                    executed_tool.get("type") == "function"
+                    and executed_tool.get("name") == "python"
+                ):
+                    name = "code_interpreter"
+                server_tool_call: types.ServerToolCall = {
+                    "type": "server_tool_call",
+                    "name": name,
+                    "id": str(idx),
+                    "args": args,
+                }
+                content_blocks.append(server_tool_call)
+            if tool_output := executed_tool.get("output"):
+                tool_result: types.ServerToolResult = {
+                    "type": "server_tool_result",
+                    "tool_call_id": str(idx),
+                    "output": tool_output,
+                    "status": "success",
+                }
+                known_fields = {"type", "arguments", "index", "output"}
+                _populate_extras(tool_result, executed_tool, known_fields)
+                content_blocks.append(tool_result)
+
+    if isinstance(message.content, str) and message.content:
+        content_blocks.append({"type": "text", "text": message.content})
+
+    for tool_call in message.tool_calls:
+        content_blocks.append(  # noqa: PERF401
+            {
+                "type": "tool_call",
+                "name": tool_call["name"],
+                "args": tool_call["args"],
+                "id": tool_call.get("id"),
+            }
         )
-        warnings.warn(warning_message, stacklevel=2)
-        WARNED = True
-    raise NotImplementedError
+
+    return content_blocks
+
+
+def translate_content(message: AIMessage) -> list[types.ContentBlock]:
+    """Derive standard content blocks from a message with groq content."""
+    return _convert_to_v1_from_groq(message)
+
+
+def translate_content_chunk(message: AIMessageChunk) -> list[types.ContentBlock]:
+    """Derive standard content blocks from a message chunk with groq content."""
+    return _convert_to_v1_from_groq(message)
 
 
 def _register_groq_translator() -> None:
-    """Register the Groq translator with the central registry.
+    """Register the groq translator with the central registry.
 
     Run automatically when the module is imported.
     """

langchain_core/messages/block_translators/langchain_v0.py

@@ -1,6 +1,6 @@
 """Derivations of standard content blocks from LangChain v0 multimodal content."""
 
-from typing import Any, Union, cast
+from typing import Any, cast
 
 from langchain_core.messages import content as types
 
@@ -10,12 +10,12 @@ def _convert_v0_multimodal_input_to_v1(
 ) -> list[types.ContentBlock]:
     """Convert v0 multimodal blocks to v1 format.
 
-    During the `.content_blocks` parsing process, we wrap blocks not recognized as a v1
-    block as a ``'non_standard'`` block with the original block stored in the ``value``
+    During the `content_blocks` parsing process, we wrap blocks not recognized as a v1
+    block as a `'non_standard'` block with the original block stored in the `value`
     field. This function attempts to unpack those blocks and convert any v0 format
     blocks to v1 format.
 
-    If conversion fails, the block is left as a ``'non_standard'`` block.
+    If conversion fails, the block is left as a `'non_standard'` block.
 
     Args:
         content: List of content blocks to process.
@@ -45,7 +45,7 @@ def _convert_v0_multimodal_input_to_v1(
 
 def _convert_legacy_v0_content_block_to_v1(
     block: dict,
-) -> Union[types.ContentBlock, dict]:
+) -> types.ContentBlock | dict:
     """Convert a LangChain v0 content block to v1 format.
 
     Preserves unknown keys as extras to avoid data loss.

langchain_core/messages/block_translators/openai.py

@@ -5,7 +5,7 @@ from __future__ import annotations
 import json
 import warnings
 from collections.abc import Iterable
-from typing import TYPE_CHECKING, Any, Literal, Optional, Union, cast
+from typing import TYPE_CHECKING, Any, Literal, cast
 
 from langchain_core.language_models._utils import (
     _parse_data_uri,
@@ -18,7 +18,7 @@ if TYPE_CHECKING:
 
 
 def convert_to_openai_image_block(block: dict[str, Any]) -> dict:
-    """Convert ``ImageContentBlock`` to format expected by OpenAI Chat Completions."""
+    """Convert `ImageContentBlock` to format expected by OpenAI Chat Completions."""
     if "url" in block:
         return {
             "type": "image_url",
@@ -155,12 +155,12 @@ def _convert_to_v1_from_chat_completions_input(
 ) -> list[types.ContentBlock]:
     """Convert OpenAI Chat Completions format blocks to v1 format.
 
