File size: 5,742 Bytes
2e9bf0c
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
import os
import re

import cn2an
from pypinyin import lazy_pinyin, Style


from .symbols import punctuation
from .tone_sandhi import ToneSandhi

current_file_path = os.path.dirname(__file__)
pinyin_to_symbol_map = {
    line.split("\t")[0]: line.strip().split("\t")[1]
    for line in open(os.path.join(current_file_path, "opencpop-strict.txt")).readlines()
}

import jieba.posseg as psg


rep_map = {
    ":": ",",
    ";": ",",
    ",": ",",
    "。": ".",
    "!": "!",
    "?": "?",
    "\n": ".",
    "·": ",",
    "、": ",",
    "...": "…",
    "$": ".",
    "“": "'",
    "”": "'",
    "‘": "'",
    "’": "'",
    "(": "'",
    ")": "'",
    "(": "'",
    ")": "'",
    "《": "'",
    "》": "'",
    "【": "'",
    "】": "'",
    "[": "'",
    "]": "'",
    "—": "-",
    "~": "-",
    "~": "-",
    "「": "'",
    "」": "'",
}

tone_modifier = ToneSandhi()


def replace_punctuation(text):
    text = text.replace("嗯", "恩").replace("呣", "母")
    pattern = re.compile("|".join(re.escape(p) for p in rep_map.keys()))

    replaced_text = pattern.sub(lambda x: rep_map[x.group()], text)

    replaced_text = re.sub(
        r"[^\u4e00-\u9fa5" + "".join(punctuation) + r"]+", "", replaced_text
    )

    return replaced_text


def g2p(text):
    pattern = r"(?<=[{0}])\s*".format("".join(punctuation))
    sentences = [i for i in re.split(pattern, text) if i.strip() != ""]
    phones, tones, word2ph = _g2p(sentences)
    assert sum(word2ph) == len(phones)
    assert len(word2ph) == len(text)  # Sometimes it will crash,you can add a try-catch.
    phones = ["_"] + phones + ["_"]
    tones = [0] + tones + [0]
    word2ph = [1] + word2ph + [1]
    return phones, tones, word2ph


def _get_initials_finals(word):
    initials = []
    finals = []
    orig_initials = lazy_pinyin(word, neutral_tone_with_five=True, style=Style.INITIALS)
    orig_finals = lazy_pinyin(
        word, neutral_tone_with_five=True, style=Style.FINALS_TONE3
    )
    for c, v in zip(orig_initials, orig_finals):
        initials.append(c)
        finals.append(v)
    return initials, finals


def _g2p(segments):
    phones_list = []
    tones_list = []
    word2ph = []
    for seg in segments:
        # Replace all English words in the sentence
        seg = re.sub("[a-zA-Z]+", "", seg)
        seg_cut = psg.lcut(seg)
        initials = []
        finals = []
        seg_cut = tone_modifier.pre_merge_for_modify(seg_cut)
        for word, pos in seg_cut:
            if pos == "eng":
                continue
            sub_initials, sub_finals = _get_initials_finals(word)
            sub_finals = tone_modifier.modified_tone(word, pos, sub_finals)
            initials.append(sub_initials)
            finals.append(sub_finals)

            # assert len(sub_initials) == len(sub_finals) == len(word)
        initials = sum(initials, [])
        finals = sum(finals, [])
        #
        for c, v in zip(initials, finals):
            raw_pinyin = c + v
            # NOTE: post process for pypinyin outputs
            # we discriminate i, ii and iii
            if c == v:
                assert c in punctuation
                phone = [c]
                tone = "0"
                word2ph.append(1)
            else:
                v_without_tone = v[:-1]
                tone = v[-1]

                pinyin = c + v_without_tone
                assert tone in "12345"

                if c:
                    # 多音节
                    v_rep_map = {
                        "uei": "ui",
                        "iou": "iu",
                        "uen": "un",
                    }
                    if v_without_tone in v_rep_map.keys():
                        pinyin = c + v_rep_map[v_without_tone]
                else:
                    # 单音节
                    pinyin_rep_map = {
                        "ing": "ying",
                        "i": "yi",
                        "in": "yin",
                        "u": "wu",
                    }
                    if pinyin in pinyin_rep_map.keys():
                        pinyin = pinyin_rep_map[pinyin]
                    else:
                        single_rep_map = {
                            "v": "yu",
                            "e": "e",
                            "i": "y",
                            "u": "w",
                        }
                        if pinyin[0] in single_rep_map.keys():
                            pinyin = single_rep_map[pinyin[0]] + pinyin[1:]

                assert pinyin in pinyin_to_symbol_map.keys(), (pinyin, seg, raw_pinyin)
                phone = pinyin_to_symbol_map[pinyin].split(" ")
                word2ph.append(len(phone))

            phones_list += phone
            tones_list += [int(tone)] * len(phone)
    return phones_list, tones_list, word2ph


def text_normalize(text):
    numbers = re.findall(r"\d+(?:\.?\d+)?", text)
    for number in numbers:
        text = text.replace(number, cn2an.an2cn(number), 1)
    text = replace_punctuation(text)
    return text


def get_bert_feature(text, word2ph):
    from text import chinese_bert

    return chinese_bert.get_bert_feature(text, word2ph)


if __name__ == "__main__":
    from text.chinese_bert import get_bert_feature

    text = "啊!但是《原神》是由,米哈\游自主,  [研发]的一款全.新开放世界.冒险游戏"
    text = text_normalize(text)
    print(text)
    phones, tones, word2ph = g2p(text)
    bert = get_bert_feature(text, word2ph)

    print(phones, tones, word2ph, bert.shape)


# # 示例用法
# text = "这是一个示例文本:,你好!这是一个测试...."
# print(g2p_paddle(text))  # 输出: 这是一个示例文本你好这是一个测试