Input several values into the flow via API

[andruum]

Hi!

I'm trying to figure out how to use Langflow with API

Currently I face an issue - I can't put two values in the API call.

It works fine if I have only one TextInput Block and use input_value as a field in my API call.

But if I use two TextInput Blocks - It doesn't work.

What should I do and what is the correct way to do so?

[dosubot[bot]]

Hey @andruum!

I'm here to help you with any bugs, questions or if you want to become a contributor. Just ask!

To handle multiple TextInput blocks in an API call when using Langflow, you can define a custom component with multiple input fields and corresponding output methods. Here is an example:

from langflow.base.io.text import TextComponent
from langflow.io import MessageTextInput, IntInput, Output
from langflow.schema.message import Message

class MultipleTextInputComponent(TextComponent):
    display_name = "Multiple Text Input"
    description = "Get multiple text inputs from the Playground."
    icon = "type"

    inputs = [
        MessageTextInput(
            name="input_value_1",
            display_name="Text 1",
            info="First text input.",
        ),
        MessageTextInput(
            name="input_value_2",
            display_name="Text 2",
            info="Second text input.",
        ),
        IntInput(
            name="number",
            display_name="Number",
            info="An integer input.",
        ),
    ]
    outputs = [
        Output(display_name="Text 1 Response", name="text_1_response", method="text_1_response"),
        Output(display_name="Text 2 Response", name="text_2_response", method="text_2_response"),
        Output(display_name="Number Response", name="number_response", method="number_response"),
    ]

    def text_1_response(self) -> Message:
        return Message(text=self.input_value_1)

    def text_2_response(self) -> Message:
        return Message(text=self.input_value_2)

    def number_response(self) -> str:
        return f"This is my int input multiplied by 2: {self.number * 2}"

In this example, the MultipleTextInputComponent class has three input fields: two MessageTextInput fields for string inputs and one IntInput for an integer input. It also has three output methods: text_1_response, text_2_response, and number_response, which process the inputs and return the respective outputs. This setup allows you to handle multiple TextInput blocks in an API call ^[1].

Additionally, you can refer to the ChatInput class in Langflow for more advanced handling of multiple inputs. This class demonstrates how to handle various input types such as MultilineInput, DropdownInput, FileInput, and MessageTextInput, which can be useful if you need to extend the functionality of your custom component ^[1].

To continue talking to Dosu, mention @dosu.

[andruum]

@dosu this approach seems not to work further as I try to input these values into a prompt with multiple input variables. It seems like a problem inside flow

[dosubot[bot]]

To correctly input multiple values into a prompt with multiple input variables using Langflow, you need to define the input variables in the PromptInput and then use them in the prompt template. Here is an example of how you can set this up:

1. Define the input variables in the PromptInput list.
2. Use these variables in your prompt template.

Here is a code example:

from langflow.custom import Component
from langflow.io import Output, PromptInput
from langflow.schema.message import Message

class PromptComponent(Component):
    display_name: str = "Prompt"
    description: str = "Create a prompt template with dynamic variables."
    icon = "prompts"
    trace_type = "prompt"

    inputs = [
        PromptInput(name="template", display_name="Template"),
    ]

    outputs = [
        Output(display_name="Prompt Message", name="prompt", method="build_prompt"),
    ]

    async def build_prompt(self) -> Message:
        prompt = await Message.from_template_and_variables(**self._attributes)
        self.status = prompt.text
        return prompt

    def post_code_processing(self, new_build_config: dict, current_build_config: dict):
        frontend_node = super().post_code_processing(new_build_config, current_build_config)
        template = frontend_node["template"]["template"]["value"]
        _ = process_prompt_template(
            template=template,
            name="template",
            custom_fields=frontend_node["custom_fields"],
            frontend_node_template=frontend_node["template"],
        )
        update_template_values(frontend_template=frontend_node, raw_template=current_build_config["template"])
        return frontend_node

In this example, you can create a template like:

"Hello {variable1}, your order number is {variable2}."

When you input values for variable1 and variable2, they will be dynamically inserted into the template.

