chore: import upstream snapshot with attribution

fun_text_processing/README.md

@@ -0,0 +1,32 @@
+**Fundamental Text Processing (FunTextProcessing)**
+==========================
+
+### Introduction
+
+FunTextProcessing is a Python toolkit for fundamental text processing in ASR including text processing , inverse text processing, num2words, which is included in the `FunASR`.
+
+### Highlights
+
+- FunTextProcessing supports inverse text processing (ITN), text processing (TN), number to words (num2words).
+- FunTextProcessing supports multilingual, 10+ languages for ITN, 5 languages for TN, 50+ languages for num2words.
+
+
+### Example
+#### Inverse Text Processing (ITN)
+Given text inputs, such as speech recognition results, use `fun_text_processing/inverse_text_normalization/inverse_normalize.py` to output ITN results. You may refer to the following example scripts.
+
+```
+test_file=fun_text_processing/inverse_text_normalization/id/id_itn_test_input.txt
+
+python fun_text_processing/inverse_text_normalization/inverse_normalize.py --input_file $test_file --cache_dir ./itn_model/ --output_file output.txt --language=id
+```
+
+
+### Acknowledge
+1. We borrowed a lot of codes from [NeMo](https://github.com/NVIDIA/NeMo).
+2. We refered the codes from [WeTextProcessing](https://github.com/wenet-e2e/WeTextProcessing) for Chinese inverse text normalization. 
+3. We borrowed a lot of codes from  [num2words](https://pypi.org/project/num2words/) library for convert the number to words function in some languages.
+
+### License
+
+This project is licensed under the [The MIT License](https://opensource.org/licenses/MIT). FunTextProcessing also contains various third-party components and some code modified from other repos under other open source licenses. 

fun_text_processing/__init__.py

@@ -0,0 +1 @@
+

fun_text_processing/install_pynini.sh

@@ -0,0 +1,6 @@
+#!/bin/bash
+if [[ $OSTYPE == 'darwin'* ]]; then
+  conda install -c conda-forge -y pynini=2.1.5
+else
+  pip install pynini==2.1.5
+fi

fun_text_processing/inverse_text_normalization/__init__.py

@@ -0,0 +1 @@
+

fun_text_processing/inverse_text_normalization/de/__init__.py

@@ -0,0 +1,7 @@
+from fun_text_processing.inverse_text_normalization.en.taggers.tokenize_and_classify import (
+    ClassifyFst,
+)
+from fun_text_processing.inverse_text_normalization.en.verbalizers.verbalize import VerbalizeFst
+from fun_text_processing.inverse_text_normalization.en.verbalizers.verbalize_final import (
+    VerbalizeFinalFst,
+)

fun_text_processing/inverse_text_normalization/de/taggers/cardinal.py

@@ -0,0 +1,65 @@
+import pynini
+from fun_text_processing.text_normalization.en.graph_utils import DAMO_SIGMA, GraphFst
+from pynini.lib import pynutil
+
+
+class CardinalFst(GraphFst):
+    """
+    Finite state transducer for classifying cardinals. Numbers below ten are not converted.
+    Allows both compound numeral strings or separated by whitespace.
+    "und" (en: "and") can be inserted between "hundert" and following number or "tausend" and following single or double digit number.
+
+        e.g. minus drei und zwanzig -> cardinal { negative: "-" integer: "23" } }
+        e.g. minus dreiundzwanzig -> cardinal { integer: "23" } }
+        e.g. dreizehn -> cardinal { integer: "13" } }
+        e.g. ein hundert -> cardinal { integer: "100" } }
+        e.g. einhundert -> cardinal { integer: "100" } }
+        e.g. ein tausend -> cardinal { integer: "1000" } }
+        e.g. eintausend -> cardinal { integer: "1000" } }
+        e.g. ein tausend zwanzig -> cardinal { integer: "1020" } }
+
+    Args:
+        tn_cardinal_tagger: TN cardinal tagger
+    """
+
+    def __init__(self, tn_cardinal_tagger: GraphFst, deterministic: bool = True):
+        super().__init__(name="cardinal", kind="classify", deterministic=deterministic)
+
+        # add_space_between_chars = pynini.cdrewrite(pynini.closure(insert_space, 0, 1), DAMO_CHAR, DAMO_CHAR, DAMO_SIGMA)
+        optional_delete_space = pynini.closure(DAMO_SIGMA | pynutil.delete(" "))
+
+        graph = (tn_cardinal_tagger.graph @ optional_delete_space).invert().optimize()
+        self.graph_hundred_component_at_least_one_none_zero_digit = (
+            (
+                tn_cardinal_tagger.graph_hundred_component_at_least_one_none_zero_digit
+                @ optional_delete_space
+            )
+            .invert()
+            .optimize()
+        )
+
+        self.graph_ties = (
+            (tn_cardinal_tagger.two_digit_non_zero @ optional_delete_space).invert().optimize()
+        )
+        # this is to make sure if there is an ambiguity with decimal, decimal is chosen, e.g. 1000000 vs. 1 million
+        graph = pynutil.add_weight(graph, weight=0.001)
+        self.graph_no_exception = graph
+        self.digit = (
+            pynini.arcmap(tn_cardinal_tagger.digit, map_type="rmweight").invert().optimize()
+        )
+        graph_exception = pynini.project(self.digit, "input")
+        self.graph = (pynini.project(graph, "input") - graph_exception.arcsort()) @ graph
+
+        self.optional_minus_graph = pynini.closure(
+            pynutil.insert("negative: ") + pynini.cross("minus ", '"-" '), 0, 1
+        )
+
+        final_graph = (
+            self.optional_minus_graph
+            + pynutil.insert('integer: "')
+            + self.graph
+            + pynutil.insert('"')
+        )
+
+        final_graph = self.add_tokens(final_graph)
+        self.fst = final_graph.optimize()

fun_text_processing/inverse_text_normalization/de/taggers/date.py

@@ -0,0 +1,84 @@
+import pynini
+from fun_text_processing.text_normalization.en.graph_utils import (
+    DAMO_DIGIT,
+    DAMO_NOT_QUOTE,
+    DAMO_SIGMA,
+    GraphFst,
+    convert_space,
+)
+from pynini.lib import pynutil
+
+
+class DateFst(GraphFst):
+    """
+    Finite state transducer for classifying date, in the form of (day) month (year) or year
+        e.g. vierundzwanzigster juli zwei tausend dreizehn -> tokens { name: "24. Jul. 2013" }
+        e.g. neunzehnachtzig -> tokens { name: "1980" }
+        e.g. vierzehnter januar -> tokens { name: "14. Jan." }
+        e.g. zweiter dritter -> tokens { name: "02.03." }
+        e.g. januar neunzehnachtzig -> tokens { name: "Jan. 1980" }
+        e.g. zwanzigzwanzig -> tokens { name: "2020" }
+
+    Args:
+        itn_cardinal_tagger: ITN cardinal tagger
+        tn_date_tagger: TN date tagger
+        tn_date_verbalizer: TN date verbalizer
+    """
+
+    def __init__(
+        self,
+        itn_cardinal_tagger: GraphFst,
+        tn_date_tagger: GraphFst,
+        tn_date_verbalizer: GraphFst,
+        deterministic: bool = True,
+    ):
+        super().__init__(name="date", kind="classify", deterministic=deterministic)
+
+        add_leading_zero_to_double_digit = (DAMO_DIGIT + DAMO_DIGIT) | (
+            pynutil.insert("0") + DAMO_DIGIT
+        )
+        optional_delete_space = pynini.closure(DAMO_SIGMA | pynutil.delete(" ", weight=0.0001))
+        tagger = tn_date_verbalizer.graph.invert().optimize()
+
+        delete_day_marker = (
+            pynutil.delete('day: "') + pynini.closure(DAMO_NOT_QUOTE, 1) + pynutil.delete('"')
+        ) @ itn_cardinal_tagger.graph_no_exception
+
+        month_as_number = (
+            pynutil.delete('month: "')
+            + itn_cardinal_tagger.graph_no_exception
+            + pynutil.delete('"')
+        )
+        month_as_string = (
+            pynutil.delete('month: "') + tn_date_tagger.month_abbr.invert() + pynutil.delete('"')
+        )
+
+        convert_year = (tn_date_tagger.year @ optional_delete_space).invert().optimize()
+        delete_year_marker = (
+            pynutil.delete('year: "') + pynini.closure(DAMO_NOT_QUOTE, 1) + pynutil.delete('"')
+        ) @ convert_year
+
+        # day. month as string (year)
+        verbalizer = (
+            pynini.closure(delete_day_marker + pynutil.insert(".") + pynini.accep(" "), 0, 1)
+            + month_as_string
+            + pynini.closure(pynini.accep(" ") + delete_year_marker, 0, 1)
+        )
+
+        # day. month as number (year)
+        verbalizer |= (
+            delete_day_marker @ add_leading_zero_to_double_digit
+            + pynutil.insert(".")
+            + pynutil.delete(" ")
+            + month_as_number @ add_leading_zero_to_double_digit
+            + pynutil.insert(".")
+            + pynini.closure(pynutil.delete(" ") + delete_year_marker, 0, 1)
+        )
+
+        # year
+        verbalizer |= delete_year_marker
+
+        final_graph = tagger @ verbalizer
+
+        graph = pynutil.insert('name: "') + convert_space(final_graph) + pynutil.insert('"')
+        self.fst = graph.optimize()

fun_text_processing/inverse_text_normalization/de/taggers/decimal.py

@@ -0,0 +1,52 @@
+import pynini
+from fun_text_processing.text_normalization.de.taggers.decimal import get_quantity, quantities
+from fun_text_processing.text_normalization.en.graph_utils import DAMO_SIGMA, GraphFst
+from pynini.lib import pynutil
+
+
+class DecimalFst(GraphFst):
+    """
+    Finite state transducer for classifying decimal
+        e.g. minus elf komma zwei null null sechs billionen -> decimal { negative: "true" integer_part: "11"  fractional_part: "2006" quantity: "billionen" }
+        e.g. eine billion -> decimal { integer_part: "1" quantity: "billion" }
+    Args:
+        itn_cardinal_tagger: ITN Cardinal tagger
+        tn_decimal_tagger: TN decimal tagger
+    """
+
+    def __init__(
+        self, itn_cardinal_tagger: GraphFst, tn_decimal_tagger: GraphFst, deterministic: bool = True
+    ):
+        super().__init__(name="decimal", kind="classify", deterministic=deterministic)
+
+        self.graph = tn_decimal_tagger.graph.invert().optimize()
+
+        delete_point = pynutil.delete(" komma")
+
+        allow_spelling = pynini.cdrewrite(
+            pynini.cross("eine ", "eins ") + quantities, "[BOS]", "[EOS]", DAMO_SIGMA
+        )
+
+        graph_fractional = pynutil.insert('fractional_part: "') + self.graph + pynutil.insert('"')
+        graph_integer = (
+            pynutil.insert('integer_part: "')
+            + itn_cardinal_tagger.graph_no_exception
+            + pynutil.insert('"')
+        )
+        final_graph_wo_sign = graph_integer + delete_point + pynini.accep(" ") + graph_fractional
+
+        self.final_graph_wo_negative = (
+            allow_spelling
+            @ (
+                final_graph_wo_sign
+                | get_quantity(
+                    final_graph_wo_sign,
+                    itn_cardinal_tagger.graph_hundred_component_at_least_one_none_zero_digit,
+                )
+            ).optimize()
+        )
+
+        final_graph = itn_cardinal_tagger.optional_minus_graph + self.final_graph_wo_negative
+        final_graph += pynutil.insert(" preserve_order: true")
+        final_graph = self.add_tokens(final_graph)
+        self.fst = final_graph.optimize()

fun_text_processing/inverse_text_normalization/de/taggers/electronic.py

@@ -0,0 +1,29 @@
+import pynini
+from fun_text_processing.text_normalization.en.graph_utils import GraphFst
+from pynini.lib import pynutil
+
+
+class ElectronicFst(GraphFst):
+    """
+    Finite state transducer for classifying electronic: email addresses, etc.
+        e.g. c d f eins at a b c punkt e d u -> tokens { name: "cdf1.abc.edu" }
+
+    Args:
+        tn_electronic_tagger: TN eletronic tagger
+        tn_electronic_verbalizer: TN eletronic verbalizer
+    """
+
+    def __init__(
+        self,
+        tn_electronic_tagger: GraphFst,
+        tn_electronic_verbalizer: GraphFst,
+        deterministic: bool = True,
+    ):
+        super().__init__(name="electronic", kind="classify", deterministic=deterministic)
+
+        tagger = pynini.invert(tn_electronic_verbalizer.graph).optimize()
+        verbalizer = pynini.invert(tn_electronic_tagger.graph).optimize()
+        final_graph = tagger @ verbalizer
+
+        graph = pynutil.insert('name: "') + final_graph + pynutil.insert('"')
+        self.fst = graph.optimize()

fun_text_processing/inverse_text_normalization/de/taggers/fraction.py

@@ -0,0 +1,66 @@
+import pynini
+from fun_text_processing.text_normalization.en.graph_utils import (
+    DAMO_NOT_QUOTE,
+    GraphFst,
+    convert_space,
+    delete_space,
+)
+from pynini.lib import pynutil
+
+
+class FractionFst(GraphFst):
+    """
+    Finite state transducer for classifying fraction
+        e.g. ein halb -> tokens { name: "1/2" }
+        e.g. ein ein halb -> tokens { name: "1 1/2" }
+        e.g. drei zwei ein hundertstel -> tokens { name: "3 2/100" }
+
+    Args:
+        itn_cardinal_tagger: ITN cardinal tagger
+        tn_fraction_verbalizer: TN fraction verbalizer
+    """
+
+    def __init__(
+        self,
+        itn_cardinal_tagger: GraphFst,
+        tn_fraction_verbalizer: GraphFst,
+        deterministic: bool = True,
+    ):
+        super().__init__(name="fraction", kind="classify", deterministic=deterministic)
+        tagger = tn_fraction_verbalizer.graph.invert().optimize()
+
+        delete_optional_sign = pynini.closure(
+            pynutil.delete("negative: ") + pynini.cross('"true" ', "-"), 0, 1
+        )
+        delete_integer_marker = (
+            pynutil.delete('integer_part: "')
+            + pynini.closure(DAMO_NOT_QUOTE, 1)
+            + pynutil.delete('"')
+        ) @ itn_cardinal_tagger.graph_no_exception
+
+        delete_numerator_marker = (
+            pynutil.delete('numerator: "') + pynini.closure(DAMO_NOT_QUOTE, 1) + pynutil.delete('"')
+        ) @ itn_cardinal_tagger.graph_no_exception
+
+        delete_denominator_marker = (
+            pynutil.insert("/")
+            + (
+                pynutil.delete('denominator: "')
+                + pynini.closure(DAMO_NOT_QUOTE, 1)
+                + pynutil.delete('"')
+            )
+            @ itn_cardinal_tagger.graph_no_exception
+        )
+
+        graph = (
+            pynini.closure(delete_integer_marker + pynini.accep(" "), 0, 1)
+            + delete_numerator_marker
+            + delete_space
+            + delete_denominator_marker
+        ).optimize()
+        verbalizer = delete_optional_sign + graph
+
+        self.graph = tagger @ verbalizer
+
+        graph = pynutil.insert('name: "') + convert_space(self.graph) + pynutil.insert('"')
+        self.fst = graph.optimize()

fun_text_processing/inverse_text_normalization/de/taggers/measure.py

@@ -0,0 +1,98 @@
+import pynini
+from fun_text_processing.text_normalization.de.taggers.measure import (
+    singular_to_plural,
+    unit_singular,
+)
+from fun_text_processing.text_normalization.en.graph_utils import (
+    DAMO_SIGMA,
+    GraphFst,
+    convert_space,
+    delete_extra_space,
+    delete_space,
+)
+from pynini.lib import pynutil
+
+
+class MeasureFst(GraphFst):
+    """
+    Finite state transducer for classifying measure. Allows for plural form for unit.
+        e.g. minus elf kilogramm -> measure { cardinal { negative: "true" integer: "11" } units: "kg" }
+        e.g. drei stunden -> measure { cardinal { integer: "3" } units: "h" }
+        e.g. ein halb kilogramm -> measure { decimal { integer_part: "1/2" } units: "kg" }
+        e.g. eins komma zwei kilogramm -> measure { decimal { integer_part: "1" fractional_part: "2" } units: "kg" }
+
+    Args:
+        itn_cardinal_tagger: ITN Cardinal tagger
+        itn_decimal_tagger: ITN Decimal tagger
+        itn_fraction_tagger: ITN Fraction tagger
+    """
+
+    def __init__(
+        self,
+        itn_cardinal_tagger: GraphFst,
+        itn_decimal_tagger: GraphFst,
+        itn_fraction_tagger: GraphFst,
+        deterministic: bool = True,
+    ):
+        super().__init__(name="measure", kind="classify", deterministic=deterministic)
+
+        cardinal_graph = (
+            pynini.cdrewrite(
+                pynini.cross(pynini.union("ein", "eine"), "eins"), "[BOS]", "[EOS]", DAMO_SIGMA
+            )
+            @ itn_cardinal_tagger.graph_no_exception
+        )
+
+        graph_unit_singular = pynini.invert(unit_singular)  # singular -> abbr
+        unit = (
+            pynini.invert(singular_to_plural()) @ graph_unit_singular
+        ) | graph_unit_singular  # plural -> abbr
+        unit = convert_space(unit)
+        graph_unit_singular = convert_space(graph_unit_singular)
+
+        optional_graph_negative = pynini.closure(
+            pynutil.insert("negative: ") + pynini.cross("minus", '"true"') + delete_extra_space,
+            0,
+            1,
+        )
+
+        unit_misc = pynutil.insert("/") + pynutil.delete("pro") + delete_space + graph_unit_singular
+
+        unit = (
+            pynutil.insert('units: "')
+            + (unit | unit_misc | pynutil.add_weight(unit + delete_space + unit_misc, 0.01))
+            + pynutil.insert('"')
+        )
+
+        subgraph_decimal = (
+            pynutil.insert("decimal { ")
+            + optional_graph_negative
+            + itn_decimal_tagger.final_graph_wo_negative
+            + pynutil.insert(" }")
+            + delete_extra_space
+            + unit
+        )
+
+        subgraph_fraction = (
+            pynutil.insert("decimal { ")
+            + optional_graph_negative
+            + pynutil.insert('integer_part: "')
+            + itn_fraction_tagger.graph
+            + pynutil.insert('" }')
+            + delete_extra_space
+            + unit
+        )
+
+        subgraph_cardinal = (
+            pynutil.insert("cardinal { ")
+            + optional_graph_negative
+            + pynutil.insert('integer: "')
+            + cardinal_graph
+            + pynutil.insert('"')
+            + pynutil.insert(" }")
+            + delete_extra_space
+            + unit
+        )
+        final_graph = subgraph_cardinal | subgraph_decimal | subgraph_fraction
+        final_graph = self.add_tokens(final_graph)
+        self.fst = final_graph.optimize()