-    During the `.content_blocks` parsing process, we wrap blocks not recognized as a v1
-    block as a ``'non_standard'`` block with the original block stored in the ``value``
+    During the `content_blocks` parsing process, we wrap blocks not recognized as a v1
+    block as a `'non_standard'` block with the original block stored in the `value`
     field. This function attempts to unpack those blocks and convert any blocks that
     might be OpenAI format to v1 ContentBlocks.
 
-    If conversion fails, the block is left as a ``'non_standard'`` block.
+    If conversion fails, the block is left as a `'non_standard'` block.
 
     Args:
         content: List of content blocks to process.
@@ -263,7 +263,7 @@ _FUNCTION_CALL_IDS_MAP_KEY = "__openai_function_call_ids__"
 
 
 def _convert_from_v03_ai_message(message: AIMessage) -> AIMessage:
-    """Convert v0 AIMessage into ``output_version="responses/v1"`` format."""
+    """Convert v0 AIMessage into `output_version="responses/v1"` format."""
     from langchain_core.messages import AIMessageChunk  # noqa: PLC0415
 
     # Only update ChatOpenAI v0.3 AIMessages
@@ -401,7 +401,7 @@ def _convert_from_v03_ai_message(message: AIMessage) -> AIMessage:
 
 def _convert_openai_format_to_data_block(
     block: dict,
-) -> Union[types.ContentBlock, dict[Any, Any]]:
+) -> types.ContentBlock | dict[Any, Any]:
     """Convert OpenAI image/audio/file content block to respective v1 multimodal block.
 
     We expect that the incoming block is verified to be in OpenAI Chat Completions
@@ -677,9 +677,9 @@ def _convert_to_v1_from_responses(message: AIMessage) -> list[types.ContentBlock
                 yield cast("types.ImageContentBlock", new_block)
 
             elif block_type == "function_call":
-                tool_call_block: Optional[
-                    Union[types.ToolCall, types.InvalidToolCall, types.ToolCallChunk]
-                ] = None
+                tool_call_block: (
+                    types.ToolCall | types.InvalidToolCall | types.ToolCallChunk | None
+                ) = None
                 call_id = block.get("call_id", "")
 
                 from langchain_core.messages import AIMessageChunk  # noqa: PLC0415
@@ -726,7 +726,7 @@ def _convert_to_v1_from_responses(message: AIMessage) -> list[types.ContentBlock
                 if "index" in block:
                     web_search_call["index"] = f"lc_wsc_{block['index']}"
 
-                sources: Optional[dict[str, Any]] = None
+                sources: dict[str, Any] | None = None
                 if "action" in block and isinstance(block["action"], dict):
                     if "sources" in block["action"]:
                         sources = block["action"]["sources"]

langchain_core/messages/chat.py

@@ -19,7 +19,7 @@ class ChatMessage(BaseMessage):
     """The speaker / role of the Message."""
 
     type: Literal["chat"] = "chat"
-    """The type of the message (used during serialization). Defaults to "chat"."""
+    """The type of the message (used during serialization)."""
 
 
 class ChatMessageChunk(ChatMessage, BaseMessageChunk):
@@ -29,11 +29,7 @@ class ChatMessageChunk(ChatMessage, BaseMessageChunk):
     # to make sure that the chunk variant can be discriminated from the
     # non-chunk variant.
     type: Literal["ChatMessageChunk"] = "ChatMessageChunk"  # type: ignore[assignment]
-    """The type of the message (used during serialization).
-
-    Defaults to ``'ChatMessageChunk'``.
-
-    """
+    """The type of the message (used during serialization)."""
 
     @override
     def __add__(self, other: Any) -> BaseMessageChunk:  # type: ignore[override]