beamlit 0.0.41rc85__tar.gz → 0.0.42__tar.gz

beamlit-0.0.42/LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2025 Beamlit, Inc
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

{beamlit-0.0.41rc85 → beamlit-0.0.42}/PKG-INFO

@@ -1,8 +1,9 @@
 Metadata-Version: 2.4
 Name: beamlit
-Version: 0.0.41rc85
+Version: 0.0.42
 Summary: Add your description here
 Author-email: cploujoux <ch.ploujoux@gmail.com>
+License-File: LICENSE
 Requires-Python: >=3.12
 Requires-Dist: asgi-correlation-id<5.0.0,>=4.3.4
 Requires-Dist: attrs>=21.3.0
@@ -17,7 +18,7 @@ Requires-Dist: langchain-mistralai>=0.2.5
 Requires-Dist: langchain-openai<0.4.0,>=0.3.0
 Requires-Dist: langchain-xai>=0.2.0
 Requires-Dist: langgraph<0.3.0,>=0.2.40
-Requires-Dist: mcp>=1.1.2
+Requires-Dist: mcp>=1.2.1
 Requires-Dist: opentelemetry-api>=1.28.2
 Requires-Dist: opentelemetry-exporter-otlp>=1.28.2
 Requires-Dist: opentelemetry-instrumentation-anthropic>=0.35.0

{beamlit-0.0.41rc85 → beamlit-0.0.42}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "beamlit"
-version = "0.0.41.pre-85"
+version = "0.0.42"
 description = "Add your description here"
 readme = "README.md"
 authors = [
@@ -25,7 +25,6 @@ dependencies = [
     "langgraph>=0.2.40,<0.3.0",
     "fastapi[standard]>=0.115.4,<0.116.0",
     "asgi-correlation-id>=4.3.4,<5.0.0",
-    "mcp>=1.1.2",
     "opentelemetry-sdk>=1.28.2",
     "opentelemetry-exporter-otlp>=1.28.2",
     "opentelemetry-api>=1.28.2",
@@ -56,6 +55,7 @@ dependencies = [
     "langchain-anthropic>=0.3.4",
     "langchain-mistralai>=0.2.5",
     "langchain-deepseek-official>=0.1.0",
+    "mcp>=1.2.1",
 ]
 [dependency-groups]
 dev = [

{beamlit-0.0.41rc85 → beamlit-0.0.42}/src/beamlit/agents/chat.py

@@ -1,11 +1,12 @@
 from logging import getLogger
 from typing import Tuple, Union
 
+from langchain_core.language_models import BaseChatModel
+
 from beamlit.api.models import get_model
 from beamlit.authentication import get_authentication_headers, new_client
 from beamlit.common.settings import get_settings
 from beamlit.models import Model
-from langchain_core.language_models import BaseChatModel
 
 logger = getLogger(__name__)
 

{beamlit-0.0.41rc85 → beamlit-0.0.42}/src/beamlit/common/settings.py

@@ -4,8 +4,12 @@ from typing import Tuple, Type, Union
 from langchain_core.language_models.chat_models import BaseChatModel
 from langgraph.graph.graph import CompiledGraph
 from pydantic import Field
-from pydantic_settings import (BaseSettings, PydanticBaseSettingsSource,
-                               SettingsConfigDict, YamlConfigSettingsSource)
+from pydantic_settings import (
+    BaseSettings,
+    PydanticBaseSettingsSource,
+    SettingsConfigDict,
+    YamlConfigSettingsSource,
+)
 
 from beamlit.common.logger import init as init_logger
 from beamlit.models import Agent, Function, Model

{beamlit-0.0.41rc85 → beamlit-0.0.42}/src/beamlit/functions/__init__.py

@@ -1,5 +1,6 @@
 """Functions package providing function decorators and utilities."""
 
-from .decorator import function, get_functions, kit
+from .common import get_functions
+from .decorator import function, kit
 
 __all__ = ["function", "kit", "get_functions"]

beamlit-0.0.41rc85/src/beamlit/functions/decorator.py → beamlit-0.0.42/src/beamlit/functions/common.py

@@ -1,14 +1,11 @@
 """Decorators for creating function tools with Beamlit and LangChain integration."""
 import ast
 import asyncio
-import functools
 import importlib.util
 import os
-from collections.abc import Callable
 from logging import getLogger
 from typing import Union
 
-from fastapi import Request
 from langchain_core.tools import StructuredTool
 from langchain_core.tools.base import create_schema_from_function
 
@@ -17,7 +14,7 @@ from beamlit.client import AuthenticatedClient
 from beamlit.common import slugify
 from beamlit.common.settings import get_settings
 from beamlit.functions.remote.remote import RemoteToolkit
-from beamlit.models import AgentChain, Function, FunctionKit
+from beamlit.models import AgentChain
 
 logger = getLogger(__name__)
 