fun_text_processing/inverse_text_normalization/de/taggers/money.py

@@ -0,0 +1,90 @@
+import pynini
+from fun_text_processing.text_normalization.de.taggers.money import (
+    maj_singular,
+    min_plural,
+    min_singular,
+)
+from fun_text_processing.text_normalization.en.graph_utils import (
+    DAMO_DIGIT,
+    DAMO_SIGMA,
+    GraphFst,
+    convert_space,
+    delete_extra_space,
+    delete_space,
+    insert_space,
+)
+from pynini.lib import pynutil
+
+
+class MoneyFst(GraphFst):
+    """
+    Finite state transducer for classifying money
+        e.g. elf euro und vier cent -> money { integer_part: "11" fractional_part: 04 currency: "€" }
+
+    Args:
+        itn_cardinal_tagger: ITN Cardinal Tagger
+        itn_decimal_tagger: ITN Decimal Tagger
+    """
+
+    def __init__(
+        self,
+        itn_cardinal_tagger: GraphFst,
+        itn_decimal_tagger: GraphFst,
+        deterministic: bool = True,
+    ):
+        super().__init__(name="money", kind="classify", deterministic=deterministic)
+        cardinal_graph = (
+            pynini.cdrewrite(
+                pynini.cross(pynini.union("ein", "eine"), "eins"), "[BOS]", "[EOS]", DAMO_SIGMA
+            )
+            @ itn_cardinal_tagger.graph_no_exception
+        )
+        graph_decimal_final = itn_decimal_tagger.final_graph_wo_negative
+
+        graph_unit = pynini.invert(maj_singular)
+        graph_unit = pynutil.insert('currency: "') + convert_space(graph_unit) + pynutil.insert('"')
+
+        add_leading_zero_to_double_digit = (DAMO_DIGIT + DAMO_DIGIT) | (
+            pynutil.insert("0") + DAMO_DIGIT
+        )
+        min_unit = pynini.project(min_singular | min_plural, "output")
+        # elf euro (und) vier cent, vier cent
+        cents_standalone = (
+            pynutil.insert('fractional_part: "')
+            + cardinal_graph @ add_leading_zero_to_double_digit
+            + delete_space
+            + pynutil.delete(min_unit)
+            + pynutil.insert('"')
+        )
+
+        optional_cents_standalone = pynini.closure(
+            delete_space
+            + pynini.closure(pynutil.delete("und") + delete_space, 0, 1)
+            + insert_space
+            + cents_standalone,
+            0,
+            1,
+        )
+        # elf euro vierzig, only after integer
+        optional_cents_suffix = pynini.closure(
+            delete_extra_space
+            + pynutil.insert('fractional_part: "')
+            + pynutil.add_weight(cardinal_graph @ add_leading_zero_to_double_digit, -0.7)
+            + pynutil.insert('"'),
+            0,
+            1,
+        )
+
+        graph_integer = (
+            pynutil.insert('integer_part: "')
+            + cardinal_graph
+            + pynutil.insert('"')
+            + delete_extra_space
+            + graph_unit
+            + (optional_cents_standalone | optional_cents_suffix)
+        )
+        graph_decimal = graph_decimal_final + delete_extra_space + graph_unit
+        graph_decimal |= pynutil.insert('currency: "€" integer_part: "0" ') + cents_standalone
+        final_graph = graph_integer | graph_decimal
+        final_graph = self.add_tokens(final_graph)
+        self.fst = final_graph.optimize()

fun_text_processing/inverse_text_normalization/de/taggers/ordinal.py

@@ -0,0 +1,33 @@
+import pynini
+from fun_text_processing.text_normalization.en.graph_utils import DAMO_NOT_QUOTE, GraphFst
+from pynini.lib import pynutil
+
+
+class OrdinalFst(GraphFst):
+    """
+    Finite state transducer for classifying ordinal
+        e.g. dreizehnter -> tokens { name: "13." }
+
+    Args:
+        itn_cardinal_tagger: ITN Cardinal Tagger
+        tn_ordinal_verbalizer: TN Ordinal Verbalizer
+    """
+
+    def __init__(
+        self,
+        itn_cardinal_tagger: GraphFst,
+        tn_ordinal_verbalizer: GraphFst,
+        deterministic: bool = True,
+    ):
+        super().__init__(name="ordinal", kind="classify", deterministic=deterministic)
+
+        tagger = tn_ordinal_verbalizer.graph.invert().optimize()
+
+        graph = (
+            pynutil.delete('integer: "') + pynini.closure(DAMO_NOT_QUOTE, 1) + pynutil.delete('"')
+        ) @ itn_cardinal_tagger.graph
+
+        final_graph = tagger @ graph + pynutil.insert(".")
+
+        graph = pynutil.insert('name: "') + final_graph + pynutil.insert('"')
+        self.fst = graph.optimize()

fun_text_processing/inverse_text_normalization/de/taggers/telephone.py

@@ -0,0 +1,43 @@
+import pynini
+from fun_text_processing.text_normalization.en.graph_utils import (
+    GraphFst,
+    convert_space,
+    insert_space,
+)
+from pynini.lib import pynutil
+
+
+class TelephoneFst(GraphFst):
+    """
+    Finite state transducer for classifying telephone numbers, e.g.
+        null vier eins eins eins zwei drei vier eins zwei drei vier -> tokens { name: "(0411) 1234-1234" }
+
+    Args:
+        tn_cardinal_tagger: TN Cardinal Tagger
+    """
+
+    def __init__(self, tn_cardinal_tagger: GraphFst, deterministic: bool = True):
+        super().__init__(name="telephone", kind="classify", deterministic=deterministic)
+        separator = pynini.accep(" ")  # between components
+        digit = pynini.union(*list(map(str, range(1, 10)))) @ tn_cardinal_tagger.two_digit_non_zero
+        zero = pynini.cross("0", "null")
+
+        number_part = (
+            pynutil.delete("(")
+            + zero
+            + insert_space
+            + pynini.closure(digit + insert_space, 2, 2)
+            + digit
+            + pynutil.delete(")")
+            + separator
+            + pynini.closure(digit + insert_space, 3, 3)
+            + digit
+            + pynutil.delete("-")
+            + insert_space
+            + pynini.closure(digit + insert_space, 3, 3)
+            + digit
+        )
+        graph = convert_space(pynini.invert(number_part))
+        final_graph = pynutil.insert('name: "') + graph + pynutil.insert('"')
+
+        self.fst = final_graph.optimize()

fun_text_processing/inverse_text_normalization/de/taggers/time.py

@@ -0,0 +1,28 @@
+import pynini
+from fun_text_processing.text_normalization.en.graph_utils import DAMO_SIGMA, GraphFst
+from pynini.lib import pynutil
+
+
+class TimeFst(GraphFst):
+    """
+    Finite state transducer for classifying time
+        e.g. acht uhr e s t-> time { hours: "8" zone: "e s t" }
+        e.g. dreizehn uhr -> time { hours: "13" }
+        e.g. dreizehn uhr zehn -> time { hours: "13" minutes: "10" }
+        e.g. viertel vor zwölf -> time { minutes: "45" hours: "11" }
+        e.g. viertel nach zwölf -> time { minutes: "15" hours: "12" }
+        e.g. halb zwölf -> time { minutes: "30" hours: "11" }
+        e.g. drei vor zwölf -> time { minutes: "57" hours: "11" }
+        e.g. drei nach zwölf -> time { minutes: "3" hours: "12" }
+        e.g. drei uhr zehn minuten zehn sekunden -> time { hours: "3" hours: "10" sekunden: "10"}
+
+    Args:
+        tn_time_verbalizer: TN time verbalizer
+    """
+
+    def __init__(self, tn_time_verbalizer: GraphFst, deterministic: bool = True):
+        super().__init__(name="time", kind="classify", deterministic=deterministic)
+        # lazy way to make sure compounds work
+        optional_delete_space = pynini.closure(DAMO_SIGMA | pynutil.delete(" ", weight=0.0001))
+        graph = (tn_time_verbalizer.graph @ optional_delete_space).invert().optimize()
+        self.fst = self.add_tokens(graph).optimize()

fun_text_processing/inverse_text_normalization/de/taggers/tokenize_and_classify.py

@@ -0,0 +1,162 @@
+import os
+
+import pynini
+from fun_text_processing.inverse_text_normalization.de.taggers.cardinal import CardinalFst
+from fun_text_processing.inverse_text_normalization.de.taggers.date import DateFst
+from fun_text_processing.inverse_text_normalization.de.taggers.decimal import DecimalFst
+from fun_text_processing.inverse_text_normalization.de.taggers.electronic import ElectronicFst
+from fun_text_processing.inverse_text_normalization.de.taggers.fraction import FractionFst
+from fun_text_processing.inverse_text_normalization.de.taggers.measure import MeasureFst
+from fun_text_processing.inverse_text_normalization.de.taggers.money import MoneyFst
+from fun_text_processing.inverse_text_normalization.de.taggers.ordinal import OrdinalFst
+from fun_text_processing.inverse_text_normalization.de.taggers.telephone import TelephoneFst
+from fun_text_processing.inverse_text_normalization.de.taggers.time import TimeFst
+from fun_text_processing.inverse_text_normalization.de.taggers.whitelist import WhiteListFst
+from fun_text_processing.inverse_text_normalization.en.taggers.punctuation import PunctuationFst
+from fun_text_processing.inverse_text_normalization.en.taggers.word import WordFst
+from fun_text_processing.text_normalization.de.taggers.cardinal import (
+    CardinalFst as TNCardinalTagger,
+)
+from fun_text_processing.text_normalization.de.taggers.date import DateFst as TNDateTagger
+from fun_text_processing.text_normalization.de.taggers.decimal import DecimalFst as TNDecimalTagger
+from fun_text_processing.text_normalization.de.taggers.electronic import (
+    ElectronicFst as TNElectronicTagger,
+)
+from fun_text_processing.text_normalization.de.taggers.whitelist import (
+    WhiteListFst as TNWhitelistTagger,
+)
+from fun_text_processing.text_normalization.de.verbalizers.date import DateFst as TNDateVerbalizer
+from fun_text_processing.text_normalization.de.verbalizers.electronic import (
+    ElectronicFst as TNElectronicVerbalizer,
+)
+from fun_text_processing.text_normalization.de.verbalizers.fraction import (
+    FractionFst as TNFractionVerbalizer,
+)
+from fun_text_processing.text_normalization.de.verbalizers.ordinal import (
+    OrdinalFst as TNOrdinalVerbalizer,
+)
+from fun_text_processing.text_normalization.de.verbalizers.time import TimeFst as TNTimeVerbalizer
+from fun_text_processing.text_normalization.en.graph_utils import (
+    GraphFst,
+    delete_extra_space,
+    delete_space,
+    generator_main,
+)
+from pynini.lib import pynutil
+
+import logging
+
+
+class ClassifyFst(GraphFst):
+    """
+    Final class that composes all other classification grammars. This class can process an entire sentence, that is lower cased.
+    For deployment, this grammar will be compiled and exported to OpenFst Finate State Archiv (FAR) File.
+    More details to deployment at NeMo/tools/text_processing_deployment.
+
+    Args:
+        cache_dir: path to a dir with .far grammar file. Set to None to avoid using cache.
+        overwrite_cache: set to True to overwrite .far files
+    """
+
+    def __init__(
+        self, cache_dir: str = None, overwrite_cache: bool = False, deterministic: bool = True
+    ):
+        super().__init__(name="tokenize_and_classify", kind="classify", deterministic=deterministic)
+
+        far_file = None
+        if cache_dir is not None and cache_dir != "None":
+            os.makedirs(cache_dir, exist_ok=True)
+            far_file = os.path.join(cache_dir, "_de_itn.far")
+        if not overwrite_cache and far_file and os.path.exists(far_file):
+            self.fst = pynini.Far(far_file, mode="r")["tokenize_and_classify"]
+            logging.info(f"ClassifyFst.fst was restored from {far_file}.")
+        else:
+            logging.info(f"Creating ClassifyFst grammars.")
+            tn_cardinal_tagger = TNCardinalTagger(deterministic=False)
+            tn_date_tagger = TNDateTagger(cardinal=tn_cardinal_tagger, deterministic=False)
+            tn_decimal_tagger = TNDecimalTagger(cardinal=tn_cardinal_tagger, deterministic=False)
+            tn_ordinal_verbalizer = TNOrdinalVerbalizer(deterministic=False)
+            tn_fraction_verbalizer = TNFractionVerbalizer(
+                ordinal=tn_ordinal_verbalizer, deterministic=False
+            )
+            tn_time_verbalizer = TNTimeVerbalizer(
+                cardinal_tagger=tn_cardinal_tagger, deterministic=False
+            )
+            tn_date_verbalizer = TNDateVerbalizer(
+                ordinal=tn_ordinal_verbalizer, deterministic=False
+            )
+            tn_electronic_tagger = TNElectronicTagger(deterministic=False)
+            tn_electronic_verbalizer = TNElectronicVerbalizer(deterministic=False)
+            tn_whitelist_tagger = TNWhitelistTagger(input_case="cased", deterministic=False)
+
+            cardinal = CardinalFst(tn_cardinal_tagger=tn_cardinal_tagger)
+            cardinal_graph = cardinal.fst
+
+            ordinal = OrdinalFst(
+                itn_cardinal_tagger=cardinal, tn_ordinal_verbalizer=tn_ordinal_verbalizer
+            )
+            ordinal_graph = ordinal.fst
+            decimal = DecimalFst(itn_cardinal_tagger=cardinal, tn_decimal_tagger=tn_decimal_tagger)
+            decimal_graph = decimal.fst
+
+            fraction = FractionFst(
+                itn_cardinal_tagger=cardinal, tn_fraction_verbalizer=tn_fraction_verbalizer
+            )
+            fraction_graph = fraction.fst
+
+            measure_graph = MeasureFst(
+                itn_cardinal_tagger=cardinal,
+                itn_decimal_tagger=decimal,
+                itn_fraction_tagger=fraction,
+            ).fst
+            date_graph = DateFst(
+                itn_cardinal_tagger=cardinal,
+                tn_date_verbalizer=tn_date_verbalizer,
+                tn_date_tagger=tn_date_tagger,
+            ).fst
+            word_graph = WordFst().fst
+            time_graph = TimeFst(tn_time_verbalizer=tn_time_verbalizer).fst
+            money_graph = MoneyFst(itn_cardinal_tagger=cardinal, itn_decimal_tagger=decimal).fst
+            whitelist_graph = WhiteListFst(tn_whitelist_tagger=tn_whitelist_tagger).fst
+            punct_graph = PunctuationFst().fst
+            electronic_graph = ElectronicFst(
+                tn_electronic_tagger=tn_electronic_tagger,
+                tn_electronic_verbalizer=tn_electronic_verbalizer,
+            ).fst
+            telephone_graph = TelephoneFst(tn_cardinal_tagger=tn_cardinal_tagger).fst
+
+            classify = (
+                pynutil.add_weight(cardinal_graph, 1.1)
+                | pynutil.add_weight(whitelist_graph, 1.0)
+                | pynutil.add_weight(time_graph, 1.1)
+                | pynutil.add_weight(date_graph, 1.1)
+                | pynutil.add_weight(decimal_graph, 1.1)
+                | pynutil.add_weight(measure_graph, 1.1)
+                | pynutil.add_weight(ordinal_graph, 1.1)
+                | pynutil.add_weight(fraction_graph, 1.1)
+                | pynutil.add_weight(money_graph, 1.1)
+                | pynutil.add_weight(telephone_graph, 1.1)
+                | pynutil.add_weight(electronic_graph, 1.1)
+                | pynutil.add_weight(word_graph, 100)
+            )
+
+            punct = (
+                pynutil.insert("tokens { ")
+                + pynutil.add_weight(punct_graph, weight=1.1)
+                + pynutil.insert(" }")
+            )
+            token = pynutil.insert("tokens { ") + classify + pynutil.insert(" }")
+            token_plus_punct = (
+                pynini.closure(punct + pynutil.insert(" "))
+                + token
+                + pynini.closure(pynutil.insert(" ") + punct)
+            )
+
+            graph = token_plus_punct + pynini.closure(delete_extra_space + token_plus_punct)
+            graph = delete_space + graph + delete_space
+
+            self.fst = graph.optimize()
+
+            if far_file:
+                generator_main(far_file, {"tokenize_and_classify": self.fst})
+                logging.info(f"ClassifyFst grammars are saved to {far_file}.")