Additionally, you can use the validate_prompt function to ensure that your prompt template is correctly set up with valid input variables. This function checks for invalid characters and ensures that the prompt template is valid:

from langflow.api.v1.base import check_input_variables
from langflow.interface.utils import extract_input_variables_from_prompt
from langflow.template.field.prompt import DefaultPromptField

def validate_prompt(prompt_template: str, silent_errors: bool = False) -> list[str]:
    input_variables = extract_input_variables_from_prompt(prompt_template)

    input_variables = check_input_variables(input_variables)
    if any(var in INVALID_NAMES for var in input_variables):
        raise ValueError(f"Invalid input variables. None of the variables can be named {', '.join(input_variables)}. ")

    try:
        PromptTemplate(template=prompt_template, input_variables=input_variables)
    except Exception as exc:
        logger.error(f"Invalid prompt: {exc}")
        if not silent_errors:
            raise ValueError(f"Invalid prompt: {exc}") from exc

    return input_variables

This function will help you ensure that your prompt template is correctly formatted and that all input variables are valid ^{[[1]](https://github.com/langflow-ai/langflow/blob/dev/src/backend/base/langflow/initial_setup/starter_projects/Langflow Prompt Chaining.json)[2][3]}.

To continue talking to Dosu, mention @dosu.

[andruum]

@dosu how can I pass several input variables to prompt via API?

[dosubot[bot]]

To pass several input variables to a prompt via API using Langflow, follow these steps:

1. Define the Prompt Template: Create a template string with placeholders for the input variables.
2. Validate the Prompt Template: Use the validate_prompt function to ensure all input variables are correctly formatted.

Example

Define the Prompt Template

template = "Hello, {name}. You have {count} new messages."

Validate the Prompt Template

from langflow.base.prompts.api_utils import validate_prompt

try:
    input_variables = validate_prompt(template)
    print("Input variables are valid:", input_variables)
except ValueError as e:
    print("Validation error:", e)

Supporting Code

The validate_prompt function performs the following steps:

1. Extract Input Variables: It uses the extract_input_variables_from_prompt function to extract the input variables from the template.
2. Check for Invalid Characters: It checks for invalid characters in the input variables using the check_input_variables function.
3. Validate Input Variables: It ensures that none of the input variables have invalid names or characters and raises an error if any issues are found.

Here is the relevant code for validation:

from collections import defaultdict
from typing import Any, Dict, List, Optional

from fastapi import HTTPException
from langchain_core.prompts import PromptTemplate
from loguru import logger

from langflow.api.v1.base import INVALID_NAMES, check_input_variables
from langflow.interface.utils import extract_input_variables_from_prompt
from langflow.template.field.prompt import DefaultPromptField

def validate_prompt(prompt_template: str, silent_errors: bool = False) -> list[str]:
    input_variables = extract_input_variables_from_prompt(prompt_template)

    # Check if there are invalid characters in the input_variables
    input_variables = check_input_variables(input_variables)
    if any(var in INVALID_NAMES for var in input_variables):
        raise ValueError(f"Invalid input variables. None of the variables can be named {', '.join(input_variables)}. ")

    try:
        PromptTemplate(template=prompt_template, input_variables=input_variables)
    except Exception as exc:
        logger.error(f"Invalid prompt: {exc}")
        if not silent_errors:
            raise ValueError(f"Invalid prompt: {exc}") from exc

    return input_variables

This code ensures that all input variables in the prompt template are correctly formatted and valid ^[1][2][3].

To continue talking to Dosu, mention @dosu.

[andruum]

@dosu how to call API after that, if I have {name} and {count} input variables for my flow?

[dosubot[bot]]

To call the API with {name} and {count} input variables for your flow, follow these steps:

1. Define the Prompt Template: Create a template string with placeholders for the input variables.
2. Pass the Variables in the API Request: Include the values for name and count in your API request.

Example

Define the Prompt Template

{
  "template": "Hello {name}, you have {count} new messages."
}

API Request

Make an API request with the template and variables. Here is an example request:

{
  "template": "Hello {name}, you have {count} new messages.",
  "variables": {
    "name": "John",
    "count": 5
  }
}

API Response

The API will process the template and return the processed prompt. Example response:

{
  "processed_prompt": "Hello John, you have 5 new messages."
}

By following these steps, you can pass {name} and {count} to the API and get the processed prompt as a response ^[1][2].

To continue talking to Dosu, mention @dosu.