@@ -144,38 +141,3 @@ def get_functions(
 
     return functions
 
-
-
-def kit(bl_kit: FunctionKit = None, **kwargs: dict) -> Callable:
-    """Create function tools with Beamlit and LangChain integration."""
-
-    def wrapper(func: Callable) -> Callable:
-        if bl_kit and not func.__doc__ and bl_kit.description:
-            func.__doc__ = bl_kit.description
-        return func
-
-    return wrapper
-
-
-def function(*args, function: Function | dict = None, kit=False, **kwargs: dict) -> Callable:
-    """Create function tools with Beamlit and LangChain integration."""
-    if function is not None and not isinstance(function, dict):
-        raise Exception(
-            'function must be a dictionary, example: @function(function={"metadata": {"name": "my_function"}})'
-        )
-    if isinstance(function, dict):
-        function = Function(**function)
-
-    def wrapper(func: Callable) -> Callable:
-        if function and not func.__doc__ and function.spec and function.spec.description:
-            func.__doc__ = function.spec.description
-
-        @functools.wraps(func)
-        async def wrapped(*args, **kwargs):
-            if len(args) > 0 and isinstance(args[0], Request):
-                body = await args[0].json()
-                args = [body.get(param) for param in func.__code__.co_varnames[:func.__code__.co_argcount]]
-            return await func(*args, **kwargs) if asyncio.iscoroutinefunction(func) else func(*args, **kwargs)
-        return wrapped
-
-    return wrapper

beamlit-0.0.42/src/beamlit/functions/decorator.py

@@ -0,0 +1,45 @@
+"""Decorators for creating function tools with Beamlit and LangChain integration."""
+import asyncio
+import functools
+from collections.abc import Callable
+from logging import getLogger
+
+from fastapi import Request
+
+from beamlit.models import Function, FunctionKit
+
+logger = getLogger(__name__)
+
+def kit(bl_kit: FunctionKit = None, **kwargs: dict) -> Callable:
+    """Create function tools with Beamlit and LangChain integration."""
+
+    def wrapper(func: Callable) -> Callable:
+        if bl_kit and not func.__doc__ and bl_kit.description:
+            func.__doc__ = bl_kit.description
+        return func
+
+    return wrapper
+
+
+def function(*args, function: Function | dict = None, kit=False, **kwargs: dict) -> Callable:
+    """Create function tools with Beamlit and LangChain integration."""
+    if function is not None and not isinstance(function, dict):
+        raise Exception(
+            'function must be a dictionary, example: @function(function={"metadata": {"name": "my_function"}})'
+        )
+    if isinstance(function, dict):
+        function = Function(**function)
+
+    def wrapper(func: Callable) -> Callable:
+        if function and not func.__doc__ and function.spec and function.spec.description:
+            func.__doc__ = function.spec.description
+
+        @functools.wraps(func)
+        async def wrapped(*args, **kwargs):
+            if len(args) > 0 and isinstance(args[0], Request):
+                body = await args[0].json()
+                args = [body.get(param) for param in func.__code__.co_varnames[:func.__code__.co_argcount]]
+            return await func(*args, **kwargs) if asyncio.iscoroutinefunction(func) else func(*args, **kwargs)
+        return wrapped
+
+    return wrapper

{beamlit-0.0.41rc85 → beamlit-0.0.42}/src/beamlit/functions/mcp/mcp.py

@@ -1,5 +1,4 @@
 import asyncio
-import urllib.parse
 import warnings
 from typing import Any, Callable
 
@@ -7,39 +6,59 @@ import pydantic
 import pydantic_core
 import requests
 import typing_extensions as t
-from langchain_core.tools.base import BaseTool, BaseToolkit, ToolException
-from mcp import ListToolsResult
-
 from beamlit.authentication.authentication import AuthenticatedClient
 from beamlit.common.settings import get_settings
+from langchain_core.tools.base import BaseTool, BaseToolkit, ToolException
+from mcp.types import CallToolResult, ListToolsResult
+from pydantic.json_schema import JsonSchemaValue
+from pydantic_core import core_schema as cs
 
 settings = get_settings()
 