fun_text_processing/inverse_text_normalization/de/taggers/whitelist.py

@@ -0,0 +1,19 @@
+import pynini
+from fun_text_processing.text_normalization.en.graph_utils import GraphFst, convert_space
+from pynini.lib import pynutil
+
+
+class WhiteListFst(GraphFst):
+    """
+    Finite state transducer for classifying whitelisted tokens
+        e.g. misses -> tokens { name: "Mrs." }
+    Args:
+        tn_whitelist_tagger: TN whitelist tagger
+    """
+
+    def __init__(self, tn_whitelist_tagger: GraphFst, deterministic: bool = True):
+        super().__init__(name="whitelist", kind="classify", deterministic=deterministic)
+
+        whitelist = pynini.invert(tn_whitelist_tagger.graph)
+        graph = pynutil.insert('name: "') + convert_space(whitelist) + pynutil.insert('"')
+        self.fst = graph.optimize()

fun_text_processing/inverse_text_normalization/de/verbalizers/__init__.py

@@ -0,0 +1 @@
+

fun_text_processing/inverse_text_normalization/de/verbalizers/cardinal.py

@@ -0,0 +1,23 @@
+import pynini
+from fun_text_processing.text_normalization.en.graph_utils import DAMO_NOT_QUOTE, GraphFst
+from pynini.lib import pynutil
+
+
+class CardinalFst(GraphFst):
+    """
+    Finite state transducer for verbalizing cardinal
+        e.g. cardinal { integer: "23" negative: "-" } -> -23
+
+    Args:
+        tn_cardinal_verbalizer: TN cardinal verbalizer
+    """
+
+    def __init__(self, tn_cardinal_verbalizer: GraphFst, deterministic: bool = True):
+        super().__init__(name="cardinal", kind="verbalize", deterministic=deterministic)
+        self.numbers = tn_cardinal_verbalizer.numbers
+        optional_sign = pynini.closure(
+            pynutil.delete('negative: "') + DAMO_NOT_QUOTE + pynutil.delete('" '), 0, 1
+        )
+        graph = optional_sign + self.numbers
+        delete_tokens = self.delete_tokens(graph)
+        self.fst = delete_tokens.optimize()

fun_text_processing/inverse_text_normalization/de/verbalizers/decimal.py

@@ -0,0 +1,37 @@
+import pynini
+from fun_text_processing.text_normalization.en.graph_utils import (
+    DAMO_NOT_QUOTE,
+    GraphFst,
+    delete_preserve_order,
+)
+from pynini.lib import pynutil
+
+
+class DecimalFst(GraphFst):
+    """
+    Finite state transducer for verbalizing decimal, e.g.
+        decimal { negative: "true" integer_part: "12"  fractional_part: "5006" quantity: "billion" } -> -12.5006 billion
+
+    Args:
+        tn_decimal_verbalizer: TN decimal verbalizer
+    """
+
+    def __init__(self, tn_decimal_verbalizer: GraphFst, deterministic: bool = True):
+        super().__init__(name="decimal", kind="verbalize", deterministic=deterministic)
+        delete_space = pynutil.delete(" ")
+        optional_sign = pynini.closure(
+            pynutil.delete('negative: "') + DAMO_NOT_QUOTE + pynutil.delete('"') + delete_space,
+            0,
+            1,
+        )
+        optional_integer = pynini.closure(tn_decimal_verbalizer.integer, 0, 1)
+        optional_fractional = pynini.closure(
+            delete_space + pynutil.insert(",") + tn_decimal_verbalizer.fractional_default, 0, 1
+        )
+        graph = (
+            optional_integer + optional_fractional + tn_decimal_verbalizer.optional_quantity
+        ).optimize()
+        self.numbers = optional_sign + graph
+        graph = self.numbers + delete_preserve_order
+        delete_tokens = self.delete_tokens(graph)
+        self.fst = delete_tokens.optimize()

fun_text_processing/inverse_text_normalization/de/verbalizers/measure.py

@@ -0,0 +1,50 @@
+import pynini
+from fun_text_processing.text_normalization.en.graph_utils import DAMO_CHAR, GraphFst, delete_space
+from pynini.lib import pynutil
+
+
+class MeasureFst(GraphFst):
+    """
+    Finite state transducer for verbalizing measure, e.g.
+        measure { cardinal { negative: "true" integer: "12" } units: "kg" } -> -12 kg
+        measure { decimal { integer_part: "1/2" } units: "kg" } -> 1/2 kg
+        measure { decimal { integer_part: "1" fractional_part: "2" quantity: "million" } units: "kg" } -> 1,2 million kg
+
+    Args:
+        decimal: ITN Decimal verbalizer
+        cardinal: ITN Cardinal verbalizer
+    """
+
+    def __init__(self, decimal: GraphFst, cardinal: GraphFst, deterministic: bool = True):
+        super().__init__(name="measure", kind="verbalize", deterministic=deterministic)
+        optional_sign = pynini.closure(pynini.cross('negative: "true"', "-"), 0, 1)
+        unit = (
+            pynutil.delete("units:")
+            + delete_space
+            + pynutil.delete('"')
+            + pynini.closure(DAMO_CHAR - " ", 1)
+            + pynutil.delete('"')
+            + delete_space
+        )
+        graph_decimal = (
+            pynutil.delete("decimal {")
+            + delete_space
+            + optional_sign
+            + delete_space
+            + decimal.numbers
+            + delete_space
+            + pynutil.delete("}")
+        )
+        graph_cardinal = (
+            pynutil.delete("cardinal {")
+            + delete_space
+            + optional_sign
+            + delete_space
+            + cardinal.numbers
+            + delete_space
+            + pynutil.delete("}")
+        )
+
+        graph = (graph_cardinal | graph_decimal) + delete_space + pynutil.insert(" ") + unit
+        delete_tokens = self.delete_tokens(graph)
+        self.fst = delete_tokens.optimize()

fun_text_processing/inverse_text_normalization/de/verbalizers/money.py

@@ -0,0 +1,26 @@
+import pynini
+from fun_text_processing.text_normalization.en.graph_utils import DAMO_CHAR, GraphFst, delete_space
+from pynini.lib import pynutil
+
+
+class MoneyFst(GraphFst):
+    """
+    Finite state transducer for verbalizing money, e.g.
+        money { integer_part: "12" fractional_part: "05" currency: "$" } -> $12.05
+
+    Args:
+        decimal: ITN Decimal verbalizer
+    """
+
+    def __init__(self, decimal: GraphFst, deterministic: bool = True):
+        super().__init__(name="money", kind="verbalize", deterministic=deterministic)
+        unit = (
+            pynutil.delete("currency:")
+            + delete_space
+            + pynutil.delete('"')
+            + pynini.closure(DAMO_CHAR - " ", 1)
+            + pynutil.delete('"')
+        )
+        graph = unit + delete_space + decimal.numbers
+        delete_tokens = self.delete_tokens(graph)
+        self.fst = delete_tokens.optimize()

fun_text_processing/inverse_text_normalization/de/verbalizers/time.py

@@ -0,0 +1,52 @@
+import pynini
+from fun_text_processing.text_normalization.en.graph_utils import (
+    DAMO_ALPHA,
+    DAMO_DIGIT,
+    GraphFst,
+    delete_space,
+)
+from pynini.lib import pynutil
+
+
+class TimeFst(GraphFst):
+    """
+    Finite state transducer for verbalizing time, e.g.
+        time { hours: "8" minutes: "30" zone: "e s t" } -> 08:30 Uhr est
+        time { hours: "8" } -> 8 Uhr
+        time { hours: "8" minutes: "30" seconds: "10" } -> 08:30:10 Uhr
+    """
+
+    def __init__(self, deterministic: bool = True):
+        super().__init__(name="time", kind="verbalize", deterministic=deterministic)
+
+        add_leading_zero_to_double_digit = (DAMO_DIGIT + DAMO_DIGIT) | (
+            pynutil.insert("0") + DAMO_DIGIT
+        )
+        hour = pynutil.delete('hours: "') + pynini.closure(DAMO_DIGIT, 1) + pynutil.delete('"')
+        minute = pynutil.delete('minutes: "') + pynini.closure(DAMO_DIGIT, 1) + pynutil.delete('"')
+
+        second = pynutil.delete('seconds: "') + pynini.closure(DAMO_DIGIT, 1) + pynutil.delete('"')
+        zone = (
+            pynutil.delete('zone: "')
+            + pynini.closure(DAMO_ALPHA + delete_space)
+            + DAMO_ALPHA
+            + pynutil.delete('"')
+        )
+        optional_zone = pynini.closure(pynini.accep(" ") + zone, 0, 1)
+        graph = (
+            delete_space
+            + pynutil.insert(":")
+            + (minute @ add_leading_zero_to_double_digit)
+            + pynini.closure(
+                delete_space + pynutil.insert(":") + (second @ add_leading_zero_to_double_digit),
+                0,
+                1,
+            )
+            + pynutil.insert(" Uhr")
+            + optional_zone
+        )
+        graph_h = hour + pynutil.insert(" Uhr") + optional_zone
+        graph_hm = hour @ add_leading_zero_to_double_digit + graph
+        graph_hms = hour @ add_leading_zero_to_double_digit + graph
+        final_graph = graph_hm | graph_hms | graph_h
+        self.fst = self.delete_tokens(final_graph).optimize()

fun_text_processing/inverse_text_normalization/de/verbalizers/verbalize.py

@@ -0,0 +1,35 @@
+from fun_text_processing.inverse_text_normalization.de.verbalizers.cardinal import CardinalFst
+from fun_text_processing.inverse_text_normalization.de.verbalizers.decimal import DecimalFst
+from fun_text_processing.inverse_text_normalization.de.verbalizers.measure import MeasureFst
+from fun_text_processing.inverse_text_normalization.de.verbalizers.money import MoneyFst
+from fun_text_processing.inverse_text_normalization.de.verbalizers.time import TimeFst
+from fun_text_processing.text_normalization.de.verbalizers.cardinal import (
+    CardinalFst as TNCardinalVerbalizer,
+)
+from fun_text_processing.text_normalization.de.verbalizers.decimal import (
+    DecimalFst as TNDecimalVerbalizer,
+)
+from fun_text_processing.text_normalization.en.graph_utils import GraphFst
+
+
+class VerbalizeFst(GraphFst):
+    """
+    Composes other verbalizer grammars.
+    For deployment, this grammar will be compiled and exported to OpenFst Finate State Archiv (FAR) File.
+    More details to deployment at NeMo/tools/text_processing_deployment.
+    """
+
+    def __init__(self, deterministic: bool = True):
+        super().__init__(name="verbalize", kind="verbalize", deterministic=deterministic)
+        tn_cardinal_verbalizer = TNCardinalVerbalizer(deterministic=False)
+        tn_decimal_verbalizer = TNDecimalVerbalizer(deterministic=False)
+
+        cardinal = CardinalFst(tn_cardinal_verbalizer=tn_cardinal_verbalizer)
+        cardinal_graph = cardinal.fst
+        decimal = DecimalFst(tn_decimal_verbalizer=tn_decimal_verbalizer)
+        decimal_graph = decimal.fst
+        measure_graph = MeasureFst(decimal=decimal, cardinal=cardinal).fst
+        money_graph = MoneyFst(decimal=decimal).fst
+        time_graph = TimeFst().fst
+        graph = time_graph | money_graph | measure_graph | decimal_graph | cardinal_graph
+        self.fst = graph

fun_text_processing/inverse_text_normalization/de/verbalizers/verbalize_final.py

@@ -0,0 +1,33 @@
+import pynini
+from fun_text_processing.inverse_text_normalization.de.verbalizers.verbalize import VerbalizeFst
+from fun_text_processing.inverse_text_normalization.en.verbalizers.word import WordFst
+from fun_text_processing.text_normalization.en.graph_utils import (
+    GraphFst,
+    delete_extra_space,
+    delete_space,
+)
+from pynini.lib import pynutil
+
+
+class VerbalizeFinalFst(GraphFst):
+    """
+    Finite state transducer that verbalizes an entire sentence, e.g.
+    tokens { name: "jetzt" } tokens { name: "ist" } tokens { time { hours: "12" minutes: "30" } } -> jetzt ist 12:30 Uhr
+    """
+
+    def __init__(self, deterministic: bool = True):
+        super().__init__(name="verbalize_final", kind="verbalize", deterministic=deterministic)
+        verbalize = VerbalizeFst().fst
+        word = WordFst().fst
+        types = verbalize | word
+        graph = (
+            pynutil.delete("tokens")
+            + delete_space
+            + pynutil.delete("{")
+            + delete_space
+            + types
+            + delete_space
+            + pynutil.delete("}")
+        )
+        graph = delete_space + pynini.closure(graph + delete_extra_space) + graph + delete_space
+        self.fst = graph

fun_text_processing/inverse_text_normalization/en/__init__.py

@@ -0,0 +1,7 @@
+from fun_text_processing.inverse_text_normalization.en.taggers.tokenize_and_classify import (
+    ClassifyFst,
+)
+from fun_text_processing.inverse_text_normalization.en.verbalizers.verbalize import VerbalizeFst
+from fun_text_processing.inverse_text_normalization.en.verbalizers.verbalize_final import (
+    VerbalizeFinalFst,
+)

