Skip to content

Commit

Permalink
Update lyrics storage format & fix
Browse files Browse the repository at this point in the history
  • Loading branch information
@chenmozhijin
chenmozhijin committed Mar 6, 2024
1 parent d046bf6 commit 25d23ca
Showing 1 changed file with 45 additions and 43 deletions.
88 changes: 45 additions & 43 deletions utils/lyrics.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -61,11 +61,11 @@ def get_clear_lyric(lyric: str) -> str:


def qrc2list(qrc: str) -> tuple[dict, list]:
"""将qrc转换为列表[[行起始时间, 行结束时间, [字起始时间, 字结束时间, 字内容]]]"""
"""将qrc转换为列表[(行起始时间, 行结束时间, [(字起始时间, 字结束时间, 字内容)])]"""
qrc = re.findall(r'<Lyric_1 LyricType="1" LyricContent="(.*?)"/>', qrc, re.DOTALL)[0]
qrc_lines = qrc.split('\n')
tags = {}
lrc_list = []
lrc_list: list[tuple[int, int, list[tuple[int, int, str]]]] = []
wrods_split_pattern = re.compile(r'(?:\[\d+,\d+\])?((?:(?!\(\d+,\d+\)).)+)\((\d+),(\d+)\)') # 逐字匹配
line_split_pattern = re.compile(r'^\[(\d+),(\d+)\](.*)$') # 逐行匹配
tag_split_pattern = re.compile(r"^\[(\w+):([^\]]*)\]$")
Expand All @@ -75,14 +75,14 @@ def qrc2list(qrc: str) -> tuple[dict, list]:
line_split_content = re.findall(line_split_pattern, line)
if line_split_content: # 判断是否为歌词行
line_start_time, line_duration, line_content = line_split_content[0]
lrc_list.append([int(line_start_time), int(line_start_time) + int(line_duration), []])
lrc_list.append((int(line_start_time), int(line_start_time) + int(line_duration), []))
wrods_split_content = re.findall(wrods_split_pattern, line)
if wrods_split_content: # 判断是否为逐字歌词
for text, starttime, duration in wrods_split_content:
if text != "\r":
lrc_list[-1][2].append([int(starttime), int(starttime) + int(duration), text])
lrc_list[-1][2].append((int(starttime), int(starttime) + int(duration), text))
else: # 如果不是逐字歌词
lrc_list[-1][2].append([int(line_start_time), int(line_start_time) + int(line_duration), line_content])
lrc_list[-1][2].append((int(line_start_time), int(line_start_time) + int(line_duration), line_content))
else:
tag_split_content = re.findall(tag_split_pattern, line)
if tag_split_content:
Expand All @@ -92,8 +92,8 @@ def qrc2list(qrc: str) -> tuple[dict, list]:


def yrc2list(yrc: str) -> list:
"""将yrc转换为列表[[行起始时间, 行结束时间, [字起始时间, 字结束时间, 字内容]]]"""
lrc_list = []
"""将yrc转换为列表[(行起始时间, 行结束时间, [(字起始时间, 字结束时间, 字内容)])]"""
lrc_list: list[tuple[int, int, list[tuple[int, int, str]]]] = []

line_split_pattern = re.compile(r'^\[(\d+),(\d+)\](.*)$') # 逐行匹配
wrods_split_pattern = re.compile(r'(?:\[\d+,\d+\])?\((\d+),(\d+),\d+\)((?:.(?!\d+,\d+,\d+\)))*)') # 逐字匹配
Expand All @@ -106,22 +106,22 @@ def yrc2list(yrc: str) -> list:
if not line_split_content:
continue
line_start_time, line_duration, line_content = line_split_content[0]
lrc_list.append([int(line_start_time), int(line_start_time) + int(line_duration), []])
lrc_list.append((int(line_start_time), int(line_start_time) + int(line_duration), []))

wrods_split_content = re.findall(wrods_split_pattern, line_content)
if not wrods_split_content:
lrc_list[-1][2].append([int(line_start_time), int(line_start_time) + int(line_duration), line_content])
lrc_list[-1][2].append((int(line_start_time), int(line_start_time) + int(line_duration), line_content))
continue

for word_start_time, word_duration, word_content in wrods_split_content:
lrc_list[-1][2].append([int(word_start_time), int(word_start_time) + int(word_duration), word_content])
lrc_list[-1][2].append((int(word_start_time), int(word_start_time) + int(word_duration), word_content))

return lrc_list


def lrc2list(lrc: str) -> tuple[dict, list]:
"""将lrc转换为列表[[行起始时间, 行结束时间, [字起始时间, 字结束时间, 字内容]]]"""
lrc_list = []
"""将lrc转换为列表[(行起始时间, 行结束时间, [(字起始时间, 字结束时间, 字内容)])]"""
lrc_list: list[tuple[int, int | None, list[tuple[int, int | None, str]]]] = []
tags = {}

