storytold/storyteller-ml
2
0
Fork 0
mirror of https://github.com/storytold/storyteller-ml.git synced 2026-10-29 09:18:21 +00:00
Code Issues Packages Projects Releases Wiki Activity

wip

Browse Source
This commit is contained in:
Brandon Thomas
2022-07-04 00:40:12 -04:00
parent 1bfc4f9959
commit 7529181864
5 changed files with 42 additions and 32 deletions
Download Patch File Download Diff File Expand all files Collapse all files
tts/text/input_parsing/parser.py
+4 -4
View File
@@ -17,11 +17,11 @@ class ArpabetSequence:
def __init__(self, sequence: str):
self.sequence = sequence
class GraphemeSquence:
class GraphemeSequence:
def __init__(self, sequence: str):
self.sequence = sequence
ParsedSequence = List[Union[ArpabetSequence, GraphemeSquence]]
ParsedSequence = List[Union[ArpabetSequence, GraphemeSequence]]
def parse_grapheme_and_arpabet_sequence(input_text: str) -> ParsedSequence:
"""
@@ -35,7 +35,7 @@ def parse_grapheme_and_arpabet_sequence(input_text: str) -> ParsedSequence:
input_text = input_text.strip()
m = _curly_re.match(input_text)
if not m:
sequence.append(GraphemeSquence(input_text))
sequence.append(GraphemeSequence(input_text))
break
before = m.group(1).strip()
@@ -43,7 +43,7 @@ def parse_grapheme_and_arpabet_sequence(input_text: str) -> ParsedSequence:
after = m.group(3).strip()
if before:
sequence.append(GraphemeSquence(before))
sequence.append(GraphemeSequence(before))
if bracketed:
sequence.append(ArpabetSequence(bracketed))
tts/text/input_parsing/test_parser.py
+9 -9
View File
@@ -1,11 +1,11 @@
from tts.text.input_parsing.parser import ArpabetSequence, GraphemeSquence, parse_grapheme_and_arpabet_sequence
from tts.text.input_parsing.parser import ArpabetSequence, GraphemeSequence, parse_grapheme_and_arpabet_sequence
class TestParser:
def test_only_text(self):
parsed = parse_grapheme_and_arpabet_sequence("This is a test")
assert len(parsed) == 1
assert isinstance(parsed[0], GraphemeSquence)
assert isinstance(parsed[0], GraphemeSequence)
def test_only_arpabet(self):
parsed = parse_grapheme_and_arpabet_sequence("{T EH1 S T IH0 NG}")
@@ -15,31 +15,31 @@ class TestParser:
def test_text_and_arpabet(self):
parsed = parse_grapheme_and_arpabet_sequence("This is a {T EH1 S T}")
assert len(parsed) == 2
assert isinstance(parsed[0], GraphemeSquence)
assert isinstance(parsed[0], GraphemeSequence)
assert isinstance(parsed[1], ArpabetSequence)
def test_text_and_arpabet_interleved_1(self):
parsed = parse_grapheme_and_arpabet_sequence("{DH IH1 S} is a {T EH1 S T}")
assert len(parsed) == 3
assert isinstance(parsed[0], ArpabetSequence)
assert isinstance(parsed[1], GraphemeSquence)
assert isinstance(parsed[1], GraphemeSequence)
assert isinstance(parsed[2], ArpabetSequence)
def test_text_and_arpabet_interleved_2(self):
parsed = parse_grapheme_and_arpabet_sequence("Check out {DH IH1 S} {T EH1 S T} of the {S IH1 S T AH0 M}")
assert len(parsed) == 5
assert isinstance(parsed[0], GraphemeSquence)
assert isinstance(parsed[0], GraphemeSequence)
assert isinstance(parsed[1], ArpabetSequence)
assert isinstance(parsed[2], ArpabetSequence)
assert isinstance(parsed[3], GraphemeSquence)
assert isinstance(parsed[3], GraphemeSequence)
assert isinstance(parsed[4], ArpabetSequence)
def test_text_and_arpabet_interleved_2(self):
parsed = parse_grapheme_and_arpabet_sequence("Check out {DH IH1 S} {T EH1 S T} of the {S IH1 S T AH0 M}.")
assert len(parsed) == 6
assert isinstance(parsed[0], GraphemeSquence)
assert isinstance(parsed[0], GraphemeSequence)
assert isinstance(parsed[1], ArpabetSequence)
assert isinstance(parsed[2], ArpabetSequence)
assert isinstance(parsed[3], GraphemeSquence)
assert isinstance(parsed[3], GraphemeSequence)
assert isinstance(parsed[4], ArpabetSequence)
assert isinstance(parsed[5], GraphemeSquence)
assert isinstance(parsed[5], GraphemeSequence)