fun_text_processing/inverse_text_normalization/en/clean_eval_data.py

@@ -0,0 +1,384 @@
+from argparse import ArgumentParser
+from typing import List
+
+import regex as re
+from fun_text_processing.text_normalization.data_loader_utils import (
+    EOS_TYPE,
+    Instance,
+    load_files,
+    training_data_to_sentences,
+)
+
+
+"""
+This file is for evaluation purposes.
+filter_loaded_data() cleans data (list of instances) for inverse text normalization. Filters and cleaners can be specified for each semiotic class individually.
+For example, normalized text should only include characters and whitespace characters but no punctuation. 
+            Cardinal unnormalized instances should contain at least one integer and all other characters are removed.
+"""
+
+
+class Filter:
+    """
+    Filter class
+
+    Args:
+        class_type: semiotic class used in dataset
+        process_func: function to transform text
+        filter_func:  function to filter text
+
+    """
+
+    def __init__(self, class_type: str, process_func: object, filter_func: object):
+        self.class_type = class_type
+        self.process_func = process_func
+        self.filter_func = filter_func
+
+    def filter(self, instance: Instance) -> bool:
+        """
+        filter function
+
+        Args:
+            filters given instance with filter function
+
+        Returns: True if given instance fulfills criteria or does not belong to class type
+        """
+        if instance.token_type != self.class_type:
+            return True
+        return self.filter_func(instance)
+
+    def process(self, instance: Instance) -> Instance:
+        """
+        process function
+
+        Args:
+            processes given instance with process function
+
+        Returns: processed instance if instance belongs to expected class type or original instance
+        """
+        if instance.token_type != self.class_type:
+            return instance
+        return self.process_func(instance)
+
+
+def filter_cardinal_1(instance: Instance) -> bool:
+    ok = re.search(r"[0-9]", instance.un_normalized)
+    return ok
+
+
+def process_cardinal_1(instance: Instance) -> Instance:
+    un_normalized = instance.un_normalized
+    normalized = instance.normalized
+    un_normalized = re.sub(r"[^0-9]", "", un_normalized)
+    normalized = re.sub(r"[^a-z ]", "", normalized)
+    return Instance(
+        token_type=instance.token_type, un_normalized=un_normalized, normalized=normalized
+    )
+
+
+def filter_ordinal_1(instance: Instance) -> bool:
+    ok = re.search(r"(st|nd|rd|th)\s*$", instance.un_normalized)
+    return ok
+
+
+def process_ordinal_1(instance: Instance) -> Instance:
+    un_normalized = instance.un_normalized
+    normalized = instance.normalized
+    un_normalized = re.sub(r"[,\s]", "", un_normalized)
+    normalized = re.sub(r"[^a-z ]", "", normalized)
+    return Instance(
+        token_type=instance.token_type, un_normalized=un_normalized, normalized=normalized
+    )
+
+
+def filter_decimal_1(instance: Instance) -> bool:
+    ok = re.search(r"[0-9]", instance.un_normalized)
+    return ok
+
+
+def process_decimal_1(instance: Instance) -> Instance:
+    un_normalized = instance.un_normalized
+    un_normalized = re.sub(r",", "", un_normalized)
+    normalized = instance.normalized
+    normalized = re.sub(r"[^a-z ]", "", normalized)
+    return Instance(
+        token_type=instance.token_type, un_normalized=un_normalized, normalized=normalized
+    )
+
+
+def filter_measure_1(instance: Instance) -> bool:
+    ok = True
+    return ok
+
+
+def process_measure_1(instance: Instance) -> Instance:
+    un_normalized = instance.un_normalized
+    normalized = instance.normalized
+    un_normalized = re.sub(r",", "", un_normalized)
+    un_normalized = re.sub(r"m2", "m²", un_normalized)
+    un_normalized = re.sub(r"(\d)([^\d.\s])", r"\1 \2", un_normalized)
+    normalized = re.sub(r"[^a-z\s]", "", normalized)
+    normalized = re.sub(r"per ([a-z\s]*)s$", r"per \1", normalized)
+    normalized = re.sub(r"[^a-z ]", "", normalized)
+    return Instance(
+        token_type=instance.token_type, un_normalized=un_normalized, normalized=normalized
+    )
+
+
+def filter_money_1(instance: Instance) -> bool:
+    ok = re.search(r"[0-9]", instance.un_normalized)
+    return ok
+
+
+def process_money_1(instance: Instance) -> Instance:
+    un_normalized = instance.un_normalized
+    normalized = instance.normalized
+    un_normalized = re.sub(r",", "", un_normalized)
+    un_normalized = re.sub(r"a\$", r"$", un_normalized)
+    un_normalized = re.sub(r"us\$", r"$", un_normalized)
+    un_normalized = re.sub(r"(\d)m\s*$", r"\1 million", un_normalized)
+    un_normalized = re.sub(r"(\d)bn?\s*$", r"\1 billion", un_normalized)
+    normalized = re.sub(r"[^a-z ]", "", normalized)
+    return Instance(
+        token_type=instance.token_type, un_normalized=un_normalized, normalized=normalized
+    )
+
+
+def filter_time_1(instance: Instance) -> bool:
+    ok = re.search(r"[0-9]", instance.un_normalized)
+    return ok
+
+
+def process_time_1(instance: Instance) -> Instance:
+    un_normalized = instance.un_normalized
+    un_normalized = re.sub(r": ", ":", un_normalized)
+    un_normalized = re.sub(r"(\d)\s?a\s?m\s?", r"\1 a.m.", un_normalized)
+    un_normalized = re.sub(r"(\d)\s?p\s?m\s?", r"\1 p.m.", un_normalized)
+    normalized = instance.normalized
+    normalized = re.sub(r"[^a-z ]", "", normalized)
+    return Instance(
+        token_type=instance.token_type, un_normalized=un_normalized, normalized=normalized
+    )
+
+
+def filter_plain_1(instance: Instance) -> bool:
+    ok = True
+    return ok
+
+
+def process_plain_1(instance: Instance) -> Instance:
+    un_normalized = instance.un_normalized
+    normalized = instance.normalized
+    return Instance(
+        token_type=instance.token_type, un_normalized=un_normalized, normalized=normalized
+    )
+
+
+def filter_punct_1(instance: Instance) -> bool:
+    ok = True
+    return ok
+
+
+def process_punct_1(instance: Instance) -> Instance:
+    un_normalized = instance.un_normalized
+    normalized = instance.normalized
+    return Instance(
+        token_type=instance.token_type, un_normalized=un_normalized, normalized=normalized
+    )
+
+
+def filter_date_1(instance: Instance) -> bool:
+    ok = True
+    return ok
+
+
+def process_date_1(instance: Instance) -> Instance:
+    un_normalized = instance.un_normalized
+    un_normalized = re.sub(r",", "", un_normalized)
+    normalized = instance.normalized
+    normalized = re.sub(r"[^a-z ]", "", normalized)
+    return Instance(
+        token_type=instance.token_type, un_normalized=un_normalized, normalized=normalized
+    )
+
+
+def filter_letters_1(instance: Instance) -> bool:
+    ok = True
+    return ok
+
+
+def process_letters_1(instance: Instance) -> Instance:
+    un_normalized = instance.un_normalized
+    normalized = instance.normalized
+    normalized = re.sub(r"[^a-z ]", "", normalized)
+    return Instance(
+        token_type=instance.token_type, un_normalized=un_normalized, normalized=normalized
+    )
+
+
+def filter_verbatim_1(instance: Instance) -> bool:
+    ok = True
+    return ok
+
+
+def process_verbatim_1(instance: Instance) -> Instance:
+    un_normalized = instance.un_normalized
+    normalized = instance.normalized
+    return Instance(
+        token_type=instance.token_type, un_normalized=un_normalized, normalized=normalized
+    )
+
+
+def filter_digit_1(instance: Instance) -> bool:
+    ok = re.search(r"[0-9]", instance.un_normalized)
+    return ok
+
+
+def process_digit_1(instance: Instance) -> Instance:
+    un_normalized = instance.un_normalized
+    normalized = instance.normalized
+    normalized = re.sub(r"[^a-z ]", "", normalized)
+    return Instance(
+        token_type=instance.token_type, un_normalized=un_normalized, normalized=normalized
+    )
+
+
+def filter_telephone_1(instance: Instance) -> bool:
+    ok = re.search(r"[0-9]", instance.un_normalized)
+    return ok
+
+
+def process_telephone_1(instance: Instance) -> Instance:
+    un_normalized = instance.un_normalized
+    normalized = instance.normalized
+    normalized = re.sub(r"[^a-z ]", "", normalized)
+    return Instance(
+        token_type=instance.token_type, un_normalized=un_normalized, normalized=normalized
+    )
+
+
+def filter_electronic_1(instance: Instance) -> bool:
+    ok = re.search(r"[0-9]", instance.un_normalized)
+    return ok
+
+
+def process_electronic_1(instance: Instance) -> Instance:
+    un_normalized = instance.un_normalized
+    normalized = instance.normalized
+    normalized = re.sub(r"[^a-z ]", "", normalized)
+    return Instance(
+        token_type=instance.token_type, un_normalized=un_normalized, normalized=normalized
+    )
+
+
+def filter_fraction_1(instance: Instance) -> bool:
+    ok = re.search(r"[0-9]", instance.un_normalized)
+    return ok
+
+
+def process_fraction_1(instance: Instance) -> Instance:
+    un_normalized = instance.un_normalized
+    normalized = instance.normalized
+    normalized = re.sub(r"[^a-z ]", "", normalized)
+    return Instance(
+        token_type=instance.token_type, un_normalized=un_normalized, normalized=normalized
+    )
+
+
+def filter_address_1(instance: Instance) -> bool:
+    ok = True
+    return ok
+
+
+def process_address_1(instance: Instance) -> Instance:
+    un_normalized = instance.un_normalized
+    normalized = instance.normalized
+    normalized = re.sub(r"[^a-z ]", "", normalized)
+    return Instance(
+        token_type=instance.token_type, un_normalized=un_normalized, normalized=normalized
+    )
+
+
+filters = []
+filters.append(
+    Filter(class_type="CARDINAL", process_func=process_cardinal_1, filter_func=filter_cardinal_1)
+)
+filters.append(
+    Filter(class_type="ORDINAL", process_func=process_ordinal_1, filter_func=filter_ordinal_1)
+)
+filters.append(
+    Filter(class_type="DECIMAL", process_func=process_decimal_1, filter_func=filter_decimal_1)
+)
+filters.append(
+    Filter(class_type="MEASURE", process_func=process_measure_1, filter_func=filter_measure_1)
+)
+filters.append(Filter(class_type="MONEY", process_func=process_money_1, filter_func=filter_money_1))
+filters.append(Filter(class_type="TIME", process_func=process_time_1, filter_func=filter_time_1))
+
+filters.append(Filter(class_type="DATE", process_func=process_date_1, filter_func=filter_date_1))
+filters.append(Filter(class_type="PLAIN", process_func=process_plain_1, filter_func=filter_plain_1))
+filters.append(Filter(class_type="PUNCT", process_func=process_punct_1, filter_func=filter_punct_1))
+filters.append(
+    Filter(class_type="LETTERS", process_func=process_letters_1, filter_func=filter_letters_1)
+)
+filters.append(
+    Filter(class_type="VERBATIM", process_func=process_verbatim_1, filter_func=filter_verbatim_1)
+)
+filters.append(Filter(class_type="DIGIT", process_func=process_digit_1, filter_func=filter_digit_1))
+filters.append(
+    Filter(class_type="TELEPHONE", process_func=process_telephone_1, filter_func=filter_telephone_1)
+)
+filters.append(
+    Filter(
+        class_type="ELECTRONIC", process_func=process_electronic_1, filter_func=filter_electronic_1
+    )
+)
+filters.append(
+    Filter(class_type="FRACTION", process_func=process_fraction_1, filter_func=filter_fraction_1)
+)
+filters.append(
+    Filter(class_type="ADDRESS", process_func=process_address_1, filter_func=filter_address_1)
+)
+filters.append(Filter(class_type=EOS_TYPE, process_func=lambda x: x, filter_func=lambda x: True))
+
+
+def filter_loaded_data(data: List[Instance], verbose: bool = False) -> List[Instance]:
+    """
+    Filters list of instances
+
+    Args:
+        data: list of instances
+
+    Returns: filtered and transformed list of instances
+    """
+    updates_instances = []
+    for instance in data:
+        updated_instance = False
+        for fil in filters:
+            if fil.class_type == instance.token_type and fil.filter(instance):
+                instance = fil.process(instance)
+                updated_instance = True
+        if updated_instance:
+            if verbose:
+                print(instance)
+            updates_instances.append(instance)
+    return updates_instances
+
+
+def parse_args():
+    parser = ArgumentParser()
+    parser.add_argument(
+        "--input", help="input file path", type=str, default="./en_with_types/output-00001-of-00100"
+    )
+    parser.add_argument("--verbose", help="print filtered instances", action="store_true")
+    return parser.parse_args()
+
+
+if __name__ == "__main__":
+    args = parse_args()
+    file_path = args.input
+
+    print("Loading training data: " + file_path)
+    instance_list = load_files([file_path])  # List of instances
+    filtered_instance_list = filter_loaded_data(instance_list, args.verbose)
+    training_data_to_sentences(filtered_instance_list)

fun_text_processing/inverse_text_normalization/en/data/__init__.py

@@ -0,0 +1 @@
+

fun_text_processing/inverse_text_normalization/en/data/currency.tsv

@@ -0,0 +1,35 @@
+$	dollar
+$	us dollar
+$	united states dollar
+£	british pound
+€	euro
+₩	won
+nzd	new zealand dollar
+rs	rupee
+chf	swiss franc
+dkk	danish kroner
+fim	finnish markka
+aed	arab emirates dirham
+¥	yen
+czk	czech koruna
+mro	mauritanian ouguiya
+pkr	pakistani rupee
+crc costa rican colon
+hk$	hong kong dollar
+npr	nepalese rupee
+awg aruban florin
+nok	norwegian kroner
+tzs tanzanian shilling
+sek swedish kronor
+cyp	cypriot pound
+r	real
+sar	saudi riyal
+cve	cape verde escudo
+rsd	serbian dinar
+dm german mark
+shp saint helena pounds
+php	philippine peso
+cad	canadian dollar
+ssp	south sudanese pound
+scr	seychelles rupee
+mvr	maldivian rufiyaa

fun_text_processing/inverse_text_normalization/en/data/electronic/__init__.py

@@ -0,0 +1 @@
+

fun_text_processing/inverse_text_normalization/en/data/electronic/domain.tsv

@@ -0,0 +1,10 @@
+com
+uk
+fr
+net
+br
+in
+ru
+de
+it
+ai

fun_text_processing/inverse_text_normalization/en/data/electronic/server_name.tsv

@@ -0,0 +1,17 @@
+g mail	gmail
+gmail
+n vidia	nvidia
+nvidia
+outlook
+hotmail
+yahoo
+aol
+gmx
+msn
+live
+yandex
+orange
+wanadoo
+web
+comcast
+google

fun_text_processing/inverse_text_normalization/en/data/electronic/symbols.tsv

@@ -0,0 +1,22 @@
+.	dot
+-	dash
+-	hyphen
+_	underscore
+!	exclamation mark
+#	number sign
+$	dollar sign
+%	percent sign
+&	ampersand
+'	quote
+*	asterisk
++	plus
+/	slash
+=	equal sign
+?	question mark
+^	circumflex
+`	right single quote
+{	left brace
+|	vertical bar
+}	right brace
+~	tilde
+,	comma

fun_text_processing/inverse_text_normalization/en/data/magnitudes.tsv

@@ -0,0 +1,4 @@
+k	thousand
+m	million
+b	billion
+t	trillion

fun_text_processing/inverse_text_normalization/en/data/measurements.tsv

@@ -0,0 +1,109 @@
+f	fahrenheit
+c	celsius
+km	kilometer
+m	meter
+cm	centimeter
+mm	millimeter
+ha	hectare
+mi	mile
+m²	square meter
+km²	square kilometer
+ft	foot
+%	percent
+hz	hertz
+kw	kilowatt
+hp	horsepower
+mg	milligram
+kg	kilogram
+ghz	gigahertz
+khz	kilohertz
+mhz	megahertz
+v	volt
+h	hour
+mc	mega coulomb
+s	second
+nm	nanometer
+rpm	revolution per minute
+min	minute
+mA	milli ampere
+%	per cent
+kwh	kilo watt hour
+m³	cubic meter
+mph	mile per hour
+tw	tera watt
+mv	milli volt
+mw	megawatt
+μm	micrometer
+"	inch
+tb	terabyte
+cc	c c
+g	gram
+da	dalton
+atm	atmosphere
+ω	ohm
+db	decibel
+ps	peta second
+oz	ounce
+hl	hecto liter
+μg	microgram
+pg	petagram
+gb	gigabyte
+kb	kilobit
+ev	electron volt
+mb	megabyte
+kb	kilobyte
+kbps	kilobit per second
+mbps	megabit per second
+st	stone
+kl	kilo liter
+tj	tera joule
+kv	kilo volt
+mv	mega volt
+kn	kilonewton
+mm	megameter
+au	astronomical unit
+yd	yard
+rad	radian
+lm	lumen
+hs	hecto second
+mol	mole
+gpa	giga pascal
+ml	milliliter
+gw	gigawatt
+ma	mega ampere
+kt	knot
+kgf	kilogram force
+ng	nano gram
+ns	nanosecond
+ms	mega siemens
+bar	bar
+gl	giga liter
+μs	microsecond
+da	deci ampere
+pa	pascal
+ds	deci second
+ms	milli second
+dm	deci meter
+dm³	cubic deci meter
+amu	atomic mass unit
+mb	megabit
+mf	mega farad
+bq	becquerel
+pb	petabit
+mm²	square millimeter
+cm²	square centimeter
+sq mi	square mile
+sq ft	square foot
+kpa	kilopascal
+cd	candela
+tl	tera liter
+ms	mega second
+mpa	megapascal
+pm	peta meter
+pb	peta byte
+gwh	giga watt hour
+kcal	kilo calory
+gy	gray
+sv	sievert
+cwt	hundredweight
+cc	c c

fun_text_processing/inverse_text_normalization/en/data/months.tsv

@@ -0,0 +1,12 @@
+january
+february
+march
+april
+may
+june
+july
+august
+september
+october
+november
+december

fun_text_processing/inverse_text_normalization/en/data/numbers/__init__.py

@@ -0,0 +1 @@
+

fun_text_processing/inverse_text_normalization/en/data/numbers/digit.tsv

@@ -0,0 +1,9 @@
+one	1
+two	2
+three	3
+four	4
+five	5
+six	6
+seven	7
+eight	8
+nine	9

fun_text_processing/inverse_text_normalization/en/data/numbers/hundred.tsv

@@ -0,0 +1 @@
+hundred

fun_text_processing/inverse_text_normalization/en/data/numbers/teen.tsv

@@ -0,0 +1,10 @@
+ten	10
+eleven	11
+twelve	12
+thirteen	13
+fourteen	14
+fifteen	15
+sixteen	16
+seventeen	17
+eighteen	18
+nineteen	19

fun_text_processing/inverse_text_normalization/en/data/numbers/thousands.tsv

@@ -0,0 +1,22 @@
+thousand
+million
+billion
+trillion
+quadrillion
+quintillion
+sextillion
+septillion
+octillion
+nonillion
+decillion
+undecillion
+duodecillion
+tredecillion
+quattuordecillion
+quindecillion
+sexdecillion
+septendecillion
+octodecillion
+novemdecillion
+vigintillion
+centillion

fun_text_processing/inverse_text_normalization/en/data/numbers/ties.tsv

@@ -0,0 +1,9 @@
+twenty	2
+thirty	3
+forty	4
+fourty	4
+fifty	5
+sixty	6
+seventy	7
+eighty	8
+ninety	9

fun_text_processing/inverse_text_normalization/en/data/numbers/zero.tsv

@@ -0,0 +1 @@
+zero	0

fun_text_processing/inverse_text_normalization/en/data/ordinals/__init__.py

@@ -0,0 +1 @@
+

fun_text_processing/inverse_text_normalization/en/data/ordinals/digit.tsv

@@ -0,0 +1,9 @@
+first	one
+second	two
+third	three
+fourth	four
+fifth	five
+sixth	sixth
+seventh	seven
+eighth	eight
+ninth	nine

fun_text_processing/inverse_text_normalization/en/data/ordinals/teen.tsv

@@ -0,0 +1 @@
+twelfth	twelve

fun_text_processing/inverse_text_normalization/en/data/time/__init__.py

@@ -0,0 +1 @@
+

fun_text_processing/inverse_text_normalization/en/data/time/minute_to.tsv

@@ -0,0 +1,59 @@
+1	59
+2	58
+3	57
+4	56
+5	55
+6	54
+7	53
+8	52
+9	51
+10	50
+11	49
+12	48
+13	47
+14	46
+15	45
+16	44
+17	43
+18	42
+19	41
+20	40
+21	39
+22	38
+23	37
+24	36
+25	35
+26	34
+27	33
+28	32
+29	31
+30	30
+31	29
+32	28
+33	27
+34	26
+35	25
+36	24
+37	23
+38	22
+39	21
+40	20
+41	19
+42	18
+43	17
+44	16
+45	15
+46	14
+47	13
+48	12
+49	11
+50	10
+51	9
+52	8
+53	7
+54	6
+55	5
+56	4
+57	3
+58	2
+59	1

fun_text_processing/inverse_text_normalization/en/data/time/time_suffix.tsv

@@ -0,0 +1,8 @@
+p m	p.m.
+pm	p.m.
+p.m.
+p.m	p.m.
+am	a.m.
+a.m.
+a.m	a.m.
+a m	a.m.

