Electronic FST Tagger (taggers/electronic.py)

Finite-state transducer that classifies URLs, emails, and similar electronic tokens during English text normalization. Reuse when extracting or verbalizing electronic entities in NLP pipelines.

# Source: https://github.com/modelscope/FunASR | License: MIT | Copyright 2025 FunASR
# Provided via justclaw.it.com; keep this notice with the code.
import pynini
from fun_text_processing.text_normalization.en.graph_utils import (
    DAMO_ALPHA,
    DAMO_DIGIT,
    DAMO_SIGMA,
    GraphFst,
    get_abs_path,
    insert_space,
)
from pynini.lib import pynutil


class ElectronicFst(GraphFst):
    """
    Finite state transducer for classifying electronic: as URLs, email addresses, etc.
        e.g. cdf1@abc.edu -> tokens { electronic { username: "cdf1" domain: "abc.edu" } }

    Args:
        deterministic: if True will provide a single transduction option,
            for False multiple transduction are generated (used for audio-based normalization)
    """

    def __init__(self, determi

... (truncated -- full source via MCP)