-def create_dynamic_schema(name: str, schema: dict[str, t.Any]) -> type[pydantic.BaseModel]:
-    field_definitions = {}
-    for k, v in schema["properties"].items():
-        field_type = str
-        if v["type"] == "number":
-            field_type = float
-        elif v["type"] == "integer":
-            field_type = int
-        elif v["type"] == "boolean":
-            field_type = bool
-        description = v.get("description") or ""
-        default_ = v.get("default")
-        fields = {}
-        if default_ is not None:
-            fields["default"] = default_
-        if description is not None:
-            fields["description"] = description
-        field_definitions[k] = (
-            field_type,
-            pydantic.Field(**fields)
-        )
-    return pydantic.create_model(
-        f"{name}Schema",
-        **field_definitions
-    )
+TYPE_MAP = {
+    "integer": int,
+    "number": float,
+    "array": list,
+    "boolean": bool,
+    "string": str,
+    "null": type(None),
+}
+
+FIELD_DEFAULTS = {
+    int: 0,
+    float: 0.0,
+    list: [],
+    bool: False,
+    str: "",
+    type(None): None,
+}
+
+def configure_field(name: str, type_: dict[str, t.Any], required: list[str]) -> tuple[type, t.Any]:
+    field_type = TYPE_MAP[type_["type"]]
+    default_ = FIELD_DEFAULTS.get(field_type) if name not in required else ...
+    return field_type, default_
+
+def create_schema_model(name: str, schema: dict[str, t.Any]) -> type[pydantic.BaseModel]:
+    # Create a new model class that returns our JSON schema.
+    # LangChain requires a BaseModel class.
+    class SchemaBase(pydantic.BaseModel):
+        model_config = pydantic.ConfigDict(extra="allow")
+
+        @t.override
+        @classmethod
+        def __get_pydantic_json_schema__(
+            cls, core_schema: cs.CoreSchema, handler: pydantic.GetJsonSchemaHandler
+        ) -> JsonSchemaValue:
+            return schema
+
+    # Since this langchain patch, we need to synthesize pydantic fields from the schema
+    # https://github.com/langchain-ai/langchain/commit/033ac417609297369eb0525794d8b48a425b8b33
+    required = schema.get("required", [])
+    fields: dict[str, t.Any] = {
+        name: configure_field(name, type_, required) for name, type_ in schema["properties"].items()
+    }
+
+    return pydantic.create_model(f"{name}Schema", __base__=SchemaBase, **fields)
 
 
 
@@ -84,9 +103,10 @@ class MCPTool(BaseTool):
     async def _arun(self, *args: t.Any, **kwargs: t.Any) -> t.Any:
         result = self.client.call_tool(self.name, arguments=kwargs)
         response = result.json()
-        content = pydantic_core.to_json(response["content"]).decode()
-        if response["isError"]:
-            raise ToolException(content)
+        result = CallToolResult(**response)
+        if result.isError:
+            raise ToolException(result.content)
+        content = pydantic_core.to_json(result.content).decode()
         return content
 
     @t.override
@@ -123,7 +143,7 @@ class MCPToolkit(BaseToolkit):
                 client=self.client,
                 name=tool.name,
                 description=tool.description or "",
-                args_schema=create_dynamic_schema(tool.name, tool.inputSchema),
+                args_schema=create_schema_model(tool.name, tool.inputSchema),
             )
             # list_tools returns a PaginatedResult, but I don't see a way to pass the cursor to retrieve more tools
             for tool in self._tools.tools

{beamlit-0.0.41rc85 → beamlit-0.0.42}/src/beamlit/functions/remote/remote.py

@@ -5,6 +5,8 @@ from typing import Callable
 
 import pydantic
 import typing_extensions as t
+from langchain_core.tools.base import BaseTool, ToolException
+
 from beamlit.api.functions import get_function, list_functions
 from beamlit.authentication.authentication import AuthenticatedClient
 from beamlit.common.settings import get_settings
@@ -12,7 +14,6 @@ from beamlit.errors import UnexpectedStatus
 from beamlit.functions.mcp.mcp import MCPClient, MCPToolkit
 from beamlit.models import Function, StoreFunctionParameter
 from beamlit.run import RunClient
-from langchain_core.tools.base import BaseTool, ToolException
 
 
 def create_dynamic_schema(name: str, parameters: list[StoreFunctionParameter]) -> type[pydantic.BaseModel]:

{beamlit-0.0.41rc85 → beamlit-0.0.42}/src/integrationtest/agent.py

@@ -2,7 +2,9 @@ import asyncio
 import uuid
 
 from beamlit.agents import agent
+from beamlit.common import init
 
+settings = init()
 
 @agent(
     agent={
@@ -15,33 +17,29 @@ from beamlit.agents import agent
             "model": "gpt-4o-mini",
         },
     },
-    remote_functions=["math"],
+    remote_functions=["brave-search"],
 )
 async def main(
-    agent, chat_model, tools, body, headers=None, query_params=None, **_
+    input, agent,
 ):
     agent_config = {"configurable": {"thread_id": str(uuid.uuid4())}}
-    json = await body.json()
-    # if "inputs" in json:
-    # body["input"] = json["inputs"]
 