tag_split_pattern = re.compile(r"^\[(\w+):([^\]]*)\]$")
Expand Down Expand Up @@ -150,11 +150,11 @@ def lrc2list(lrc: str) -> tuple[dict, list]:
m, s, ms, line_content, m2, s2, ms2 = line_split_withend_content[0]
line_start_time = time2ms(m, s, ms)
line_end_time = time2ms(m2, s2, ms2)
lrc_list.append([line_start_time, line_end_time, []])
lrc_list.append((line_start_time, line_end_time, []))

wrods_split_contents = re.findall(wrods_split_pattern, line_content)
if not wrods_split_contents:
lrc_list[-1][2].append([line_start_time, line_end_time, line_content])
lrc_list[-1][2].append((line_start_time, line_end_time, line_content))
continue

line = []
Expand All @@ -163,30 +163,30 @@ def lrc2list(lrc: str) -> tuple[dict, list]:
word_start_time = time2ms(m, s, ms)
if line:
line[-1][1] = word_start_time
lrc_list.append([word_start_time, None, word_content])
lrc_list.append((word_start_time, None, word_content))

lrc_list.append(line)
return tags, lrc_list


def plaintext2list(plaintext: str) -> list[list[None, None, list[None, None, str]]]:
lrc_list = []
lrc_list: list[tuple[None, None, list[tuple[None, None, str]]]] = []
for line in plaintext.splitlines():
lrc_list.append([None, None, [[None, None, line]]])
lrc_list.append((None, None, [(None, None, line)]))
return lrc_list


def krc2dict(krc: str) -> tuple[dict, dict]:
"""将明文krc转换为字典{歌词类型: [[行起始时间, 行结束时间, [字起始时间, 字结束时间, 字内容]]]}"""
lrc_dict: dict[str: list[list[int, int, list[list[int, int, str]]]]] = {}
"""将明文krc转换为字典{歌词类型: [(行起始时间, 行结束时间, [(字起始时间, 字结束时间, 字内容)])]}"""
lrc_dict: dict[str: list[tuple[int, int, list[tuple[int, int, str]]]]] = {}
tag_split_pattern = re.compile(r"^\[(\w+):([^\]]*)\]$")
tags: dict[str: str] = {}

line_split_pattern = re.compile(r'^\[(\d+),(\d+)\](.*)$') # 逐行匹配
wrods_split_pattern = re.compile(r'(?:\[\d+,\d+\])?<(\d+),(\d+),\d+>((?:.(?!\d+,\d+,\d+>))*)') # 逐字匹配
orig_list: list[list[int, int, list[list[int, int, str]]]] = [] # 原文歌词
roma_list: list[list[int, int, list[list[int, int, str]]]] = []
ts_list: list[list[int, int, list[list[int, int, str]]]] = []
orig_list: list[tuple[int, int, list[tuple[int, int, str]]]] = [] # 原文歌词
roma_list: list[tuple[int, int, list[tuple[int, int, str]]]] = []
ts_list: list[tuple[int, int, list[tuple[int, int, str]]]] = []

for i in krc.splitlines():
line = i.strip()
Expand All @@ -202,28 +202,28 @@ def krc2dict(krc: str) -> tuple[dict, dict]:
if not line_split_content:
continue
line_start_time, line_duration, line_content = line_split_content[0]
orig_list.append([int(line_start_time), int(line_start_time) + int(line_duration), []])
orig_list.append((int(line_start_time), int(line_start_time) + int(line_duration), []))

wrods_split_content = re.findall(wrods_split_pattern, line_content)
if not wrods_split_content:
orig_list[-1][2].append([int(line_start_time), int(line_start_time) + int(line_duration), line_content])
orig_list[-1][2].append((int(line_start_time), int(line_start_time) + int(line_duration), line_content))
continue

for word_start_time, word_duration, word_content in wrods_split_content:
orig_list[-1][2].append([int(line_start_time) + int(word_start_time), int(line_start_time) + int(word_start_time) + int(word_duration), word_content])
orig_list[-1][2].append((int(line_start_time) + int(word_start_time), int(line_start_time) + int(word_start_time) + int(word_duration), word_content))

