A simple LRC parser for Python.
Project description
Lemony LRC Parser
柠檬味的 Python LRC 歌词解析器.
Lemon-flavored LRC Parser for Python.
Features
- 解析标准 LRC 歌词文件
- 支持 EnhancedLRC / SPL 的逐字歌词标签
- 支持 metadata 标签
- 支持折叠时间标签
- 支持参照行
- 支持歌词合并
- 与简单字幕格式 (SRT / WebVTT) 互转
- 时间偏移 (
apply_delta/<</>>运算符) - 字典序列化 (
to_dict()/from_dict()) - 深拷贝方法链 (
.copy()) - 可配置解析与序列化选项
- 完整的类型注解
Installation
推荐使用 uv.
It's recommended to use uv.
uv add lemony-lrc-parser
用 pip 也行.
It's okay to use pip.
pip install lemony-lrc-parser
可以使用 git 仓库来源来第一时间体验到最新最热的 bug feature.
You can use git repo as a source to catch the newest bugs features.
uv add https://github.com/NingmengLemon/lemony-lrc-parser.git
Usage
Quick Start
json, marshal or pickle -like usages
import lemony_lrc_parser as llp
lrc_text = """[ti: Never Gonna Give You Up]
[ar: Rick Astley]
[00:18.684]We're no strangers to love
[00:18.684]我们都是情场老手
[00:22.657]You know the rules and so do I
[00:22.657]你和我都知道爱情的规则
[00:27.070]A full commitment's what I'm thinking of
[00:27.070]我在想的正是一份实打实的承诺
[00:31.459]You wouldn't get this from any other guy
[00:31.459]你从其他人那里得不到的
"""
# 解析
lyrics = llp.loads(lrc_text)
# 访问 metadata
print(lyrics.metadata["ti"]) # "Never Gonna Give You Up"
print(lyrics.metadata["ar"]) # "Rick Astley"
# 遍历歌词行
for line in lyrics:
print(f"{line.start}ms: {line.text}")
# 序列化回 LRC 格式
output = llp.dumps(lyrics)
OOP Interface
Lyrics 类提供面向对象的解析和序列化入口, 也推荐使用面向对象接口:
from lemony_lrc_parser import Lyrics
lyrics = Lyrics.loads(lrc_text)
# Lyrics 同时是序列容器
print(len(lyrics)) # 行数
print(lyrics[0].text) # 第一行文本
print(lyrics[-1].text)
# 切片访问
first_three = lyrics[0:3]
# 序列化
lrc_output = lyrics.dumps()
# __str__ 等价于 dumps()
print(lyrics)
Word-level Lyrics
解析逐字 (Enhanced LRC / SPL) 歌词:
lrc_text = "[00:01.00]<00:01.00>Never <00:01.50>gonna <00:02.00>give <00:02.50>you <00:03.00>up[00:03.50]"
lyrics = llp.loads(lrc_text)
line = lyrics[0]
for word in line.content:
print(f" [{word.start} -> {word.end}] {word.content!r}")
# [1000 -> 1500] 'Never '
# [1500 -> 2000] 'gonna '
# [2000 -> 2500] 'give '
# [2500 -> 3000] 'you '
# [3000 -> None] 'up'
# 行级时间: line.start=1000, line.end=3500
Reference Lines (Translation / Transliteration)
LRC 文件中, 紧跟在带时间标签行后面的无标签行, 或与主行的时间戳相同的行, 会被解析为参考行, 常用于存放翻译或音译:
lrc_text = """[00:01.00]Hello
你好
[00:02.00]World
[00:02.00]世界
"""
lyrics = llp.loads(lrc_text)
line = lyrics[0]
print(line.text) # "Hello"
print(line.reference_lines[0][0].content) # "你好"
Combining Lyrics
将两份歌词 (如原文和翻译) 按时间标签合并:
main = llp.loads("[00:01.00]Hello\n[00:02.00]World\n")
translation = llp.loads("[00:01.00]你好\n[00:02.00]世界\n")
# combine 方法: 翻译行挂到同时间点的 reference_lines 中
combined = main.combine(translation)