-    agent_body = {"messages": [("user", json["inputs"])]}
+
+    agent_body = {"messages": [("user", input)]}
     responses = []
 
     async for chunk in agent.astream(agent_body, config=agent_config):
+        if "agent" in chunk and "messages" in chunk["agent"]:
+            print(chunk["agent"]["messages"][-1].content)
         responses.append(chunk)
     content = responses[-1]
     return content["agent"]["messages"][-1].content
 
 
 if __name__ == "__main__":
-
     async def check():
-        response = await main(
-            {"input": "What does 4+2 ?"},
-            headers={"X-Beamlit-Sub": "123"},
-            query_params={"debug": "false"},
-        )
+        input = "Generate a dog picture, a golden retriever because they are the best"
+        response = await main(input)
         print(response)
 
     asyncio.run(check())

{beamlit-0.0.41rc85 → beamlit-0.0.42}/src/integrationtest/chat.py

@@ -1,9 +1,10 @@
 import asyncio
 
+from langgraph.graph.graph import CompiledGraph
+
 from beamlit.agents import agent, get_chat_model
 from beamlit.common import init
 from beamlit.functions import function
-from langgraph.graph.graph import CompiledGraph
 
 settings = init()
 chat = get_chat_model("gpt-4o-mini") # all good

{beamlit-0.0.41rc85 → beamlit-0.0.42}/src/integrationtest/uv.lock

@@ -211,7 +211,7 @@ requires-dist = [
     { name = "langchain-openai", specifier = ">=0.3.0,<0.4.0" },
     { name = "langchain-xai", specifier = ">=0.2.0" },
     { name = "langgraph", specifier = ">=0.2.40,<0.3.0" },
-    { name = "mcp", specifier = ">=1.1.2" },
+    { name = "mcp", specifier = ">=1.2.1" },
     { name = "opentelemetry-api", specifier = ">=1.28.2" },
     { name = "opentelemetry-exporter-otlp", specifier = ">=1.28.2" },
     { name = "opentelemetry-instrumentation-anthropic", specifier = ">=0.35.0" },

{beamlit-0.0.41rc85 → beamlit-0.0.42}/uv.lock

@@ -217,7 +217,7 @@ requires-dist = [
     { name = "langchain-openai", specifier = ">=0.3.0,<0.4.0" },
     { name = "langchain-xai", specifier = ">=0.2.0" },
     { name = "langgraph", specifier = ">=0.2.40,<0.3.0" },
-    { name = "mcp", specifier = ">=1.1.2" },
+    { name = "mcp", specifier = ">=1.2.1" },
     { name = "opentelemetry-api", specifier = ">=1.28.2" },
     { name = "opentelemetry-exporter-otlp", specifier = ">=1.28.2" },
     { name = "opentelemetry-instrumentation-anthropic", specifier = ">=0.35.0" },
@@ -1062,7 +1062,7 @@ wheels = [
 
 [[package]]
 name = "mcp"
-version = "1.2.0"
+version = "1.2.1"
 source = { registry = "https://pypi.org/simple" }
 dependencies = [
     { name = "anyio" },
@@ -1074,9 +1074,9 @@ dependencies = [
     { name = "starlette" },
     { name = "uvicorn" },
 ]
-sdist = { url = "https://files.pythonhosted.org/packages/ab/a5/b08dc846ebedae9f17ced878e6975826e90e448cd4592f532f6a88a925a7/mcp-1.2.0.tar.gz", hash = "sha256:2b06c7ece98d6ea9e6379caa38d74b432385c338fb530cb82e2c70ea7add94f5", size = 102973 }
+sdist = { url = "https://files.pythonhosted.org/packages/fc/30/51e4555826126e3954fa2ab1e934bf74163c5fe05e98f38ca4d0f8abbf63/mcp-1.2.1.tar.gz", hash = "sha256:c9d43dbfe943aa1530e2be8f54b73af3ebfb071243827b4483d421684806cb45", size = 103968 }
 wheels = [
-    { url = "https://files.pythonhosted.org/packages/af/84/fca78f19ac8ce6c53ba416247c71baa53a9e791e98d3c81edbc20a77d6d1/mcp-1.2.0-py3-none-any.whl", hash = "sha256:1d0e77d8c14955a5aea1f5aa1f444c8e531c09355c829b20e42f7a142bc0755f", size = 66468 },
+    { url = "https://files.pythonhosted.org/packages/4c/0d/6770742a84c8aa1d36c0d628896a380584c5759612e66af7446af07d8775/mcp-1.2.1-py3-none-any.whl", hash = "sha256:579bf9c9157850ebb1344f3ca6f7a3021b0123c44c9f089ef577a7062522f0fd", size = 66453 },
 ]
 
 [[package]]