fun_text_processing/inverse_text_normalization/en/data/time/time_zone.tsv

@@ -0,0 +1,7 @@
+cst	c s t
+cet	c e t
+pst	p s t
+est	e s t
+pt	p t
+et	e t
+gmt	g m t

fun_text_processing/inverse_text_normalization/en/data/time/to_hour.tsv

@@ -0,0 +1,12 @@
+one	12
+two	1
+three	2
+four	3
+five	4
+six	5
+seven	6
+eigh	7
+nine	8
+ten	9
+eleven	10
+twelve	11

fun_text_processing/inverse_text_normalization/en/data/whitelist.tsv

@@ -0,0 +1,12 @@
+e.g.	for example
+dr.	doctor
+mr.	mister
+mrs.	misses
+st.	saint
+7-eleven	seven eleven
+es3	e s three
+s&p	s and p
+ASAP	a s a p
+AT&T	a t and t
+LLP	l l p
+ATM	a t m

fun_text_processing/inverse_text_normalization/en/taggers/__init__.py

@@ -0,0 +1 @@
+

fun_text_processing/inverse_text_normalization/en/taggers/cardinal.py

@@ -0,0 +1,138 @@
+import pynini
+from fun_text_processing.inverse_text_normalization.en.utils import get_abs_path, num_to_word
+from fun_text_processing.text_normalization.en.graph_utils import (
+    DAMO_ALPHA,
+    DAMO_DIGIT,
+    DAMO_SIGMA,
+    DAMO_SPACE,
+    GraphFst,
+    delete_space,
+)
+from pynini.lib import pynutil
+
+
+class CardinalFst(GraphFst):
+    """
+    Finite state transducer for classifying cardinals
+        e.g. minus twenty three -> cardinal { integer: "23" negative: "-" } }
+    Numbers below thirteen are not converted.
+    """
+
+    def __init__(self):
+        super().__init__(name="cardinal", kind="classify")
+        graph_zero = pynini.string_file(get_abs_path("data/numbers/zero.tsv"))
+        graph_digit = pynini.string_file(get_abs_path("data/numbers/digit.tsv"))
+        graph_ties = pynini.string_file(get_abs_path("data/numbers/ties.tsv"))
+        graph_teen = pynini.string_file(get_abs_path("data/numbers/teen.tsv"))
+
+        graph_hundred = pynini.cross("hundred", "")
+
+        graph_hundred_component = pynini.union(
+            graph_digit + delete_space + graph_hundred, pynutil.insert("0")
+        )
+        graph_hundred_component += delete_space
+        graph_hundred_component += pynini.union(
+            graph_teen | pynutil.insert("00"),
+            (graph_ties | pynutil.insert("0")) + delete_space + (graph_digit | pynutil.insert("0")),
+        )
+
+        graph_hundred_component_at_least_one_none_zero_digit = graph_hundred_component @ (
+            pynini.closure(DAMO_DIGIT) + (DAMO_DIGIT - "0") + pynini.closure(DAMO_DIGIT)
+        )
+        self.graph_hundred_component_at_least_one_none_zero_digit = (
+            graph_hundred_component_at_least_one_none_zero_digit
+        )
+
+        graph_thousands = pynini.union(
+            graph_hundred_component_at_least_one_none_zero_digit
+            + delete_space
+            + pynutil.delete("thousand"),
+            pynutil.insert("000", weight=0.1),
+        )
+
+        graph_million = pynini.union(
+            graph_hundred_component_at_least_one_none_zero_digit
+            + delete_space
+            + pynutil.delete("million"),
+            pynutil.insert("000", weight=0.1),
+        )
+        graph_billion = pynini.union(
+            graph_hundred_component_at_least_one_none_zero_digit
+            + delete_space
+            + pynutil.delete("billion"),
+            pynutil.insert("000", weight=0.1),
+        )
+        graph_trillion = pynini.union(
+            graph_hundred_component_at_least_one_none_zero_digit
+            + delete_space
+            + pynutil.delete("trillion"),
+            pynutil.insert("000", weight=0.1),
+        )
+        graph_quadrillion = pynini.union(
+            graph_hundred_component_at_least_one_none_zero_digit
+            + delete_space
+            + pynutil.delete("quadrillion"),
+            pynutil.insert("000", weight=0.1),
+        )
+        graph_quintillion = pynini.union(
+            graph_hundred_component_at_least_one_none_zero_digit
+            + delete_space
+            + pynutil.delete("quintillion"),
+            pynutil.insert("000", weight=0.1),
+        )
+        graph_sextillion = pynini.union(
+            graph_hundred_component_at_least_one_none_zero_digit
+            + delete_space
+            + pynutil.delete("sextillion"),
+            pynutil.insert("000", weight=0.1),
+        )
+
+        graph = pynini.union(
+            graph_sextillion
+            + delete_space
+            + graph_quintillion
+            + delete_space
+            + graph_quadrillion
+            + delete_space
+            + graph_trillion
+            + delete_space
+            + graph_billion
+            + delete_space
+            + graph_million
+            + delete_space
+            + graph_thousands
+            + delete_space
+            + graph_hundred_component,
+            graph_zero,
+        )
+
+        graph = graph @ pynini.union(
+            pynutil.delete(pynini.closure("0"))
+            + pynini.difference(DAMO_DIGIT, "0")
+            + pynini.closure(DAMO_DIGIT),
+            "0",
+        )
+
+        labels_exception = [num_to_word(x) for x in range(0, 13)]
+        graph_exception = pynini.union(*labels_exception)
+
+        graph = (
+            pynini.cdrewrite(pynutil.delete("and"), DAMO_SPACE, DAMO_SPACE, DAMO_SIGMA)
+            @ (DAMO_ALPHA + DAMO_SIGMA)
+            @ graph
+        )
+
+        self.graph_no_exception = graph
+
+        self.graph = (pynini.project(graph, "input") - graph_exception.arcsort()) @ graph
+
+        optional_minus_graph = pynini.closure(
+            pynutil.insert("negative: ") + pynini.cross("minus", '"-"') + DAMO_SPACE, 0, 1
+        )
+
+        final_graph = (
+            optional_minus_graph + pynutil.insert('integer: "') + self.graph + pynutil.insert('"')
+        )
+
+        final_graph = self.add_tokens(final_graph)
+        self.fst = final_graph.optimize()

fun_text_processing/inverse_text_normalization/en/taggers/date.py

@@ -0,0 +1,150 @@
+import pynini
+from fun_text_processing.inverse_text_normalization.en.utils import get_abs_path
+from fun_text_processing.text_normalization.en.graph_utils import (
+    DAMO_ALPHA,
+    DAMO_DIGIT,
+    GraphFst,
+    delete_extra_space,
+    delete_space,
+)
+from pynini.lib import pynutil
+
+graph_teen = pynini.string_file(get_abs_path("data/numbers/teen.tsv")).optimize()
+graph_digit = pynini.string_file(get_abs_path("data/numbers/digit.tsv")).optimize()
+ties_graph = pynini.string_file(get_abs_path("data/numbers/ties.tsv")).optimize()
+
+
+def _get_month_graph():
+    """
+    Transducer for month, e.g. march -> march
+    """
+    month_graph = pynini.string_file(get_abs_path("data/months.tsv"))
+    return month_graph
+
+
+def _get_ties_graph():
+    """
+    Transducer for 20-99 e.g
+    twenty three -> 23
+    """
+    graph = ties_graph + (delete_space + graph_digit | pynutil.insert("0"))
+    return graph
+
+
+def _get_range_graph():
+    """
+    Transducer for decades (1**0s, 2**0s), centuries (2*00s, 1*00s), millennia (2000s)
+    """
+    graph_ties = _get_ties_graph()
+    graph = (graph_ties | graph_teen) + delete_space + pynini.cross("hundreds", "00s")
+    graph |= pynini.cross("two", "2") + delete_space + pynini.cross("thousands", "000s")
+    graph |= (
+        (graph_ties | graph_teen)
+        + delete_space
+        + (pynini.closure(DAMO_ALPHA, 1) + (pynini.cross("ies", "y") | pynutil.delete("s")))
+        @ (graph_ties | pynini.cross("ten", "10"))
+        + pynutil.insert("s")
+    )
+    graph @= pynini.union("1", "2") + DAMO_DIGIT + DAMO_DIGIT + DAMO_DIGIT + "s"
+    return graph
+
+
+def _get_year_graph():
+    """
+    Transducer for year, e.g. twenty twenty -> 2020
+    """
+
+    def _get_digits_graph():
+        zero = pynini.cross((pynini.accep("oh") | pynini.accep("o")), "0")
+        graph = zero + delete_space + graph_digit
+        graph.optimize()
+        return graph
+
+    def _get_thousands_graph():
+        graph_ties = _get_ties_graph()
+        graph_hundred_component = (
+            graph_digit + delete_space + pynutil.delete("hundred")
+        ) | pynutil.insert("0")
+        graph = (
+            graph_digit
+            + delete_space
+            + pynutil.delete("thousand")
+            + delete_space
+            + graph_hundred_component
+            + delete_space
+            + (graph_teen | graph_ties)
+        )
+        return graph
+
+    graph_ties = _get_ties_graph()
+    graph_digits = _get_digits_graph()
+    graph_thousands = _get_thousands_graph()
+    year_graph = (
+        # 20 19, 40 12, 2012 - assuming no limit on the year
+        (graph_teen + delete_space + (graph_ties | graph_digits | graph_teen))
+        | (graph_ties + delete_space + (graph_ties | graph_digits | graph_teen))
+        | graph_thousands
+    )
+    year_graph.optimize()
+    return year_graph
+
+
+class DateFst(GraphFst):
+    """
+    Finite state transducer for classifying date,
+        e.g. january fifth twenty twelve -> date { month: "january" day: "5" year: "2012" preserve_order: true }
+        e.g. the fifth of january twenty twelve -> date { day: "5" month: "january" year: "2012" preserve_order: true }
+        e.g. twenty twenty -> date { year: "2012" preserve_order: true }
+
+    Args:
+        ordinal: OrdinalFst
+    """
+
+    def __init__(self, ordinal: GraphFst):
+        super().__init__(name="date", kind="classify")
+
+        ordinal_graph = ordinal.graph
+        year_graph = _get_year_graph()
+        YEAR_WEIGHT = 0.001
+        year_graph = pynutil.add_weight(year_graph, YEAR_WEIGHT)
+        month_graph = _get_month_graph()
+
+        month_graph = pynutil.insert('month: "') + month_graph + pynutil.insert('"')
+
+        day_graph = (
+            pynutil.insert('day: "') + pynutil.add_weight(ordinal_graph, -0.7) + pynutil.insert('"')
+        )
+        graph_year = (
+            delete_extra_space
+            + pynutil.insert('year: "')
+            + pynutil.add_weight(year_graph, -YEAR_WEIGHT)
+            + pynutil.insert('"')
+        )
+        optional_graph_year = pynini.closure(
+            graph_year,
+            0,
+            1,
+        )
+        graph_mdy = month_graph + (
+            (delete_extra_space + day_graph)
+            | graph_year
+            | (delete_extra_space + day_graph + graph_year)
+        )
+        graph_dmy = (
+            pynutil.delete("the")
+            + delete_space
+            + day_graph
+            + delete_space
+            + pynutil.delete("of")
+            + delete_extra_space
+            + month_graph
+            + optional_graph_year
+        )
+        graph_year = (
+            pynutil.insert('year: "') + (year_graph | _get_range_graph()) + pynutil.insert('"')
+        )
+
+        final_graph = graph_mdy | graph_dmy | graph_year
+        final_graph += pynutil.insert(" preserve_order: true")
+        final_graph = self.add_tokens(final_graph)
+        self.fst = final_graph.optimize()

fun_text_processing/inverse_text_normalization/en/taggers/decimal.py

@@ -0,0 +1,100 @@
+import pynini
+from fun_text_processing.inverse_text_normalization.en.utils import get_abs_path
+from fun_text_processing.text_normalization.en.graph_utils import (
+    DAMO_DIGIT,
+    GraphFst,
+    delete_extra_space,
+    delete_space,
+)
+from pynini.lib import pynutil
+
+
+def get_quantity(
+    decimal: "pynini.FstLike", cardinal_up_to_hundred: "pynini.FstLike"
+) -> "pynini.FstLike":
+    """
+    Returns FST that transforms either a cardinal or decimal followed by a quantity into a numeral,
+    e.g. one million -> integer_part: "1" quantity: "million"
+    e.g. one point five million -> integer_part: "1" fractional_part: "5" quantity: "million"
+
+    Args:
+        decimal: decimal FST
+        cardinal_up_to_hundred: cardinal FST
+    """
+    numbers = cardinal_up_to_hundred @ (
+        pynutil.delete(pynini.closure("0"))
+        + pynini.difference(DAMO_DIGIT, "0")
+        + pynini.closure(DAMO_DIGIT)
+    )
+    suffix = pynini.union(
+        "million", "billion", "trillion", "quadrillion", "quintillion", "sextillion"
+    )
+    res = (
+        pynutil.insert('integer_part: "')
+        + numbers
+        + pynutil.insert('"')
+        + delete_extra_space
+        + pynutil.insert('quantity: "')
+        + suffix
+        + pynutil.insert('"')
+    )
+    res |= (
+        decimal
+        + delete_extra_space
+        + pynutil.insert('quantity: "')
+        + (suffix | "thousand")
+        + pynutil.insert('"')
+    )
+    return res
+
+
+class DecimalFst(GraphFst):
+    """
+    Finite state transducer for classifying decimal
+        e.g. minus twelve point five o o six billion -> decimal { negative: "true" integer_part: "12"  fractional_part: "5006" quantity: "billion" }
+        e.g. one billion -> decimal { integer_part: "1" quantity: "billion" }
+    Args:
+        cardinal: CardinalFst
+    """
+
+    def __init__(self, cardinal: GraphFst):
+        super().__init__(name="decimal", kind="classify")
+
+        cardinal_graph = cardinal.graph_no_exception
+
+        graph_decimal = pynini.string_file(get_abs_path("data/numbers/digit.tsv"))
+        graph_decimal |= pynini.string_file(get_abs_path("data/numbers/zero.tsv")) | pynini.cross(
+            "o", "0"
+        )
+
+        graph_decimal = pynini.closure(graph_decimal + delete_space) + graph_decimal
+        self.graph = graph_decimal
+
+        point = pynutil.delete("point")
+
+        optional_graph_negative = pynini.closure(
+            pynutil.insert("negative: ") + pynini.cross("minus", '"true"') + delete_extra_space,
+            0,
+            1,
+        )
+
+        graph_fractional = (
+            pynutil.insert('fractional_part: "') + graph_decimal + pynutil.insert('"')
+        )
+        graph_integer = pynutil.insert('integer_part: "') + cardinal_graph + pynutil.insert('"')
+        final_graph_wo_sign = (
+            pynini.closure(graph_integer + delete_extra_space, 0, 1)
+            + point
+            + delete_extra_space
+            + graph_fractional
+        )
+        final_graph = optional_graph_negative + final_graph_wo_sign
+
+        self.final_graph_wo_negative = final_graph_wo_sign | get_quantity(
+            final_graph_wo_sign, cardinal.graph_hundred_component_at_least_one_none_zero_digit
+        )
+        final_graph |= optional_graph_negative + get_quantity(
+            final_graph_wo_sign, cardinal.graph_hundred_component_at_least_one_none_zero_digit
+        )
+        final_graph = self.add_tokens(final_graph)
+        self.fst = final_graph.optimize()

fun_text_processing/inverse_text_normalization/en/taggers/electronic.py

@@ -0,0 +1,97 @@
+import pynini
+from fun_text_processing.inverse_text_normalization.en.utils import get_abs_path
+from fun_text_processing.text_normalization.en.graph_utils import DAMO_ALPHA, GraphFst, insert_space
+from pynini.lib import pynutil
+
+
+class ElectronicFst(GraphFst):
+    """
+    Finite state transducer for classifying electronic: as URLs, email addresses, etc.
+        e.g. c d f one at a b c dot e d u -> tokens { electronic { username: "cdf1" domain: "abc.edu" } }
+    """
+
+    def __init__(self):
+        super().__init__(name="electronic", kind="classify")
+
+        delete_extra_space = pynutil.delete(" ")
+        alpha_num = (
+            DAMO_ALPHA
+            | pynini.string_file(get_abs_path("data/numbers/digit.tsv"))
+            | pynini.string_file(get_abs_path("data/numbers/zero.tsv"))
+        )
+
+        symbols = pynini.string_file(get_abs_path("data/electronic/symbols.tsv")).invert()
+
+        accepted_username = alpha_num | symbols
+        process_dot = pynini.cross("dot", ".")
+        username = (
+            alpha_num + pynini.closure(delete_extra_space + accepted_username)
+        ) | pynutil.add_weight(pynini.closure(DAMO_ALPHA, 1), weight=0.0001)
+        username = pynutil.insert('username: "') + username + pynutil.insert('"')
+        single_alphanum = pynini.closure(alpha_num + delete_extra_space) + alpha_num
+        server = single_alphanum | pynini.string_file(
+            get_abs_path("data/electronic/server_name.tsv")
+        )
+        domain = single_alphanum | pynini.string_file(get_abs_path("data/electronic/domain.tsv"))
+        domain_graph = (
+            pynutil.insert('domain: "')
+            + server
+            + delete_extra_space
+            + process_dot
+            + delete_extra_space
+            + domain
+            + pynutil.insert('"')
+        )
+        graph = (
+            username
+            + delete_extra_space
+            + pynutil.delete("at")
+            + insert_space
+            + delete_extra_space
+            + domain_graph
+        )
+
+        ############# url ###
+        protocol_end = pynini.cross(pynini.union("w w w", "www"), "www")
+        protocol_start = (
+            pynini.cross("h t t p", "http") | pynini.cross("h t t p s", "https")
+        ) + pynini.cross(" colon slash slash ", "://")
+        # .com,
+        ending = (
+            delete_extra_space
+            + symbols
+            + delete_extra_space
+            + (
+                domain
+                | pynini.closure(
+                    accepted_username + delete_extra_space,
+                )
+                + accepted_username
+            )
+        )
+
+        protocol_default = (
+            (
+                (pynini.closure(delete_extra_space + accepted_username, 1) | server)
+                | pynutil.add_weight(pynini.closure(DAMO_ALPHA, 1), weight=0.0001)
+            )
+            + pynini.closure(ending, 1)
+        ).optimize()
+        protocol = (
+            pynini.closure(protocol_start, 0, 1)
+            + protocol_end
+            + delete_extra_space
+            + process_dot
+            + protocol_default
+        ).optimize()
+
+        protocol |= (
+            pynini.closure(protocol_end + delete_extra_space + process_dot, 0, 1) + protocol_default
+        )
+
+        protocol = pynutil.insert('protocol: "') + protocol.optimize() + pynutil.insert('"')
+        graph |= protocol
+        ########
+
+        final_graph = self.add_tokens(graph)
+        self.fst = final_graph.optimize()

fun_text_processing/inverse_text_normalization/en/taggers/fraction.py

@@ -0,0 +1,11 @@
+from fun_text_processing.text_normalization.en.graph_utils import GraphFst
+
+
+class FractionFst(GraphFst):
+    """
+    Finite state transducer for classifying fraction
+    """
+
+    def __init__(self):
+        super().__init__(name="fraction", kind="classify")
+        # integer_part # numerator # denominator

