feat: Add SimpleAgentComponent and related pytests (#4318)

* simple ai agent

new component:
simple ai agent

py tests for tool calling agent as well as for simple ai agent

* Update simple_agent.py

update _config Changes and code oprimised

* Update simple_agent.py

Error or inputType as None Solved

* [autofix.ci] apply automated fixes

* fix: INP001 ruff error

* Update test_tool_calling_agent.py

added pytest mark as api required

* Update simple_agent.py

added system prompt

* [autofix.ci] apply automated fixes

* updates in the input field orders

* [autofix.ci] apply automated fixes

* Update simple_agent.py

chat memory added

* [autofix.ci] apply automated fixes

* lint Errors Solved

* Squashed commit of the following:

commit e965fd3b7f
Author: Sebastián Estévez <estevezsebastian@gmail.com>
Date:   Thu Oct 31 17:16:08 2024 -0400

    fix: broken docker test (#4337)

    * Fix broken docker test

    Fixes #4336

    Update the Docker test workflow to use the `get_version` function from `langflow.utils.version`.

    * Modify the `Test image` step in `.github/workflows/docker_test.yml` to use the `get_version` function instead of directly importing `__version__`.
    * Modify the `Test backend image` step in `.github/workflows/docker_test.yml` to use the `get_version` function instead of directly importing `__version__`.

    ---

    For more details, open the [Copilot Workspace session](https://copilot-workspace.githubnext.com/langflow-ai/langflow/issues/4336?shareId=XXXX-XXXX-XXXX-XXXX).

    * version.version not util.version

commit 3279b8a1e8
Author: Lucas Oliveira <62335616+lucaseduoli@users.noreply.github.com>
Date:   Thu Oct 31 16:39:58 2024 -0300

    fix: add collapsible function to templates and fix design bugs (#4305)

    * Updated colors

    * Fixed design for small screens

    * Change border radius

    * Changed size of text on templates description

    * Fix shine effect on small screens

    * Fixed icons on starter templates

    * Updated mono font to JetBrains

    * Updated icon hit area for X

    * Added gradient wrapper and x-gradient

    * Changed colors and font weights for nav component

    * Added zoom on hover of gradient

    * Fixed input size

    * Fixed all templates to show everything

    * Hide scrollbar

    * Change text size of card

    * Removed title of the categories

    * Removed unused currentTab from templatecategory

    * Updated position of search icon

    * Updated style of inputs

    * Updated search clear button

    * Fixed bug on small screens

    * Added no results query

    * Fixed background on get started cards

    * Added focus ring on nav component

    * Added tab index to search and sidebar buttons

    * Added keyboard navigation to templates

    * Updated templatesModal to use ShadCN Sidebar

    * Implemented collapsible sidebar

    * Fix collapsible to work on mobile but be overlaying content

    * Added noise to styleUtils

    * Updated padding and sizes for mobile

    * Updated text size

    * Updated font family to inter

    * Made get started components fetch title and description from the flow

    * Updated description on get started component

    * Updated naming of sidebar

    * Updated description of start from scratch

    * Updated color of selected sidebar item

    * Changed text color for sidebar not active items

    * changed description sizes

    * changed to line clamp

    * Reduced gap between icon and category text

    * Fixed no results state

    * Fixed X icon only appearing on hover

    * Fix auto focus issue

    * fixed hover color of primary button

    * Fixed gradients to use stops if it exists and stop using random gradient

    * removed random gradient

    * Fixed design of cards in templates

    * Updated nav to go through tests

    * Fixed focus on input

    * [autofix.ci] apply automated fixes

    * New color

    * fix testes

    * Fixed starter projects test

    * ✨ (starter-projects.spec.ts): add Page import to test function parameters for better code readability and maintainability
    📝 (starter-projects.spec.ts): refactor test to include a function for waiting for template visibility, improving code readability and reducing duplication

    ---------

    Co-authored-by: Cristhian Zanforlin Lousa <cristhian.lousa@gmail.com>
    Co-authored-by: autofix-ci[bot] <114827586+autofix-ci[bot]@users.noreply.github.com>

commit a03da10750
Author: Sebastián Estévez <estevezsebastian@gmail.com>
Date:   Thu Oct 31 14:57:07 2024 -0400

    fix: langfuse upgrade and crew result usage (#4342)

    * lanfuse upgrade and crew result usage

    * ✅ (Dynamic Agent.spec.ts): Enable tests for checking the presence of required environment variables before running the test suite.

