Use API response as input data

[gossese]

how to use the response body of the API Request component as text input for the Split Text component?

I want to retrieve the full html of a website and store that in the Astra vector DB. Therefore, I parse the response to only keep the html as text. Up until here, I succeed. But I cannot provide this data output as data input for the Text Splitter Component. What should I do?

[dosubot[bot]]

Hey @gossese! I'm here to help you with any bugs, questions, or contributions. I'm Dosu, a bot designed to assist you.

To use the response body of the APIRequestComponent as text input for the Split Text component, follow these steps:

1. Make the API Request: Use the APIRequestComponent to make the HTTP request and get the response.
2. Extract the Response Body: Access the result field from the Data object returned by the make_requests method.
3. Use the Response Body as Input: Pass the extracted response body to the Split Text component.

Here is a code snippet demonstrating these steps:

# Step 1: Make the API Request
api_request_component = APIRequestComponent()
api_request_component.urls = ["https://api.example.com/data"]
api_request_component.method = "GET"
responses = await api_request_component.make_requests()

# Step 2: Extract the Response Body
response_body = responses[0].data["result"]

# Step 3: Use the Response Body as Input for Split Text Component
split_text_component = SplitTextComponent()
split_text_component.data_inputs = [Data(text=response_body)]
split_text_result = split_text_component.split_text()

In this example:

• APIRequestComponent makes an HTTP GET request to https://api.example.com/data.
• The response body is extracted from the result field of the first Data object in the responses list.
• The extracted response body is then used as the text input for the Split Text component by setting it in the data_inputs field.

This approach ensures that the response body from the API request is correctly passed as input to the Split Text component ^[1].

To continue talking to Dosu, mention @dosu.

[gossese]

@dosu When I do this, there is no data output of the Split Text component.

[dosubot[bot]]

The issue is likely due to a mismatch in the expected input type for the Split Text component. The Split Text component expects input of type Data, but the response body from the API Request component might not be in this format.

Ensure that the response body from the API Request component is properly converted into the Data format before being passed to the Split Text component. Here is an updated example:

# Step 1: Make the API Request
api_request_component = APIRequestComponent()
api_request_component.urls = ["https://api.example.com/data"]
api_request_component.method = "GET"
responses = await api_request_component.make_requests()

# Step 2: Extract the Response Body and Convert to Data Format
response_body = responses[0].data["result"]
data_input = Data(text=response_body)

# Step 3: Use the Response Body as Input for Split Text Component
split_text_component = SplitTextComponent()
split_text_component.data_inputs = [data_input]
split_text_result = split_text_component.split_text()

In this example:

• The response body is extracted and then wrapped in a Data object.
• This Data object is then passed to the Split Text component.

This should ensure that the Split Text component processes the input correctly and produces the expected output ^{[[1]](https://github.com/langflow-ai/langflow/blob/main/src/backend/base/langflow/initial_setup/starter_projects/Vector Store RAG.json)}.

To continue talking to Dosu, mention @dosu.

[gossese]

That was it! I already created a custom component to wrap the response, but I used data_input = Data(value=response_body)

The correct way is data_input = Data(text=response_body)
That way, the Split Text component accepts the input and the chunks can be sent to the DB.