fun_text_processing/inverse_text_normalization/en/taggers/measure.py

@@ -0,0 +1,97 @@
+import pynini
+from fun_text_processing.inverse_text_normalization.en.utils import get_abs_path
+from fun_text_processing.text_normalization.en.graph_utils import (
+    DAMO_SIGMA,
+    GraphFst,
+    convert_space,
+    delete_extra_space,
+    delete_space,
+    get_singulars,
+)
+from pynini.lib import pynutil
+
+
+class MeasureFst(GraphFst):
+    """
+    Finite state transducer for classifying measure
+        e.g. minus twelve kilograms -> measure { negative: "true" cardinal { integer: "12" } units: "kg" }
+
+    Args:
+        cardinal: CardinalFst
+        decimal: DecimalFst
+    """
+
+    def __init__(self, cardinal: GraphFst, decimal: GraphFst):
+        super().__init__(name="measure", kind="classify")
+
+        cardinal_graph = cardinal.graph_no_exception
+
+        graph_unit = pynini.string_file(get_abs_path("data/measurements.tsv"))
+        graph_unit_singular = pynini.invert(graph_unit)  # singular -> abbr
+        graph_unit_plural = get_singulars(graph_unit_singular)  # plural -> abbr
+
+        optional_graph_negative = pynini.closure(
+            pynutil.insert("negative: ") + pynini.cross("minus", '"true"') + delete_extra_space,
+            0,
+            1,
+        )
+
+        unit_singular = convert_space(graph_unit_singular)
+        unit_plural = convert_space(graph_unit_plural)
+        unit_misc = (
+            pynutil.insert("/")
+            + pynutil.delete("per")
+            + delete_space
+            + convert_space(graph_unit_singular)
+        )
+
+        unit_singular = (
+            pynutil.insert('units: "')
+            + (
+                unit_singular
+                | unit_misc
+                | pynutil.add_weight(unit_singular + delete_space + unit_misc, 0.01)
+            )
+            + pynutil.insert('"')
+        )
+        unit_plural = (
+            pynutil.insert('units: "')
+            + (
+                unit_plural
+                | unit_misc
+                | pynutil.add_weight(unit_plural + delete_space + unit_misc, 0.01)
+            )
+            + pynutil.insert('"')
+        )
+
+        subgraph_decimal = (
+            pynutil.insert("decimal { ")
+            + optional_graph_negative
+            + decimal.final_graph_wo_negative
+            + pynutil.insert(" }")
+            + delete_extra_space
+            + unit_plural
+        )
+        subgraph_cardinal = (
+            pynutil.insert("cardinal { ")
+            + optional_graph_negative
+            + pynutil.insert('integer: "')
+            + ((DAMO_SIGMA - "one") @ cardinal_graph)
+            + pynutil.insert('"')
+            + pynutil.insert(" }")
+            + delete_extra_space
+            + unit_plural
+        )
+        subgraph_cardinal |= (
+            pynutil.insert("cardinal { ")
+            + optional_graph_negative
+            + pynutil.insert('integer: "')
+            + pynini.cross("one", "1")
+            + pynutil.insert('"')
+            + pynutil.insert(" }")
+            + delete_extra_space
+            + unit_singular
+        )
+        final_graph = subgraph_decimal | subgraph_cardinal
+        final_graph = self.add_tokens(final_graph)
+        self.fst = final_graph.optimize()

fun_text_processing/inverse_text_normalization/en/taggers/money.py

@@ -0,0 +1,110 @@
+import pynini
+from fun_text_processing.inverse_text_normalization.en.utils import get_abs_path
+from fun_text_processing.text_normalization.en.graph_utils import (
+    DAMO_DIGIT,
+    DAMO_NOT_SPACE,
+    DAMO_SIGMA,
+    GraphFst,
+    convert_space,
+    delete_extra_space,
+    delete_space,
+    get_singulars,
+    insert_space,
+)
+from pynini.lib import pynutil
+
+
+class MoneyFst(GraphFst):
+    """
+    Finite state transducer for classifying money
+        e.g. twelve dollars and five cents -> money { integer_part: "12" fractional_part: 05 currency: "$" }
+
+    Args:
+        cardinal: CardinalFst
+        decimal: DecimalFst
+    """
+
+    def __init__(self, cardinal: GraphFst, decimal: GraphFst):
+        super().__init__(name="money", kind="classify")
+        # quantity, integer_part, fractional_part, currency
+
+        cardinal_graph = cardinal.graph_no_exception
+        # add support for missing hundred (only for 3 digit numbers)
+        # "one fifty" -> "one hundred fifty"
+        with_hundred = pynini.compose(
+            pynini.closure(DAMO_NOT_SPACE)
+            + pynini.accep(" ")
+            + pynutil.insert("hundred ")
+            + DAMO_SIGMA,
+            pynini.compose(cardinal_graph, DAMO_DIGIT**3),
+        )
+        cardinal_graph |= with_hundred
+        graph_decimal_final = decimal.final_graph_wo_negative
+
+        unit = pynini.string_file(get_abs_path("data/currency.tsv"))
+        unit_singular = pynini.invert(unit)
+        unit_plural = get_singulars(unit_singular)
+
+        graph_unit_singular = (
+            pynutil.insert('currency: "') + convert_space(unit_singular) + pynutil.insert('"')
+        )
+        graph_unit_plural = (
+            pynutil.insert('currency: "') + convert_space(unit_plural) + pynutil.insert('"')
+        )
+
+        add_leading_zero_to_double_digit = (DAMO_DIGIT + DAMO_DIGIT) | (
+            pynutil.insert("0") + DAMO_DIGIT
+        )
+        # twelve dollars (and) fifty cents, zero cents
+        cents_standalone = (
+            pynutil.insert('fractional_part: "')
+            + pynini.union(
+                pynutil.add_weight(((DAMO_SIGMA - "one") @ cardinal_graph), -0.7)
+                @ add_leading_zero_to_double_digit
+                + delete_space
+                + (pynutil.delete("cents") | pynutil.delete("cent")),
+                pynini.cross("one", "01") + delete_space + pynutil.delete("cent"),
+            )
+            + pynutil.insert('"')
+        )
+
+        optional_cents_standalone = pynini.closure(
+            delete_space
+            + pynini.closure(pynutil.delete("and") + delete_space, 0, 1)
+            + insert_space
+            + cents_standalone,
+            0,
+            1,
+        )
+        # twelve dollars fifty, only after integer
+        optional_cents_suffix = pynini.closure(
+            delete_extra_space
+            + pynutil.insert('fractional_part: "')
+            + pynutil.add_weight(cardinal_graph @ add_leading_zero_to_double_digit, -0.7)
+            + pynutil.insert('"'),
+            0,
+            1,
+        )
+
+        graph_integer = (
+            pynutil.insert('integer_part: "')
+            + ((DAMO_SIGMA - "one") @ cardinal_graph)
+            + pynutil.insert('"')
+            + delete_extra_space
+            + graph_unit_plural
+            + (optional_cents_standalone | optional_cents_suffix)
+        )
+        graph_integer |= (
+            pynutil.insert('integer_part: "')
+            + pynini.cross("one", "1")
+            + pynutil.insert('"')
+            + delete_extra_space
+            + graph_unit_singular
+            + (optional_cents_standalone | optional_cents_suffix)
+        )
+        graph_decimal = graph_decimal_final + delete_extra_space + graph_unit_plural
+        graph_decimal |= pynutil.insert('currency: "$" integer_part: "0" ') + cents_standalone
+        final_graph = graph_integer | graph_decimal
+
+        final_graph = self.add_tokens(final_graph)
+        self.fst = final_graph.optimize()

fun_text_processing/inverse_text_normalization/en/taggers/ordinal.py

@@ -0,0 +1,29 @@
+import pynini
+from fun_text_processing.inverse_text_normalization.en.utils import get_abs_path
+from fun_text_processing.text_normalization.en.graph_utils import DAMO_CHAR, GraphFst
+from pynini.lib import pynutil
+
+
+class OrdinalFst(GraphFst):
+    """
+    Finite state transducer for classifying ordinal
+        e.g. thirteenth -> ordinal { integer: "13" }
+
+    Args:
+        cardinal: CardinalFst
+    """
+
+    def __init__(self, cardinal: GraphFst):
+        super().__init__(name="ordinal", kind="classify")
+
+        cardinal_graph = cardinal.graph_no_exception
+        graph_digit = pynini.string_file(get_abs_path("data/ordinals/digit.tsv"))
+        graph_teens = pynini.string_file(get_abs_path("data/ordinals/teen.tsv"))
+        graph = pynini.closure(DAMO_CHAR) + pynini.union(
+            graph_digit, graph_teens, pynini.cross("tieth", "ty"), pynini.cross("th", "")
+        )
+
+        self.graph = graph @ cardinal_graph
+        final_graph = pynutil.insert('integer: "') + self.graph + pynutil.insert('"')
+        final_graph = self.add_tokens(final_graph)
+        self.fst = final_graph.optimize()

fun_text_processing/inverse_text_normalization/en/taggers/punctuation.py

@@ -0,0 +1,20 @@
+import pynini
+from fun_text_processing.text_normalization.en.graph_utils import GraphFst
+from pynini.lib import pynutil
+
+
+class PunctuationFst(GraphFst):
+    """
+    Finite state transducer for classifying punctuation
+        e.g. a, -> tokens { name: "a" } tokens { name: "," }
+    """
+
+    def __init__(self):
+        super().__init__(name="punctuation", kind="classify")
+
+        s = "!#$%&'()*+,-./:;<=>?@^_`{|}~"
+        punct = pynini.union(*s)
+
+        graph = pynutil.insert('name: "') + punct + pynutil.insert('"')
+
+        self.fst = graph.optimize()

fun_text_processing/inverse_text_normalization/en/taggers/telephone.py

@@ -0,0 +1,149 @@
+import pynini
+from fun_text_processing.inverse_text_normalization.en.utils import get_abs_path
+from fun_text_processing.text_normalization.en.graph_utils import (
+    DAMO_ALNUM,
+    DAMO_ALPHA,
+    DAMO_DIGIT,
+    GraphFst,
+    insert_space,
+)
+from pynini.lib import pynutil
+
+
+def get_serial_number(cardinal):
+    """
+    any alphanumerical character sequence with at least one number with length greater equal to 3
+    """
+    digit = pynini.compose(cardinal.graph_no_exception, DAMO_DIGIT)
+    character = digit | DAMO_ALPHA
+    sequence = character + pynini.closure(pynutil.delete(" ") + character, 2)
+    sequence = sequence @ (pynini.closure(DAMO_ALNUM) + DAMO_DIGIT + pynini.closure(DAMO_ALNUM))
+    return sequence.optimize()
+
+
+class TelephoneFst(GraphFst):
+    """
+    Finite state transducer for classifying telephone numbers, e.g.
+        one two three one two three five six seven eight -> { number_part: "123-123-5678" }
+
+    This class also support card number and IP format.
+        "one two three dot one double three dot o dot four o" -> { number_part: "123.133.0.40"}
+
+        "three two double seven three two one four three two one four three double zero five" ->
+            { number_part: 3277 3214 3214 3005}
+
+    Args:
+        cardinal: CardinalFst
+    """
+
+    def __init__(self, cardinal: GraphFst):
+        super().__init__(name="telephone", kind="classify")
+        # country code, number_part, extension
+        digit_to_str = (
+            pynini.invert(pynini.string_file(get_abs_path("data/numbers/digit.tsv")).optimize())
+            | pynini.cross("0", pynini.union("o", "oh", "zero")).optimize()
+        )
+
+        str_to_digit = pynini.invert(digit_to_str)
+
+        double_digit = pynini.union(
+            *[
+                pynini.cross(
+                    pynini.project(str(i) @ digit_to_str, "output")
+                    + pynini.accep(" ")
+                    + pynini.project(str(i) @ digit_to_str, "output"),
+                    pynutil.insert("double ") + pynini.project(str(i) @ digit_to_str, "output"),
+                )
+                for i in range(10)
+            ]
+        )
+        double_digit.invert()
+
+        # to handle cases like "one twenty three"
+        two_digit_cardinal = pynini.compose(cardinal.graph_no_exception, DAMO_DIGIT**2)
+        double_digit_to_digit = (
+            pynini.compose(double_digit, str_to_digit + pynutil.delete(" ") + str_to_digit)
+            | two_digit_cardinal
+        )
+
+        single_or_double_digit = (
+            pynutil.add_weight(double_digit_to_digit, -0.0001) | str_to_digit
+        ).optimize()
+        single_or_double_digit |= (
+            single_or_double_digit
+            + pynini.closure(
+                pynutil.add_weight(pynutil.delete(" ") + single_or_double_digit, 0.0001)
+            )
+        ).optimize()
+
+        number_part = pynini.compose(
+            single_or_double_digit,
+            DAMO_DIGIT**3
+            + pynutil.insert("-")
+            + DAMO_DIGIT**3
+            + pynutil.insert("-")
+            + DAMO_DIGIT**4,
+        ).optimize()
+        number_part = (
+            pynutil.insert('number_part: "') + number_part.optimize() + pynutil.insert('"')
+        )
+
+        cardinal_option = pynini.compose(single_or_double_digit, DAMO_DIGIT ** (2, 3))
+
+        country_code = (
+            pynutil.insert('country_code: "')
+            + pynini.closure(pynini.cross("plus ", "+"), 0, 1)
+            + (
+                (pynini.closure(str_to_digit + pynutil.delete(" "), 0, 2) + str_to_digit)
+                | cardinal_option
+            )
+            + pynutil.insert('"')
+        )
+
+        optional_country_code = pynini.closure(
+            country_code + pynutil.delete(" ") + insert_space, 0, 1
+        ).optimize()
+        graph = optional_country_code + number_part
+
+        # credit card number
+        space_four_digits = insert_space + DAMO_DIGIT**4
+        credit_card_graph = pynini.compose(
+            single_or_double_digit, DAMO_DIGIT**4 + space_four_digits**3
+        ).optimize()
+        graph |= (
+            pynutil.insert('number_part: "') + credit_card_graph.optimize() + pynutil.insert('"')
+        )
+
+        # SSN
+        ssn_graph = pynini.compose(
+            single_or_double_digit,
+            DAMO_DIGIT**3
+            + pynutil.insert("-")
+            + DAMO_DIGIT**2
+            + pynutil.insert("-")
+            + DAMO_DIGIT**4,
+        ).optimize()
+        graph |= pynutil.insert('number_part: "') + ssn_graph.optimize() + pynutil.insert('"')
+
+        # ip
+        digit_or_double = (
+            pynini.closure(str_to_digit + pynutil.delete(" "), 0, 1) + double_digit_to_digit
+        )
+        digit_or_double |= double_digit_to_digit + pynini.closure(
+            pynutil.delete(" ") + str_to_digit, 0, 1
+        )
+        digit_or_double |= str_to_digit + (pynutil.delete(" ") + str_to_digit) ** (0, 2)
+        digit_or_double |= cardinal_option
+        digit_or_double = digit_or_double.optimize()
+
+        ip_graph = digit_or_double + (pynini.cross(" dot ", ".") + digit_or_double) ** 3
+
+        graph |= pynutil.insert('number_part: "') + ip_graph.optimize() + pynutil.insert('"')
+        graph |= (
+            pynutil.insert('number_part: "')
+            + pynutil.add_weight(get_serial_number(cardinal=cardinal), weight=0.0001)
+            + pynutil.insert('"')
+        )
+
+        final_graph = self.add_tokens(graph)
+        self.fst = final_graph.optimize()

