Langflow Node that extracts data using RegEx

regex_extract.py

from typing import Optional, Union, List
import re
from langflow.custom import CustomComponent
from langflow.field_typing import Text
from langflow.schema import Record

class ExtractWithRegexComponent(CustomComponent):
    display_name = "Extract with Regex"
    description = "Extracts text from input based on a specified regular expression."

    field_config = {
        "input_text": {
            "display_name": "Input Text",
            "info": "The text from which to extract the pattern.",
        },
        "regex_pattern": {
            "display_name": "Regex Pattern",
            "info": "The regular expression pattern to use for extraction.",
        },
        "output_type": {
            "display_name": "Output Type",
            "info": "The type of output to return: 'text' or 'record'.",
            "input_types": ["text", "record"],
        },
    }

    def build(
        self,
        input_text: Text,
        regex_pattern: Text,
        output_type: str = "text",
    ) -> Union[Text, Record]:
        if not input_text or not regex_pattern:
            raise ValueError("Both 'input_text' and 'regex_pattern' must be provided and non-empty.")

        # Compile the regex pattern
        pattern = re.compile(regex_pattern)
        match = pattern.search(input_text)

        if match:
            extracted_text = match.group(1) if match.groups() else match.group(0)
            self.status = extracted_text
            if output_type == "record":
                return Record(data={"extracted_text": extracted_text})
            else:
                return extracted_text
        else:
            self.status = "No match found."
            if output_type == "record":
                return Record(data={"extracted_text": "No match found."})
            else:
                return "No match found."