if "language" in tags and tags["language"].strip() != "":
languages = json.loads(b64decode(tags["language"].strip()))
for language in languages["content"]:
if language["type"] == 0: # 逐字(罗马音)
for i, line in enumerate(orig_list):
roma_line = [line[0], line[1], []]
roma_line = (line[0], line[1], [])
for j, word in enumerate(line[2]):
roma_line[2].append([word[0], word[1], language["lyricContent"][i][j]])
roma_line[2].append((word[0], word[1], language["lyricContent"][i][j]))
roma_list.append(roma_line)
elif language["type"] == 1: # 逐行(翻译)
for i, line in enumerate(orig_list):
ts_list.append([line[0], line[1], [[line[0], line[1], language["lyricContent"][i][0]]]])
ts_list.append((line[0], line[1], [(line[0], line[1], language["lyricContent"][i][0])]))

tags_str = ""
for key, value in tags.items():
Expand Down Expand Up @@ -262,7 +262,7 @@ def lrclist2str(lrc_list: list) -> str:
return lrc_str


def linelist2str(line_list: list[int | None, int | None, list[int | None, int | None, str]]) -> str:
def linelist2str(line_list: list[tuple[int, int | None, list[tuple[int, int | None, str]]]]) -> str:
lrc_str = ""
if line_list[0] is None and line_list[1] is None:
lrc_str += line_list[2][0][2]
Expand Down Expand Up @@ -291,11 +291,11 @@ def linelist2str(line_list: list[int | None, int | None, list[int | None, int |
return lrc_str


def find_closest_match(list1: list, list2: list, list3: list | None = None, source: Source | None = None) -> list:
list1: list[list[int | None, int | None, list[int | None, int | None, str]]] = list1[:]
list2: list[list[int | None, int | None, list[int | None, int | None, str]]] = list2[:]
def find_closest_match(list1: list, list2: list, list3: list | None = None, source: Source | None = None) -> list[tuple[list, list]]:
list1: list[tuple[int, int | None, list[tuple[int, int | None, str]]]] = list1[:]
list2: list[tuple[int, int | None, list[tuple[int, int | None, str]]]] = list2[:]
if list3:
list3: list[list[int | None, int | None, list[int | None, int | None, str]]] = list3[:]
list3: list[tuple[int, int | None, list[tuple[int, int | None, str]]]] = list3[:]
# 存储合并结果的列表
merged_dict = {}
merged_list = []
Expand All @@ -307,7 +307,7 @@ def find_closest_match(list1: list, list2: list, list3: list | None = None, sour
for match in matchs1:
list2_line_str = "".join([word[2]for word in match[1][2] if word[2] != ""])
if list1_line_str == list2_line_str or list1_line_str == list2_line_str.replace("\u3000", ""):
merged_list.append([list1_line, match[0]])
merged_list.append((list1_line, match[0]))
matchs1.remove(match)
break
if len(merged_list) == 0:
Expand All @@ -325,22 +325,18 @@ def find_closest_match(list1: list, list2: list, list3: list | None = None, sour
if len(list12) == len(list22):
logging.info("qm/kg 匹配方法")
for i, value in enumerate(list12):
merged_list.append([value, list22[i]])
merged_list.append((value, list22[i]))
return merged_list
list12, list22 = None, None

logging.info("other 匹配方法")

for i, value in enumerate(list1):
for index, word in enumerate(value[2]):
value[2][index] = tuple(word)

list1[i] = (value[0], value[1], tuple(value[2]))

for i, value in enumerate(list2):
for index, word in enumerate(value[2]):
value[2][index] = tuple(word)
list2[i] = (value[0], value[1], tuple(value[2]))

# 遍历第一个列表中的每个时间戳和歌词
i = 0
while len(list1) > i:
Expand Down Expand Up @@ -381,10 +377,16 @@ def find_closest_match(list1: list, list2: list, list3: list | None = None, sour

sorted_items = sorted(((key, value) for key, value in merged_dict.items()), key=lambda x: x[0])

return tuple_to_list([[item[1], item[0]] for item in sorted_items])
merged_list = []
for items in sorted_items:
item1 = (items[0][0], items[0][1], list(items[0][2]))
item2 = (items[1][0], items[1][1], list(items[1][2]))
merged_list.append((item2, item1))

return merged_list


class Lyrics(dict):
class Lyrics(dict[str: list[tuple[int | None, int | None, list[tuple[int | None, int | None, str]]]]]):
def __init__(self, info: dict | None = None) -> None:
if info is None:
info = {}
Expand Down Expand Up @@ -570,7 +572,7 @@ def get_full_line(mapping_table: dict, orig_linelist: list) -> str:

if not has_content(line_str):
return ""
if orig_linelist[0]:
if orig_linelist[0] is not None:
if re.search(end_time_pattern, line_str): # 检查是否有结束时间
return f"[{ms2formattime(orig_linelist[0])}]{line_str}"
if orig_linelist[1]:
Expand Down

0 comments on commit 25d23ca

Please sign in to comment.