fun_text_processing/inverse_text_normalization/en/taggers/time.py

@@ -0,0 +1,139 @@
+import pynini
+from fun_text_processing.inverse_text_normalization.en.taggers.cardinal import CardinalFst
+from fun_text_processing.inverse_text_normalization.en.utils import get_abs_path, num_to_word
+from fun_text_processing.text_normalization.en.graph_utils import (
+    GraphFst,
+    convert_space,
+    delete_extra_space,
+    delete_space,
+    insert_space,
+)
+from pynini.lib import pynutil
+
+
+class TimeFst(GraphFst):
+    """
+    Finite state transducer for classifying time
+        e.g. twelve thirty -> time { hours: "12" minutes: "30" }
+        e.g. twelve past one -> time { minutes: "12" hours: "1" }
+        e.g. two o clock a m -> time { hours: "2" suffix: "a.m." }
+        e.g. quarter to two -> time { hours: "1" minutes: "45" }
+        e.g. quarter past two -> time { hours: "2" minutes: "15" }
+        e.g. half past two -> time { hours: "2" minutes: "30" }
+    """
+
+    def __init__(self):
+        super().__init__(name="time", kind="classify")
+        # hours, minutes, seconds, suffix, zone, style, speak_period
+
+        suffix_graph = pynini.string_file(get_abs_path("data/time/time_suffix.tsv"))
+        time_zone_graph = pynini.invert(pynini.string_file(get_abs_path("data/time/time_zone.tsv")))
+        to_hour_graph = pynini.string_file(get_abs_path("data/time/to_hour.tsv"))
+        minute_to_graph = pynini.string_file(get_abs_path("data/time/minute_to.tsv"))
+
+        # only used for < 1000 thousand -> 0 weight
+        cardinal = pynutil.add_weight(CardinalFst().graph_no_exception, weight=-0.7)
+
+        labels_hour = [num_to_word(x) for x in range(0, 24)]
+        labels_minute_single = [num_to_word(x) for x in range(1, 10)]
+        labels_minute_double = [num_to_word(x) for x in range(10, 60)]
+
+        graph_hour = pynini.union(*labels_hour) @ cardinal
+
+        graph_minute_single = pynini.union(*labels_minute_single) @ cardinal
+        graph_minute_double = pynini.union(*labels_minute_double) @ cardinal
+        graph_minute_verbose = pynini.cross("half", "30") | pynini.cross("quarter", "15")
+        oclock = pynini.cross(pynini.union("o' clock", "o clock", "o'clock", "oclock"), "")
+
+        final_graph_hour = pynutil.insert('hours: "') + graph_hour + pynutil.insert('"')
+        graph_minute = (
+            oclock + pynutil.insert("00")
+            | pynutil.delete("o") + delete_space + graph_minute_single
+            | graph_minute_double
+        )
+        final_suffix = (
+            pynutil.insert('suffix: "') + convert_space(suffix_graph) + pynutil.insert('"')
+        )
+        final_suffix = delete_space + insert_space + final_suffix
+        final_suffix_optional = pynini.closure(final_suffix, 0, 1)
+        final_time_zone_optional = pynini.closure(
+            delete_space
+            + insert_space
+            + pynutil.insert('zone: "')
+            + convert_space(time_zone_graph)
+            + pynutil.insert('"'),
+            0,
+            1,
+        )
+
+        # five o' clock
+        # two o eight, two thirty five (am/pm)
+        # two pm/am
+        graph_hm = (
+            final_graph_hour
+            + delete_extra_space
+            + pynutil.insert('minutes: "')
+            + graph_minute
+            + pynutil.insert('"')
+        )
+        # 10 past four, quarter past four, half past four
+        graph_m_past_h = (
+            pynutil.insert('minutes: "')
+            + pynini.union(graph_minute_single, graph_minute_double, graph_minute_verbose)
+            + pynutil.insert('"')
+            + delete_space
+            + pynutil.delete("past")
+            + delete_extra_space
+            + final_graph_hour
+        )
+
+        graph_quarter_time = (
+            pynutil.insert('minutes: "')
+            + pynini.cross("quarter", "45")
+            + pynutil.insert('"')
+            + delete_space
+            + pynutil.delete(pynini.union("to", "till"))
+            + delete_extra_space
+            + pynutil.insert('hours: "')
+            + to_hour_graph
+            + pynutil.insert('"')
+        )
+
+        graph_m_to_h_suffix_time = (
+            pynutil.insert('minutes: "')
+            + ((graph_minute_single | graph_minute_double).optimize() @ minute_to_graph)
+            + pynutil.insert('"')
+            + pynini.closure(
+                delete_space + pynutil.delete(pynini.union("min", "mins", "minute", "minutes")),
+                0,
+                1,
+            )
+            + delete_space
+            + pynutil.delete(pynini.union("to", "till"))
+            + delete_extra_space
+            + pynutil.insert('hours: "')
+            + to_hour_graph
+            + pynutil.insert('"')
+            + final_suffix
+        )
+
+        graph_h = (
+            final_graph_hour
+            + delete_extra_space
+            + pynutil.insert('minutes: "')
+            + (pynutil.insert("00") | graph_minute)
+            + pynutil.insert('"')
+            + final_suffix
+            + final_time_zone_optional
+        )
+        final_graph = (
+            (graph_hm | graph_m_past_h | graph_quarter_time)
+            + final_suffix_optional
+            + final_time_zone_optional
+        )
+        final_graph |= graph_h
+        final_graph |= graph_m_to_h_suffix_time
+
+        final_graph = self.add_tokens(final_graph.optimize())
+
+        self.fst = final_graph.optimize()

fun_text_processing/inverse_text_normalization/en/taggers/tokenize_and_classify.py

@@ -0,0 +1,102 @@
+import os
+
+import pynini
+from fun_text_processing.inverse_text_normalization.en.taggers.cardinal import CardinalFst
+from fun_text_processing.inverse_text_normalization.en.taggers.date import DateFst
+from fun_text_processing.inverse_text_normalization.en.taggers.decimal import DecimalFst
+from fun_text_processing.inverse_text_normalization.en.taggers.electronic import ElectronicFst
+from fun_text_processing.inverse_text_normalization.en.taggers.measure import MeasureFst
+from fun_text_processing.inverse_text_normalization.en.taggers.money import MoneyFst
+from fun_text_processing.inverse_text_normalization.en.taggers.ordinal import OrdinalFst
+from fun_text_processing.inverse_text_normalization.en.taggers.punctuation import PunctuationFst
+from fun_text_processing.inverse_text_normalization.en.taggers.telephone import TelephoneFst
+from fun_text_processing.inverse_text_normalization.en.taggers.time import TimeFst
+from fun_text_processing.inverse_text_normalization.en.taggers.whitelist import WhiteListFst
+from fun_text_processing.inverse_text_normalization.en.taggers.word import WordFst
+from fun_text_processing.text_normalization.en.graph_utils import (
+    GraphFst,
+    delete_extra_space,
+    delete_space,
+    generator_main,
+)
+from pynini.lib import pynutil
+
+import logging
+
+
+class ClassifyFst(GraphFst):
+    """
+    Final class that composes all other classification grammars. This class can process an entire sentence, that is lower cased.
+    For deployment, this grammar will be compiled and exported to OpenFst Finate State Archiv (FAR) File.
+    More details to deployment at NeMo/tools/text_processing_deployment.
+
+    Args:
+        cache_dir: path to a dir with .far grammar file. Set to None to avoid using cache.
+        overwrite_cache: set to True to overwrite .far files
+    """
+
+    def __init__(self, cache_dir: str = None, overwrite_cache: bool = False):
+        super().__init__(name="tokenize_and_classify", kind="classify")
+
+        far_file = None
+        if cache_dir is not None and cache_dir != "None":
+            os.makedirs(cache_dir, exist_ok=True)
+            far_file = os.path.join(cache_dir, "_en_itn.far")
+        if not overwrite_cache and far_file and os.path.exists(far_file):
+            self.fst = pynini.Far(far_file, mode="r")["tokenize_and_classify"]
+            logging.info(f"ClassifyFst.fst was restored from {far_file}.")
+        else:
+            logging.info(f"Creating ClassifyFst grammars.")
+            cardinal = CardinalFst()
+            cardinal_graph = cardinal.fst
+
+            ordinal = OrdinalFst(cardinal)
+            ordinal_graph = ordinal.fst
+
+            decimal = DecimalFst(cardinal)
+            decimal_graph = decimal.fst
+
+            measure_graph = MeasureFst(cardinal=cardinal, decimal=decimal).fst
+            date_graph = DateFst(ordinal=ordinal).fst
+            word_graph = WordFst().fst
+            time_graph = TimeFst().fst
+            money_graph = MoneyFst(cardinal=cardinal, decimal=decimal).fst
+            whitelist_graph = WhiteListFst().fst
+            punct_graph = PunctuationFst().fst
+            electronic_graph = ElectronicFst().fst
+            telephone_graph = TelephoneFst(cardinal).fst
+
+            classify = (
+                pynutil.add_weight(whitelist_graph, 1.01)
+                | pynutil.add_weight(time_graph, 1.1)
+                | pynutil.add_weight(date_graph, 1.09)
+                | pynutil.add_weight(decimal_graph, 1.1)
+                | pynutil.add_weight(measure_graph, 1.1)
+                | pynutil.add_weight(cardinal_graph, 1.1)
+                | pynutil.add_weight(ordinal_graph, 1.1)
+                | pynutil.add_weight(money_graph, 1.1)
+                | pynutil.add_weight(telephone_graph, 1.1)
+                | pynutil.add_weight(electronic_graph, 1.1)
+                | pynutil.add_weight(word_graph, 100)
+            )
+
+            punct = (
+                pynutil.insert("tokens { ")
+                + pynutil.add_weight(punct_graph, weight=1.1)
+                + pynutil.insert(" }")
+            )
+            token = pynutil.insert("tokens { ") + classify + pynutil.insert(" }")
+            token_plus_punct = (
+                pynini.closure(punct + pynutil.insert(" "))
+                + token
+                + pynini.closure(pynutil.insert(" ") + punct)
+            )
+
+            graph = token_plus_punct + pynini.closure(delete_extra_space + token_plus_punct)
+            graph = delete_space + graph + delete_space
+
+            self.fst = graph.optimize()
+
+            if far_file:
+                generator_main(far_file, {"tokenize_and_classify": self.fst})
+                logging.info(f"ClassifyFst grammars are saved to {far_file}.")

fun_text_processing/inverse_text_normalization/en/taggers/whitelist.py

@@ -0,0 +1,19 @@
+import pynini
+from fun_text_processing.inverse_text_normalization.en.utils import get_abs_path
+from fun_text_processing.text_normalization.en.graph_utils import GraphFst, convert_space
+from pynini.lib import pynutil
+
+
+class WhiteListFst(GraphFst):
+    """
+    Finite state transducer for classifying whitelisted tokens
+        e.g. misses -> tokens { name: "mrs." }
+    This class has highest priority among all classifier grammars. Whitelisted tokens are defined and loaded from "data/whitelist.tsv".
+    """
+
+    def __init__(self):
+        super().__init__(name="whitelist", kind="classify")
+
+        whitelist = pynini.string_file(get_abs_path("data/whitelist.tsv")).invert()
+        graph = pynutil.insert('name: "') + convert_space(whitelist) + pynutil.insert('"')
+        self.fst = graph.optimize()

fun_text_processing/inverse_text_normalization/en/taggers/word.py

@@ -0,0 +1,15 @@
+import pynini
+from fun_text_processing.text_normalization.en.graph_utils import DAMO_NOT_SPACE, GraphFst
+from pynini.lib import pynutil
+
+
+class WordFst(GraphFst):
+    """
+    Finite state transducer for classifying plain tokens, that do not belong to any special class. This can be considered as the default class.
+        e.g. sleep -> tokens { name: "sleep" }
+    """
+
+    def __init__(self):
+        super().__init__(name="word", kind="classify")
+        word = pynutil.insert('name: "') + pynini.closure(DAMO_NOT_SPACE, 1) + pynutil.insert('"')
+        self.fst = word.optimize()

fun_text_processing/inverse_text_normalization/en/utils.py

@@ -0,0 +1,33 @@
+import os
+from typing import Union
+
+import inflect
+
+_inflect = inflect.engine()
+
+
+def num_to_word(x: Union[str, int]):
+    """
+    converts integer to spoken representation
+
+    Args
+        x: integer
+
+    Returns: spoken representation
+    """
+    if isinstance(x, int):
+        x = str(x)
+        x = _inflect.number_to_words(str(x)).replace("-", " ").replace(",", "")
+    return x
+
+
+def get_abs_path(rel_path):
+    """
+    Get absolute path
+
+    Args:
+        rel_path: relative path to this file
+
+    Returns absolute path
+    """
+    return os.path.dirname(os.path.abspath(__file__)) + "/" + rel_path

fun_text_processing/inverse_text_normalization/en/verbalizers/__init__.py

@@ -0,0 +1 @@
+

fun_text_processing/inverse_text_normalization/en/verbalizers/cardinal.py

@@ -0,0 +1,38 @@
+import pynini
+from fun_text_processing.text_normalization.en.graph_utils import (
+    DAMO_NOT_QUOTE,
+    GraphFst,
+    delete_space,
+)
+from pynini.lib import pynutil
+
+
+class CardinalFst(GraphFst):
+    """
+    Finite state transducer for verbalizing cardinal
+        e.g. cardinal { integer: "23" negative: "-" } -> -23
+    """
+
+    def __init__(self):
+        super().__init__(name="cardinal", kind="verbalize")
+        optional_sign = pynini.closure(
+            pynutil.delete("negative:")
+            + delete_space
+            + pynutil.delete('"')
+            + DAMO_NOT_QUOTE
+            + pynutil.delete('"')
+            + delete_space,
+            0,
+            1,
+        )
+        graph = (
+            pynutil.delete("integer:")
+            + delete_space
+            + pynutil.delete('"')
+            + pynini.closure(DAMO_NOT_QUOTE, 1)
+            + pynutil.delete('"')
+        )
+        self.numbers = graph
+        graph = optional_sign + graph
+        delete_tokens = self.delete_tokens(graph)
+        self.fst = delete_tokens.optimize()

fun_text_processing/inverse_text_normalization/en/verbalizers/date.py

@@ -0,0 +1,70 @@
+import pynini
+from fun_text_processing.text_normalization.en.graph_utils import (
+    DAMO_NOT_QUOTE,
+    GraphFst,
+    delete_extra_space,
+    delete_space,
+)
+from pynini.lib import pynutil
+
+
+class DateFst(GraphFst):
+    """
+    Finite state transducer for verbalizing date, e.g.
+        date { month: "january" day: "5" year: "2012" preserve_order: true } -> february 5 2012
+        date { day: "5" month: "january" year: "2012" preserve_order: true } -> 5 february 2012
+    """
+
+    def __init__(self):
+        super().__init__(name="date", kind="verbalize")
+        month = (
+            pynutil.delete("month:")
+            + delete_space
+            + pynutil.delete('"')
+            + pynini.closure(DAMO_NOT_QUOTE, 1)
+            + pynutil.delete('"')
+        )
+        day = (
+            pynutil.delete("day:")
+            + delete_space
+            + pynutil.delete('"')
+            + pynini.closure(DAMO_NOT_QUOTE, 1)
+            + pynutil.delete('"')
+        )
+        year = (
+            pynutil.delete("year:")
+            + delete_space
+            + pynutil.delete('"')
+            + pynini.closure(DAMO_NOT_QUOTE, 1)
+            + delete_space
+            + pynutil.delete('"')
+        )
+
+        # month (day) year
+        graph_mdy = (
+            month
+            + pynini.closure(delete_extra_space + day, 0, 1)
+            + pynini.closure(delete_extra_space + year, 0, 1)
+        )
+
+        # (day) month year
+        graph_dmy = (
+            pynini.closure(day + delete_extra_space, 0, 1)
+            + month
+            + pynini.closure(delete_extra_space + year, 0, 1)
+        )
+
+        optional_preserve_order = pynini.closure(
+            pynutil.delete("preserve_order:") + delete_space + pynutil.delete("true") + delete_space
+            | pynutil.delete("field_order:")
+            + delete_space
+            + pynutil.delete('"')
+            + DAMO_NOT_QUOTE
+            + pynutil.delete('"')
+            + delete_space
+        )
+
+        final_graph = (graph_mdy | year | graph_dmy) + delete_space + optional_preserve_order
+
+        delete_tokens = self.delete_tokens(final_graph)
+        self.fst = delete_tokens.optimize()

fun_text_processing/inverse_text_normalization/en/verbalizers/decimal.py

@@ -0,0 +1,48 @@
+import pynini
+from fun_text_processing.text_normalization.en.graph_utils import (
+    DAMO_NOT_QUOTE,
+    GraphFst,
+    delete_space,
+)
+from pynini.lib import pynutil
+
+
+class DecimalFst(GraphFst):
+    """
+    Finite state transducer for verbalizing decimal, e.g.
+        decimal { negative: "true" integer_part: "12"  fractional_part: "5006" quantity: "billion" } -> -12.5006 billion
+    """
+
+    def __init__(self):
+        super().__init__(name="decimal", kind="verbalize")
+        optionl_sign = pynini.closure(pynini.cross('negative: "true"', "-") + delete_space, 0, 1)
+        integer = (
+            pynutil.delete("integer_part:")
+            + delete_space
+            + pynutil.delete('"')
+            + pynini.closure(DAMO_NOT_QUOTE, 1)
+            + pynutil.delete('"')
+        )
+        optional_integer = pynini.closure(integer + delete_space, 0, 1)
+        fractional = (
+            pynutil.insert(".")
+            + pynutil.delete("fractional_part:")
+            + delete_space
+            + pynutil.delete('"')
+            + pynini.closure(DAMO_NOT_QUOTE, 1)
+            + pynutil.delete('"')
+        )
+        optional_fractional = pynini.closure(fractional + delete_space, 0, 1)
+        quantity = (
+            pynutil.delete("quantity:")
+            + delete_space
+            + pynutil.delete('"')
+            + pynini.closure(DAMO_NOT_QUOTE, 1)
+            + pynutil.delete('"')
+        )
+        optional_quantity = pynini.closure(pynutil.insert(" ") + quantity + delete_space, 0, 1)
+        graph = optional_integer + optional_fractional + optional_quantity
+        self.numbers = graph
+        graph = optionl_sign + graph
+        delete_tokens = self.delete_tokens(graph)
+        self.fst = delete_tokens.optimize()

fun_text_processing/inverse_text_normalization/en/verbalizers/electronic.py

@@ -0,0 +1,45 @@
+import pynini
+from fun_text_processing.text_normalization.en.graph_utils import (
+    DAMO_NOT_QUOTE,
+    GraphFst,
+    delete_space,
+)
+from pynini.lib import pynutil
+
+
+class ElectronicFst(GraphFst):
+    """
+    Finite state transducer for verbalizing electronic
+        e.g. tokens { electronic { username: "cdf1" domain: "abc.edu" } } -> cdf1@abc.edu
+    """
+
+    def __init__(self):
+        super().__init__(name="electronic", kind="verbalize")
+        user_name = (
+            pynutil.delete("username:")
+            + delete_space
+            + pynutil.delete('"')
+            + pynini.closure(DAMO_NOT_QUOTE, 1)
+            + pynutil.delete('"')
+        )
+        domain = (
+            pynutil.delete("domain:")
+            + delete_space
+            + pynutil.delete('"')
+            + pynini.closure(DAMO_NOT_QUOTE, 1)
+            + pynutil.delete('"')
+        )
+
+        protocol = (
+            pynutil.delete("protocol:")
+            + delete_space
+            + pynutil.delete('"')
+            + pynini.closure(DAMO_NOT_QUOTE, 1)
+            + pynutil.delete('"')
+        )
+
+        graph = user_name + delete_space + pynutil.insert("@") + domain
+        graph |= protocol
+
+        delete_tokens = self.delete_tokens(graph)
+        self.fst = delete_tokens.optimize()

fun_text_processing/inverse_text_normalization/en/verbalizers/fraction.py

@@ -0,0 +1,10 @@
+from fun_text_processing.text_normalization.en.graph_utils import GraphFst
+
+
+class FractionFst(GraphFst):
+    """
+    Finite state transducer for verbalizing fraction,
+    """
+
+    def __init__(self):
+        super().__init__(name="fraction", kind="verbalize")

