Skip to content

Commit cdbfd03

Browse files
committed
minor fixes for the tagger files
Signed-off-by: Namrata Gachchi <ngachchi@nvidia.com>
1 parent 331c723 commit cdbfd03

File tree

2 files changed

+11
-3
lines changed

2 files changed

+11
-3
lines changed

nemo_text_processing/text_normalization/hi/taggers/measure.py

Lines changed: 6 additions & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -19,6 +19,11 @@
1919
from nemo_text_processing.text_normalization.hi.utils import get_abs_path
2020

2121

22+
digit = pynini.string_file(get_abs_path("data/numbers/digit.tsv"))
23+
teens_ties = pynini.string_file(get_abs_path("data/numbers/teens_and_ties.tsv"))
24+
teens_and_ties = pynutil.add_weight(teens_ties, -0.1)
25+
26+
2227
class MeasureFst(GraphFst):
2328
"""
2429
Finite state transducer for classifying measure, suppletive aware, e.g.
@@ -35,7 +40,7 @@ class MeasureFst(GraphFst):
3540
def __init__(self, cardinal: GraphFst, decimal: GraphFst):
3641
super().__init__(name="measure", kind="classify")
3742

38-
cardinal_graph = cardinal.final_graph
43+
cardinal_graph = digit | teens_and_ties | cardinal.graph_hundreds | cardinal.graph_thousands
3944
decimal_graph = decimal.final_graph_wo_negative
4045
unit_graph = pynini.string_file(get_abs_path("data/measure/unit.tsv"))
4146

nemo_text_processing/text_normalization/hi/taggers/money.py

Lines changed: 5 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -41,14 +41,17 @@ def __init__(self, cardinal: GraphFst):
4141

4242
cardinal_graph = cardinal.final_graph
4343

44+
optional_graph_negative = pynini.closure(
45+
pynutil.insert("negative: ") + pynini.cross("-", "\"true\"") + insert_space, 0, 1,
46+
)
4447
currency_major = pynutil.insert('currency_maj: "') + currency_graph + pynutil.insert('"')
4548
integer = pynutil.insert('integer_part: "') + cardinal_graph + pynutil.insert('"')
4649
fraction = pynutil.insert('fractional_part: "') + cardinal_graph + pynutil.insert('"')
4750
currency_minor = pynutil.insert('currency_min: "') + pynutil.insert("centiles") + pynutil.insert('"')
4851

49-
graph_major_only = currency_major + insert_space + integer
52+
graph_major_only = optional_graph_negative + currency_major + insert_space + integer
5053
graph_major_and_minor = (
51-
currency_major + insert_space + integer + pynini.cross(".", " ") + fraction + insert_space + currency_minor
54+
optional_graph_negative + currency_major + insert_space + integer + pynini.cross(".", " ") + fraction + insert_space + currency_minor
5255
)
5356

5457
graph_currencies = graph_major_only | graph_major_and_minor

0 commit comments

Comments
 (0)