FlowRunner Component

The CustomComponent class allows us to create components that interact with Langflow itself. In this example, we will make a component that runs other flows available in "My Collection".

We will cover how to:

• List Collection flows using the list_flows method.
• Load a flow using the load_flow method.
• Configure a dropdown input field using the options parameter.

Example Code

_32

from langflow import CustomComponent

_32

from langchain.schema import Document

_32

_32

class FlowRunner(CustomComponent):

_32

display_name = "Flow Runner"

_32

description = "Run other flows using a document as input."

_32

_32

def build_config(self):

_32

flows = self.list_flows()

_32

flow_names = [f.name for f in flows]

_32

return {"flow_name": {"options": flow_names,

_32

"display_name": "Flow Name",

_32

},

_32

"document": {"display_name": "Document"}

_32

}

_32

_32

_32

def build(self, flow_name: str, document: Document) -> Document:

_32

# List the flows

_32

flows = self.list_flows()

_32

# Get the flow that matches the selected name

_32

# You can also get the flow by id

_32

# using self.get_flow(flow_id=flow_id)

_32

tweaks = {}

_32

flow = self.get_flow(flow_name=flow_name, tweaks=tweaks)

_32

# Get the page_content from the document

_32

if document and isinstance(document, list):

_32

document = document[0]

_32

page_content = document.page_content

_32

# Use it in the flow

_32

result = flow(page_content)

_32

return Document(page_content=str(result))

_12

from langflow import CustomComponent

_12

_12

_12

class MyComponent(CustomComponent):

_12

display_name = "Custom Component"

_12

description = "This is a custom component"

_12

_12

def build_config(self):

_12

...

_12

_12

def build(self):

_12

...

_12

_12

_12

_12

_12

_12

_12

_12

The typical structure of a Custom Component is composed of display_name and description attributes, build and build_config methods.

_12

from langflow import CustomComponent

_12

_12

_12

class FlowRunner(CustomComponent):

_12

display_name = "Flow Runner"

_12

description = "Run other flows"

_12

_12

def build_config(self):

_12

...

_12

_12

def build(self):

_12

...

_12

_12

_12

_12

_12

_12

_12

_12

Let's start by defining our component's display_name and description.

_13

from langflow import CustomComponent

_13

from langchain.schema import Document

_13

_13

_13

class FlowRunner(CustomComponent):

_13

display_name = "Flow Runner"

_13

description = "Run other flows using a document as input."

_13

_13

def build_config(self):

_13

...

_13

_13

def build(self):

_13

...

_13

_13

_13

_13

_13

_13

_13

Second, we will import Document from the langchain.schema module. This will be the return type of the build method.

_13

from langflow import CustomComponent

_13

from langchain.schema import Document

_13

_13

_13

class FlowRunner(CustomComponent):

_13

display_name = "Flow Runner"

_13

description = "Run other flows using a document as input."

_13

_13

def build_config(self):

_13

...

_13

_13

def build(self, flow_name: str, document: Document) -> Document:

_13

...

_13

_13

_13

_13

_13

_13

_13

Now, let's add the parameters and the return type to the build method. The parameters added are:

• flow_name is the name of the flow we want to run.
• document is the input document to be passed to that flow.
  • Since Document is a Langchain type, it will add an input handle to the component (see more).

_14

from langflow import CustomComponent

_14

from langchain.schema import Document

_14

_14

_14

class FlowRunner(CustomComponent):

_14

display_name = "Flow Runner"

_14

description = "Run other flows using a document as input."

_14

_14

def build_config(self):

_14

...

_14

_14

def build(self, flow_name: str, document: Document) -> Document:

_14

# List the flows

_14

flows = self.list_flows()

_14

_14

_14

_14

_14

_14

We can now start writing the build method. Let's list available flows in "My Collection" using the list_flows method.

_19

from langflow import CustomComponent

_19

from langchain.schema import Document

_19

_19

_19

class FlowRunner(CustomComponent):

_19

display_name = "Flow Runner"

_19

description = "Run other flows using a document as input."

_19

_19

def build_config(self):

_19

...

_19

_19

def build(self, flow_name: str, document: Document) -> Document:

_19

# List the flows

_19

flows = self.list_flows()

_19

# Get the flow that matches the selected name

_19

# You can also get the flow by id

_19

# using self.get_flow(flow_id=flow_id)

_19

tweaks = {}

_19

flow = self.get_flow(flow_name=flow_name, tweaks=tweaks)

_19

And retrieve a flow that matches the selected name (we'll make a dropdown input field for the user to choose among flow names).

caution

From version 0.4.0, names are unique, which was not the case in previous versions. This might lead to unexpected results if using flows with the same name.