fun_text_processing/inverse_text_normalization/en/verbalizers/measure.py

@@ -0,0 +1,47 @@
+import pynini
+from fun_text_processing.text_normalization.en.graph_utils import DAMO_CHAR, GraphFst, delete_space
+from pynini.lib import pynutil
+
+
+class MeasureFst(GraphFst):
+    """
+    Finite state transducer for verbalizing measure, e.g.
+        measure { negative: "true" cardinal { integer: "12" } units: "kg" } -> -12 kg
+
+    Args:
+        decimal: DecimalFst
+        cardinal: CardinalFst
+    """
+
+    def __init__(self, decimal: GraphFst, cardinal: GraphFst):
+        super().__init__(name="measure", kind="verbalize")
+        optional_sign = pynini.closure(pynini.cross('negative: "true"', "-"), 0, 1)
+        unit = (
+            pynutil.delete("units:")
+            + delete_space
+            + pynutil.delete('"')
+            + pynini.closure(DAMO_CHAR - " ", 1)
+            + pynutil.delete('"')
+            + delete_space
+        )
+        graph_decimal = (
+            pynutil.delete("decimal {")
+            + delete_space
+            + optional_sign
+            + delete_space
+            + decimal.numbers
+            + delete_space
+            + pynutil.delete("}")
+        )
+        graph_cardinal = (
+            pynutil.delete("cardinal {")
+            + delete_space
+            + optional_sign
+            + delete_space
+            + cardinal.numbers
+            + delete_space
+            + pynutil.delete("}")
+        )
+        graph = (graph_cardinal | graph_decimal) + delete_space + pynutil.insert(" ") + unit
+        delete_tokens = self.delete_tokens(graph)
+        self.fst = delete_tokens.optimize()

fun_text_processing/inverse_text_normalization/en/verbalizers/money.py

@@ -0,0 +1,26 @@
+import pynini
+from fun_text_processing.text_normalization.en.graph_utils import DAMO_CHAR, GraphFst, delete_space
+from pynini.lib import pynutil
+
+
+class MoneyFst(GraphFst):
+    """
+    Finite state transducer for verbalizing money, e.g.
+        money { integer_part: "12" fractional_part: "05" currency: "$" } -> $12.05
+
+    Args:
+        decimal: DecimalFst
+    """
+
+    def __init__(self, decimal: GraphFst):
+        super().__init__(name="money", kind="verbalize")
+        unit = (
+            pynutil.delete("currency:")
+            + delete_space
+            + pynutil.delete('"')
+            + pynini.closure(DAMO_CHAR - " ", 1)
+            + pynutil.delete('"')
+        )
+        graph = unit + delete_space + decimal.numbers
+        delete_tokens = self.delete_tokens(graph)
+        self.fst = delete_tokens.optimize()

fun_text_processing/inverse_text_normalization/en/verbalizers/ordinal.py

@@ -0,0 +1,48 @@
+import pynini
+from fun_text_processing.text_normalization.en.graph_utils import (
+    DAMO_NOT_QUOTE,
+    DAMO_SIGMA,
+    GraphFst,
+    delete_space,
+)
+from pynini.lib import pynutil
+
+
+class OrdinalFst(GraphFst):
+    """
+    Finite state transducer for verbalizing ordinal, e.g.
+       ordinal { integer: "13" } -> 13th
+    """
+
+    def __init__(self):
+        super().__init__(name="ordinal", kind="verbalize")
+        graph = (
+            pynutil.delete("integer:")
+            + delete_space
+            + pynutil.delete('"')
+            + pynini.closure(DAMO_NOT_QUOTE, 1)
+            + pynutil.delete('"')
+        )
+        convert_eleven = pynini.cross("11", "11th")
+        convert_twelve = pynini.cross("12", "12th")
+        convert_thirteen = pynini.cross("13", "13th")
+        convert_one = pynini.cross("1", "1st")
+        convert_two = pynini.cross("2", "2nd")
+        convert_three = pynini.cross("3", "3rd")
+        convert_rest = pynutil.insert("th", weight=0.01)
+
+        suffix = pynini.cdrewrite(
+            convert_eleven
+            | convert_twelve
+            | convert_thirteen
+            | convert_one
+            | convert_two
+            | convert_three
+            | convert_rest,
+            "",
+            "[EOS]",
+            DAMO_SIGMA,
+        )
+        graph = graph @ suffix
+        delete_tokens = self.delete_tokens(graph)
+        self.fst = delete_tokens.optimize()

fun_text_processing/inverse_text_normalization/en/verbalizers/telephone.py

@@ -0,0 +1,30 @@
+import pynini
+from fun_text_processing.text_normalization.en.graph_utils import DAMO_NOT_QUOTE, GraphFst
+from pynini.lib import pynutil
+
+
+class TelephoneFst(GraphFst):
+    """
+    Finite state transducer for verbalizing telephone, e.g.
+        telephone { number_part: "123-123-5678" }
+        -> 123-123-5678
+    """
+
+    def __init__(self):
+        super().__init__(name="telephone", kind="verbalize")
+
+        number_part = (
+            pynutil.delete('number_part: "')
+            + pynini.closure(DAMO_NOT_QUOTE, 1)
+            + pynutil.delete('"')
+        )
+        optional_country_code = pynini.closure(
+            pynutil.delete('country_code: "')
+            + pynini.closure(DAMO_NOT_QUOTE, 1)
+            + pynutil.delete('"')
+            + pynini.accep(" "),
+            0,
+            1,
+        )
+        delete_tokens = self.delete_tokens(optional_country_code + number_part)
+        self.fst = delete_tokens.optimize()

fun_text_processing/inverse_text_normalization/en/verbalizers/time.py

@@ -0,0 +1,68 @@
+import pynini
+from fun_text_processing.text_normalization.en.graph_utils import (
+    DAMO_CHAR,
+    DAMO_DIGIT,
+    GraphFst,
+    delete_space,
+    insert_space,
+)
+from pynini.lib import pynutil
+
+
+class TimeFst(GraphFst):
+    """
+    Finite state transducer for verbalizing time, e.g.
+        time { hours: "12" minutes: "30" } -> 12:30
+        time { hours: "1" minutes: "12" } -> 01:12
+        time { hours: "2" suffix: "a.m." } -> 02:00 a.m.
+    """
+
+    def __init__(self):
+        super().__init__(name="time", kind="verbalize")
+        add_leading_zero_to_double_digit = (DAMO_DIGIT + DAMO_DIGIT) | (
+            pynutil.insert("0") + DAMO_DIGIT
+        )
+        hour = (
+            pynutil.delete("hours:")
+            + delete_space
+            + pynutil.delete('"')
+            + pynini.closure(DAMO_DIGIT, 1)
+            + pynutil.delete('"')
+        )
+        minute = (
+            pynutil.delete("minutes:")
+            + delete_space
+            + pynutil.delete('"')
+            + pynini.closure(DAMO_DIGIT, 1)
+            + pynutil.delete('"')
+        )
+        suffix = (
+            delete_space
+            + insert_space
+            + pynutil.delete("suffix:")
+            + delete_space
+            + pynutil.delete('"')
+            + pynini.closure(DAMO_CHAR - " ", 1)
+            + pynutil.delete('"')
+        )
+        optional_suffix = pynini.closure(suffix, 0, 1)
+        zone = (
+            delete_space
+            + insert_space
+            + pynutil.delete("zone:")
+            + delete_space
+            + pynutil.delete('"')
+            + pynini.closure(DAMO_CHAR - " ", 1)
+            + pynutil.delete('"')
+        )
+        optional_zone = pynini.closure(zone, 0, 1)
+        graph = (
+            hour @ add_leading_zero_to_double_digit
+            + delete_space
+            + pynutil.insert(":")
+            + (minute @ add_leading_zero_to_double_digit)
+            + optional_suffix
+            + optional_zone
+        )
+        delete_tokens = self.delete_tokens(graph)
+        self.fst = delete_tokens.optimize()

fun_text_processing/inverse_text_normalization/en/verbalizers/verbalize.py

@@ -0,0 +1,47 @@
+from fun_text_processing.inverse_text_normalization.en.verbalizers.cardinal import CardinalFst
+from fun_text_processing.inverse_text_normalization.en.verbalizers.date import DateFst
+from fun_text_processing.inverse_text_normalization.en.verbalizers.decimal import DecimalFst
+from fun_text_processing.inverse_text_normalization.en.verbalizers.electronic import ElectronicFst
+from fun_text_processing.inverse_text_normalization.en.verbalizers.measure import MeasureFst
+from fun_text_processing.inverse_text_normalization.en.verbalizers.money import MoneyFst
+from fun_text_processing.inverse_text_normalization.en.verbalizers.ordinal import OrdinalFst
+from fun_text_processing.inverse_text_normalization.en.verbalizers.telephone import TelephoneFst
+from fun_text_processing.inverse_text_normalization.en.verbalizers.time import TimeFst
+from fun_text_processing.inverse_text_normalization.en.verbalizers.whitelist import WhiteListFst
+from fun_text_processing.text_normalization.en.graph_utils import GraphFst
+
+
+class VerbalizeFst(GraphFst):
+    """
+    Composes other verbalizer grammars.
+    For deployment, this grammar will be compiled and exported to OpenFst Finate State Archiv (FAR) File.
+    More details to deployment at NeMo/tools/text_processing_deployment.
+    """
+
+    def __init__(self):
+        super().__init__(name="verbalize", kind="verbalize")
+        cardinal = CardinalFst()
+        cardinal_graph = cardinal.fst
+        ordinal_graph = OrdinalFst().fst
+        decimal = DecimalFst()
+        decimal_graph = decimal.fst
+        measure_graph = MeasureFst(decimal=decimal, cardinal=cardinal).fst
+        money_graph = MoneyFst(decimal=decimal).fst
+        time_graph = TimeFst().fst
+        date_graph = DateFst().fst
+        whitelist_graph = WhiteListFst().fst
+        telephone_graph = TelephoneFst().fst
+        electronic_graph = ElectronicFst().fst
+        graph = (
+            time_graph
+            | date_graph
+            | money_graph
+            | measure_graph
+            | ordinal_graph
+            | decimal_graph
+            | cardinal_graph
+            | whitelist_graph
+            | telephone_graph
+            | electronic_graph
+        )
+        self.fst = graph

fun_text_processing/inverse_text_normalization/en/verbalizers/verbalize_final.py

@@ -0,0 +1,33 @@
+import pynini
+from fun_text_processing.inverse_text_normalization.en.verbalizers.verbalize import VerbalizeFst
+from fun_text_processing.inverse_text_normalization.en.verbalizers.word import WordFst
+from fun_text_processing.text_normalization.en.graph_utils import (
+    GraphFst,
+    delete_extra_space,
+    delete_space,
+)
+from pynini.lib import pynutil
+
+
+class VerbalizeFinalFst(GraphFst):
+    """
+    Finite state transducer that verbalizes an entire sentence, e.g.
+    tokens { name: "its" } tokens { time { hours: "12" minutes: "30" } } tokens { name: "now" } -> its 12:30 now
+    """
+
+    def __init__(self):
+        super().__init__(name="verbalize_final", kind="verbalize")
+        verbalize = VerbalizeFst().fst
+        word = WordFst().fst
+        types = verbalize | word
+        graph = (
+            pynutil.delete("tokens")
+            + delete_space
+            + pynutil.delete("{")
+            + delete_space
+            + types
+            + delete_space
+            + pynutil.delete("}")
+        )
+        graph = delete_space + pynini.closure(graph + delete_extra_space) + graph + delete_space
+        self.fst = graph

fun_text_processing/inverse_text_normalization/en/verbalizers/whitelist.py

@@ -0,0 +1,27 @@
+import pynini
+from fun_text_processing.text_normalization.en.graph_utils import (
+    DAMO_CHAR,
+    DAMO_SIGMA,
+    GraphFst,
+    delete_space,
+)
+from pynini.lib import pynutil
+
+
+class WhiteListFst(GraphFst):
+    """
+    Finite state transducer for verbalizing whitelist
+        e.g. tokens { name: "mrs." } -> mrs.
+    """
+
+    def __init__(self):
+        super().__init__(name="whitelist", kind="verbalize")
+        graph = (
+            pynutil.delete("name:")
+            + delete_space
+            + pynutil.delete('"')
+            + pynini.closure(DAMO_CHAR - " ", 1)
+            + pynutil.delete('"')
+        )
+        graph = graph @ pynini.cdrewrite(pynini.cross("\u00A0", " "), "", "", DAMO_SIGMA)
+        self.fst = graph.optimize()

fun_text_processing/inverse_text_normalization/en/verbalizers/word.py

@@ -0,0 +1,29 @@
+import pynini
+from fun_text_processing.text_normalization.en.graph_utils import (
+    DAMO_CHAR,
+    DAMO_SIGMA,
+    GraphFst,
+    delete_space,
+)
+from pynini.lib import pynutil
+
+
+class WordFst(GraphFst):
+    """
+    Finite state transducer for verbalizing plain tokens
+        e.g. tokens { name: "sleep" } -> sleep
+    """
+
+    def __init__(self):
+        super().__init__(name="word", kind="verbalize")
+        chars = pynini.closure(DAMO_CHAR - " ", 1)
+        char = (
+            pynutil.delete("name:")
+            + delete_space
+            + pynutil.delete('"')
+            + chars
+            + pynutil.delete('"')
+        )
+        graph = char @ pynini.cdrewrite(pynini.cross("\u00A0", " "), "", "", DAMO_SIGMA)
+
+        self.fst = graph.optimize()

fun_text_processing/inverse_text_normalization/es/__init__.py

@@ -0,0 +1,7 @@
+from fun_text_processing.inverse_text_normalization.es.taggers.tokenize_and_classify import (
+    ClassifyFst,
+)
+from fun_text_processing.inverse_text_normalization.es.verbalizers.verbalize import VerbalizeFst
+from fun_text_processing.inverse_text_normalization.es.verbalizers.verbalize_final import (
+    VerbalizeFinalFst,
+)

fun_text_processing/inverse_text_normalization/es/data/__init__.py

@@ -0,0 +1 @@
+

fun_text_processing/inverse_text_normalization/es/data/currency_plural.tsv

@@ -0,0 +1,6 @@
+€	euros
+US$	dólares estadounidenses
+US$	dólares americanos
+$	dólares
+$	pesos
+¥	yenes

fun_text_processing/inverse_text_normalization/es/data/currency_singular.tsv

@@ -0,0 +1,6 @@
+€	euro
+US$	dólar estadounidense
+US$	dólar americano
+$	dólar
+$	peso
+¥	yen

fun_text_processing/inverse_text_normalization/es/data/electronic/__init__.py

@@ -0,0 +1 @@
+

fun_text_processing/inverse_text_normalization/es/data/electronic/domain.tsv

@@ -0,0 +1,25 @@
+com
+es
+uk
+fr
+net
+br
+in
+ru
+de
+it
+edu
+co
+ar
+bo
+cl
+co
+ec
+fk
+gf
+fy
+pe
+py
+sr
+ve
+uy

fun_text_processing/inverse_text_normalization/es/data/electronic/server_name.tsv

@@ -0,0 +1,17 @@
+gmail	g mail
+gmail
+nvidia	n vidia
+nvidia
+outlook
+hotmail
+yahoo
+aol
+gmx
+msn
+live
+yandex
+orange
+wanadoo
+web
+comcast
+bbc

fun_text_processing/inverse_text_normalization/es/data/electronic/symbols.tsv

@@ -0,0 +1,4 @@
+.	punto
+-	guion
+_	guion bajo
+/	barra

fun_text_processing/inverse_text_normalization/es/data/measurements_plural.tsv

@@ -0,0 +1,19 @@
+cm	centímetros
+g	gramos
+h	horas
+kg	kilos
+kg	kilogramos
+km	kilómetros
+km²	kilómetros cuadrados
+l	litros
+m	metros
+m²	metros cuadrados
+m³	metros cubicos
+mph	millas por hora
+ml	mililitros
+mm	milímetros
+ms	milisegundos
+min	minutos
+%	por ciento
+%	porciento
+s	segundos

fun_text_processing/inverse_text_normalization/es/data/measurements_singular.tsv

@@ -0,0 +1,19 @@
+cm	centímetro
+g	gramo
+h	hora
+kg	kilo
+kg	kilogramo
+km	kilómetro
+km²	kilómetro cuadrado
+l	litro
+m	metro
+m²	metro cuadrado
+m³	metro cubico
+mph	milla por hora
+ml	mililitro
+mm	milímetro
+ms	milisegundo
+min	minuto
+%	por ciento
+%	porciento
+s	segundo

fun_text_processing/inverse_text_normalization/es/data/months.tsv

@@ -0,0 +1,12 @@
+enero
+febrero
+marzo
+abril
+mayo
+junio
+julio
+agosto
+septiembre
+octubre
+noviembre
+diciembre

fun_text_processing/inverse_text_normalization/es/data/numbers/__init__.py

@@ -0,0 +1 @@
+

fun_text_processing/inverse_text_normalization/es/data/numbers/digit.tsv

@@ -0,0 +1,12 @@
+uno	1
+un	1
+ún	1
+una	1
+dos	2
+tres	3
+cuatro	4
+cinco	5
+seis	6
+siete	7
+ocho	8
+nueve	9

fun_text_processing/inverse_text_normalization/es/data/numbers/hundreds.tsv

@@ -0,0 +1,18 @@
+ciento	1
+cien	1
+doscientos	2
+doscientas	2
+trescientos	3
+trescientas	3
+cuatrocientos	4
+cuatrocientas	4
+quinientos	5
+quinientas	5
+seiscientos	6
+seiscientas	6
+setecientos	7
+setecientas	7
+ochocientos	8
+ochocientas	8
+novecientos	9
+novecientas	9

fun_text_processing/inverse_text_normalization/es/data/numbers/teen.tsv

@@ -0,0 +1,10 @@
+diez	10
+once	11
+doce	12
+trece	13
+catorce	14
+quince	15
+dieciséis	16
+diecisiete	17
+dieciocho	18
+diecinueve	19

fun_text_processing/inverse_text_normalization/es/data/numbers/ties.tsv

@@ -0,0 +1,9 @@
+veinte	2
+treinta	3
+cuarenta	4
+cincuenta	5
+cinquenta	5
+sesenta	6
+setenta	7
+ochenta	8
+noventa	9