    ✨ (Hierarchical Agent.spec.ts): Add integration test for Hierarchical Tasks Agent to ensure proper functionality and behavior
    ✨ (Sequential Task Agent.spec.ts): Add integration test for Sequential Tasks Agent to ensure proper functionality and behavior

    ---------

    Co-authored-by: cristhianzl <cristhian.lousa@gmail.com>

* [autofix.ci] apply automated fixes

* Update worker.py

adding  # noqa: to solve CI Error

* [autofix.ci] apply automated fixes

* Update worker.py

removed unused params that were causing the issue

---------

Co-authored-by: autofix-ci[bot] <114827586+autofix-ci[bot]@users.noreply.github.com>
Co-authored-by: italojohnny <italojohnnydosanjos@gmail.com>

src/backend/base/langflow/base/agents/agent.py

@@ -136,10 +136,10 @@ class LCAgentComponent(Component):
 
 class LCToolsAgentComponent(LCAgentComponent):
     _base_inputs = [
-        *LCAgentComponent._base_inputs,
         HandleInput(
             name="tools", display_name="Tools", input_types=["Tool", "BaseTool", "StructuredTool"], is_list=True
         ),
+        *LCAgentComponent._base_inputs,
     ]
 
     def build_agent(self) -> AgentExecutor:

src/backend/base/langflow/components/agents/agent.py

@@ -0,0 +1,156 @@
+from langflow.base.agents.agent import LCToolsAgentComponent
+from langflow.base.models.model import LCModelComponent
+from langflow.components.agents.tool_calling import ToolCallingAgentComponent
+from langflow.components.helpers.memory import MemoryComponent
+from langflow.components.models.azure_openai import AzureChatOpenAIComponent
+from langflow.components.models.openai import OpenAIModelComponent
+from langflow.io import (
+    DropdownInput,
+    MultilineInput,
+    Output,
+)
+from langflow.schema.dotdict import dotdict
+from langflow.schema.message import Message
+
+
+def set_advanced_true(component_input):
+    component_input.advanced = True
+    return component_input
+
+
+class AgentComponent(ToolCallingAgentComponent):
+    display_name: str = "Agent"
+    description: str = "Define the agent's instructions, then enter a task to complete using tools."
+    icon = "bot"
+    beta = True
+    name = "Agent"
+
+    azure_inputs = [
+        set_advanced_true(component_input) if component_input.name == "temperature" else component_input
+        for component_input in AzureChatOpenAIComponent().inputs
+        if component_input.name not in [input_field.name for input_field in LCModelComponent._base_inputs]
+    ]
+    openai_inputs = [
+        set_advanced_true(component_input) if component_input.name == "temperature" else component_input
+        for component_input in OpenAIModelComponent().inputs
+        if component_input.name not in [input_field.name for input_field in LCModelComponent._base_inputs]
+    ]
+
+    memory_inputs = [set_advanced_true(component_input) for component_input in MemoryComponent().inputs]
+
+    inputs = [
+        DropdownInput(
+            name="agent_llm",
+            display_name="Model Provider",
+            options=["Azure OpenAI", "OpenAI", "Custom"],
+            value="OpenAI",
+            real_time_refresh=True,
+            refresh_button=True,
+            input_types=[],
+        ),
+        *openai_inputs,
+        MultilineInput(
+            name="system_prompt",
+            display_name="Agent Instructions",
+            info="Initial instructions and context provided to guide the agent's behavior.",
+            value="You are a helpful assistant that can use tools to answer questions and perform tasks.",
+            advanced=False,
+        ),
+        *LCToolsAgentComponent._base_inputs,
+        *memory_inputs,
+    ]
+    outputs = [Output(name="response", display_name="Response", method="get_response")]
+
+    async def get_response(self) -> Message:
+        llm_model = self.get_llm()
+        if llm_model is None:
+            msg = "No language model selected"
+            raise ValueError(msg)
+        self.chat_history = self.get_memory_data()
+
+        agent = ToolCallingAgentComponent().set(
+            llm=llm_model,
+            tools=[self.tools],
+            chat_history=self.chat_history,
+            input_value=self.input_value,
+            system_prompt=self.system_prompt,
+        )
+
+        return await agent.message_response()
+
+    def get_memory_data(self):
+        memory_kwargs = {
+            component_input.name: getattr(self, f"{component_input.name}") for component_input in self.memory_inputs
+        }
+
+        return MemoryComponent().set(**memory_kwargs).retrieve_messages()
+
+    def get_llm(self):
+        try:
+            if self.agent_llm == "OpenAI":
+                return self._build_llm_model(OpenAIModelComponent(), self.openai_inputs)
+            if self.agent_llm == "Azure OpenAI":
+                return self._build_llm_model(AzureChatOpenAIComponent(), self.azure_inputs, prefix="azure_param_")
+        except Exception as e:
+            msg = f"Error building {self.agent_llm} language model"
+            raise ValueError(msg) from e
+        return self.agent_llm
+
+    def _build_llm_model(self, component, inputs, prefix=""):
+        return component.set(
+            **{component_input.name: getattr(self, f"{prefix}{component_input.name}") for component_input in inputs}
+        ).build_model()
+
+    def delete_fields(self, build_config, fields):
+        for field in fields:
+            build_config.pop(field, None)
+
+    def update_build_config(self, build_config: dotdict, field_value: str, field_name: str | None = None):
+        if field_name == "agent_llm":
+            openai_fields = {component_input.name: component_input for component_input in self.openai_inputs}
+            azure_fields = {
+                f"azure_param_{component_input.name}": component_input for component_input in self.azure_inputs
+            }
+
+            if field_value == "OpenAI":
+                self.delete_fields(build_config, {**azure_fields})
+                if not any(field in build_config for field in openai_fields):
+                    build_config.update(openai_fields)
+                build_config["agent_llm"]["input_types"] = []
+                build_config = self.update_input_types(build_config)
+
+            elif field_value == "Azure OpenAI":
+                self.delete_fields(build_config, {**openai_fields})
+                build_config.update(azure_fields)
+                build_config["agent_llm"]["input_types"] = []
+                build_config = self.update_input_types(build_config)
+            elif field_value == "Custom":
+                self.delete_fields(build_config, {**openai_fields})
+                self.delete_fields(build_config, {**azure_fields})
+                new_component = DropdownInput(
+                    name="agent_llm",
+                    display_name="Language Model",
+                    options=["Azure OpenAI", "OpenAI", "Custom"],
+                    value="Custom",
+                    real_time_refresh=True,
+                    input_types=["LanguageModel"],
+                )
+                build_config.update({"agent_llm": new_component.to_dict()})
+                build_config = self.update_input_types(build_config)
+            default_keys = ["code", "_type", "agent_llm", "tools", "input_value"]
+            missing_keys = [key for key in default_keys if key not in build_config]
+            if missing_keys:
+                msg = f"Missing required keys in build_config: {missing_keys}"
+                raise ValueError(msg)
+        return build_config
+
+    def update_input_types(self, build_config):
+        for key, value in build_config.items():
+            # Check if the value is a dictionary
+            if isinstance(value, dict):
+                if value.get("input_types") is None:
+                    build_config[key]["input_types"] = []
+            # Check if the value has an attribute 'input_types' and it is None
+            elif hasattr(value, "input_types") and value.input_types is None:
+                value.input_types = []
+        return build_config