_19

from langflow import CustomComponent

_19

from langchain.schema import Document

_19

_19

_19

class FlowRunner(CustomComponent):

_19

display_name = "Flow Runner"

_19

description = "Run other flows using a document as input."

_19

_19

def build_config(self):

_19

...

_19

_19

def build(self, flow_name: str, document: Document) -> Document:

_19

# List the flows

_19

flows = self.list_flows()

_19

# Get the flow that matches the selected name

_19

# You can also get the flow by id

_19

# using self.get_flow(flow_id=flow_id)

_19

tweaks = {}

_19

flow = self.get_flow(flow_name=flow_name, tweaks=tweaks)

_19

You can load this flow using get_flow and set a tweaks dictionary to customize it. Find more about tweaks in our features guidelines.

_26

from langflow import CustomComponent

_26

from langchain.schema import Document

_26

_26

_26

class FlowRunner(CustomComponent):

_26

display_name = "Flow Runner"

_26

description = "Run other flows using a document as input."

_26

_26

def build_config(self):

_26

...

_26

_26

def build(self, flow_name: str, document: Document) -> Document:

_26

# List the flows

_26

flows = self.list_flows()

_26

# Get the flow that matches the selected name

_26

# You can also get the flow by id

_26

# using self.get_flow(flow_id=flow_id)

_26

tweaks = {}

_26

flow = self.get_flow(flow_name=flow_name, tweaks=tweaks)

_26

# Get the page_content from the document

We are using a Document as input because it is a straightforward way to pass text data in Langflow (specifically because you can connect it to many loaders). Generally, a flow will take a string or a dictionary as input because that's what LangChain components expect. In case you are passing a dictionary, you need to build it according to the needs of the flow you are using.

The content of a document can be extracted using the page_content attribute, which is a string, and passed as an argument to the selected flow.

_32

from langflow import CustomComponent

_32

from langchain.schema import Document

_32

_32

_32

class FlowRunner(CustomComponent):

_32

display_name = "Flow Runner"

_32

description = "Run other flows using a document as input."

_32

_32

def build_config(self):

_32

flows = self.list_flows()

_32

flow_names = [f.name for f in flows]

_32

return {"flow_name": {"options": flow_names,

_32

"display_name": "Flow Name",

_32

},

_32

"document": {"display_name": "Document"}

_32

}

_32

_32

def build(self, flow_name: str, document: Document) -> Document:

_32

# List the flows

_32

flows = self.list_flows()

Finally, we can add field customizations through the build_config method. Here we added the options key to make the flow_name field a dropdown menu. Check out the custom component reference for a list of available keys.

caution

Make sure that the field type is str and options values are strings.

The typical structure of a Custom Component is composed of display_name and description attributes, build and build_config methods.

Let's start by defining our component's display_name and description.

Second, we will import Document from the langchain.schema module. This will be the return type of the build method.

Now, let's add the parameters and the return type to the build method. The parameters added are:

• flow_name is the name of the flow we want to run.
• document is the input document to be passed to that flow.
  • Since Document is a Langchain type, it will add an input handle to the component (see more).

We can now start writing the build method. Let's list available flows in "My Collection" using the list_flows method.

And retrieve a flow that matches the selected name (we'll make a dropdown input field for the user to choose among flow names).

caution

From version 0.4.0, names are unique, which was not the case in previous versions. This might lead to unexpected results if using flows with the same name.

You can load this flow using get_flow and set a tweaks dictionary to customize it. Find more about tweaks in our features guidelines.

We are using a Document as input because it is a straightforward way to pass text data in Langflow (specifically because you can connect it to many loaders). Generally, a flow will take a string or a dictionary as input because that's what LangChain components expect. In case you are passing a dictionary, you need to build it according to the needs of the flow you are using.

The content of a document can be extracted using the page_content attribute, which is a string, and passed as an argument to the selected flow.

Finally, we can add field customizations through the build_config method. Here we added the options key to make the flow_name field a dropdown menu. Check out the custom component reference for a list of available keys.

caution

Make sure that the field type is str and options values are strings.

_12

from langflow import CustomComponent

_12

_12

_12

class MyComponent(CustomComponent):

_12

display_name = "Custom Component"

_12

description = "This is a custom component"

_12

_12

def build_config(self):

_12

...

_12

_12

def build(self):

_12

...

_12

_12

_12

_12

_12

_12

_12

_12

Done! This is what our script and custom component looks like: