oss(ps): import from langchain instead of langchain_core where possible

src/oss/deepagents/human-in-the-loop.mdx

@@ -15,7 +15,7 @@ The `interrupt_on` parameter accepts a dictionary mapping tool names to interrup
 
 :::python
 ```python
-from langchain_core.tools import tool
+from langchain.tools import tool
 from deepagents import create_deep_agent
 from langgraph.checkpoint.memory import MemorySaver
 

src/oss/deepagents/middleware.mdx

@@ -186,7 +186,7 @@ The subagents middleware allows you to supply subagents through a `task` tool.
 
 :::python
 ```python
-from langchain_core.tools import tool
+from langchain.tools import tool
 from langchain.agents import create_agent
 from deepagents.middleware.subagents import SubAgentMiddleware
 

src/oss/langchain/agents.mdx

@@ -303,7 +303,7 @@ To customize how tool errors are handled, use the @[`@wrap_tool_call`] decorator
 ```python wrap
 from langchain.agents import create_agent
 from langchain.agents.middleware import wrap_tool_call
-from langchain_core.messages import ToolMessage
+from langchain.messages import ToolMessage
 
 
 @wrap_tool_call

src/oss/langchain/guardrails.mdx

@@ -454,7 +454,7 @@ Use "after agent" hooks to validate final outputs once before returning to the u
 ```python title="Class syntax"
 from langchain.agents.middleware import AgentMiddleware, AgentState, hook_config
 from langgraph.runtime import Runtime
-from langchain_core.messages import AIMessage
+from langchain.messages import AIMessage
 from langchain.chat_models import init_chat_model
 from typing import Any
 
@@ -511,7 +511,7 @@ result = agent.invoke({
 ```python title="Decorator syntax"
 from langchain.agents.middleware import after_agent, AgentState, hook_config
 from langgraph.runtime import Runtime
-from langchain_core.messages import AIMessage
+from langchain.messages import AIMessage
 from langchain.chat_models import init_chat_model
 from typing import Any
 

src/oss/langchain/middleware/built-in.mdx

@@ -1473,7 +1473,7 @@ The middleware uses an LLM to generate plausible responses for tool calls instea
 ```python
 from langchain.agents import create_agent
 from langchain.agents.middleware import LLMToolEmulator
-from langchain_core.tools import tool
+from langchain.tools import tool
 
 
 @tool
@@ -1974,7 +1974,7 @@ The middleware adds two search tools to agents:
 ```python
 from langchain.agents import create_agent
 from langchain.agents.middleware import FilesystemFileSearchMiddleware
-from langchain_core.messages import HumanMessage
+from langchain.messages import HumanMessage
 
 
 agent = create_agent(
@@ -2113,7 +2113,7 @@ The middleware caches content up to and including the latest message in each req
 from langchain_anthropic import ChatAnthropic
 from langchain_anthropic.middleware import AnthropicPromptCachingMiddleware
 from langchain.agents import create_agent
-from langchain_core.messages import HumanMessage
+from langchain.messages import HumanMessage
 
 
 LONG_PROMPT = """

src/oss/langchain/middleware/custom.mdx

@@ -736,7 +736,7 @@ const dynamicModelMiddleware = createMiddleware({
 ```python
 from langchain.agents.middleware import wrap_tool_call
 from langchain.tools.tool_node import ToolCallRequest
-from langchain_core.messages import ToolMessage
+from langchain.messages import ToolMessage
 from langgraph.types import Command
 from typing import Callable
 
@@ -764,7 +764,7 @@ def monitor_tool(
 ```python
 from langchain.tools.tool_node import ToolCallRequest
 from langchain.agents.middleware import AgentMiddleware
-from langchain_core.messages import ToolMessage
+from langchain.messages import ToolMessage
 from langgraph.types import Command
 from typing import Callable
 

src/oss/langchain/models.mdx

@@ -187,7 +187,7 @@ response = model.invoke(conversation)
 print(response)  # AIMessage("J'adore créer des applications.")
 ```
 ```python Message objects
-from langchain_core.messages import HumanMessage, AIMessage, SystemMessage
+from langchain.messages import HumanMessage, AIMessage, SystemMessage
 
 conversation = [
     SystemMessage("You are a helpful assistant that translates English to French."),

src/oss/langchain/supervisor.mdx

@@ -112,7 +112,7 @@ Start by defining the tools that require structured inputs. In real applications
 
 :::python
 ```python
-from langchain_core.tools import tool
+from langchain.tools import tool
 
 @tool
 def create_calendar_event(
@@ -708,7 +708,7 @@ A supervisor agent coordinates specialized sub-agents (calendar and email)
 that are wrapped as tools.
 """
 
-from langchain_core.tools import tool
+from langchain.tools import tool
 from langchain.agents import create_agent
 from langchain.chat_models import init_chat_model
 

src/oss/langgraph/thinking-in-langgraph.mdx

@@ -478,7 +478,7 @@ We'll implement each node as a simple function. Remember: nodes take state, do w
     from langgraph.graph import StateGraph, START, END
     from langgraph.types import interrupt, Command, RetryPolicy
     from langchain_openai import ChatOpenAI
-    from langchain_core.messages import HumanMessage
+    from langchain.messages import HumanMessage
 
     llm = ChatOpenAI(model="gpt-4")
 

src/oss/python/migrate/langchain-v1.mdx

@@ -458,7 +458,7 @@ from typing_extensions import Annotated
 from pydantic import BaseModel
 from langgraph.graph import StateGraph
 from langgraph.graph.messages import add_messages
-from langchain_core.messages import AnyMessage
+from langchain.messages import AnyMessage
 
 
 class AgentState(BaseModel):  # [!code highlight]