src/backend/tests/unit/components/agents/test_agent_component.py

@@ -0,0 +1,29 @@
+import os
+
+import pytest
+from langflow.components.agents.agent import AgentComponent
+from langflow.components.tools.calculator import CalculatorToolComponent
+
+
+@pytest.mark.api_key_required
+@pytest.mark.asyncio
+async def test_agent_component_with_calculator():
+    # Mock inputs
+    tools = [CalculatorToolComponent().build_tool()]  # Use the Calculator component as a tool
+    input_value = "What is 2 + 2?"
+
+    api_key = os.environ["OPENAI_API_KEY"]
+    temperature = 0.1
+
+    # Initialize the AgentComponent with mocked inputs
+    agent = AgentComponent(
+        tools=tools,
+        input_value=input_value,
+        api_key=api_key,
+        model_name="gpt-4o",
+        llm_type="OpenAI",
+        temperature=temperature,
+    )
+
+    response = await agent.get_response()
+    assert "4" in response.data.get("text")

src/backend/tests/unit/components/agents/test_tool_calling_agent.py

@@ -0,0 +1,30 @@
+import os
+
+import pytest
+from langflow.components.agents.tool_calling import ToolCallingAgentComponent
+from langflow.components.models.openai import OpenAIModelComponent
+from langflow.components.tools.calculator import CalculatorToolComponent
+
+
+@pytest.mark.api_key_required
+@pytest.mark.asyncio
+async def test_tool_calling_agent_component():
+    tools = [CalculatorToolComponent().build_tool()]  # Use the Calculator component as a tool
+    input_value = "What is 2 + 2?"
+    chat_history = []
+    api_key = os.environ["OPENAI_API_KEY"]
+    temperature = 0.1
+
+    # Default OpenAI Model Component
+    llm_component = OpenAIModelComponent().set(
+        api_key=api_key,
+        temperature=temperature,
+    )
+    llm = llm_component.build_model()
+
+    agent = ToolCallingAgentComponent()
+    agent.set(llm=llm, tools=[tools], chat_history=chat_history, input_value=input_value)
+
+    # Chat output
+    response = await agent.message_response()
+    assert "4" in response.data.get("text")

src/frontend/tests/core/integrations/Simple Agent.spec.ts

@@ -77,6 +77,8 @@ test("Simple Agent", async ({ page }) => {
     .getByTestId("popover-anchor-input-api_key")
     .fill(process.env.OPENAI_API_KEY ?? "");
 
+  await page.getByTestId("fit_view").click();
+
   await page.getByTestId("dropdown_str_model_name").click();
   await page.getByTestId("gpt-4o-1-option").click();