# 也可以用 + 运算符
combined = main + translation
for line in combined:
print(line.text) # 主歌词
for ref in line.reference_lines:
ref_text = "".join(w.content for w in ref)
print(f" -> {ref_text}") # 参考行
# other_as_refline_only=False 时, 翻译中找不到对应时间点的行会作为新行保留
combined = main.combine(translation, other_as_refline_only=False)
Dict Serialization (to_dict / from_dict)
所有数据模型都支持字典序列化, 方便 JSON 传输和 API 对接:
import lemony_lrc_parser as llp
lyrics = llp.loads("[00:01.00]Hello\n[00:02.00]World\n")
# Lyrics → dict
data = lyrics.to_dict()
# {"metadata": {}, "lines": [{"start": 1000, ...}, ...]}
# dict → Lyrics
restored = Lyrics.from_dict(data)
# 也可以对单行、单行内容、单个词元分别操作
line = lyrics[0]
line_dict = line.to_dict()
token_dict = line.content[0].to_dict()
Copy
所有数据模型都提供 .copy() 深拷贝方法, 返回独立的副本:
import lemony_lrc_parser as llp
lyrics = llp.loads("[00:01.00]Hello\n")
# 深拷贝
clone = lyrics.copy()
clone.metadata["ti"] = "New Title"
# 原始对象不受影响
print(lyrics.metadata.get("ti")) # None
Options
Parsing Options
import re
from lemony_lrc_parser import Lyrics
from lemony_lrc_parser.models import ParseOptions
lrc_text = "[00:01.000]Hello\n[00:05.000]World\n"
lyrics = Lyrics.loads(
lrc_text,
options=ParseOptions(
fill_implicit_line_end=True, # 是否填充隐式行尾时间
line_filter=r"纯音乐.*?请欣赏", # 黑名单过滤 (统一按正则理解, str 会被自动 compile)
# line_filter=re.compile(r"纯音乐.*?请欣赏"), # 也可直接传入已编译的正则
),
)
# lyrics[0].end == lyrics[1].start == 5000
line_filter 的内容是正则表达式: 传入的字符串会被 re.compile 编译, 之后用
pattern.search 匹配每行文本, 命中的行会被丢弃. 若需要精确的子串匹配 (而非正则),
请用 re.escape(...) 包一层, 例如 line_filter=re.escape("a.c").
Serialization Options
通过 SerializationOptions 控制序列化行为:
from lemony_lrc_parser import Lyrics
from lemony_lrc_parser.models import SerializationOptions
output = lyrics.dumps(
options=SerializationOptions(
with_metadata=True, # 是否输出 metadata 段
use_bracket_for_byword_tag=False, # 逐字标签使用 [...] 还是 <...> (默认)
line_tag_decimal_length=3, # 行标签毫秒位数 (默认 3)
word_tag_decimal_length=3, # 逐字标签毫秒位数 (默认 3)
line_separator="\n", # 行间分隔字符串 (默认 "\n", 设为 "" 可省去空行)
),
)
Length of Decimal Part
默认 line_tag_decimal_length=3、word_tag_decimal_length=3, 输出格式如 [00:01.000]、<00:01.050>, 保留完整的毫秒精度.
若设为 2, 小数部分表示百分秒 (如 [00:01.00]), 属于有损截断 (例如 555ms 会被截断为 55, 解析回来变成 550ms), 需要按需权衡 (部分老软件可能只支持百分秒).
Offset
使用 Lyrics.apply_delta(ms) 应用时间偏移, ms 会直接加到每个标签的时间戳上,
这意味着传入正数偏移值会导致歌词整体延后出现, 反之同理.
也可以使用重载的 >> / << 运算, 私以为这样会更好理解一些.
如果应用 offset 会导致时间戳变为负数, 将抛出 TimestampUnderflowError, 由调用方自行处理.
Applying Offset
通过 Lyrics.apply_delta(ms) 对时间戳应用偏移, 返回一个新对象:
from lemony_lrc_parser import Lyrics
lyrics = Lyrics.loads(lrc_text)
# 正数 → 歌词延后出现 (等价于 lyrics >> 500)
shifted = lyrics.apply_delta(500)
# 负数 → 歌词提前出现 (等价于 lyrics << 500)
shifted = lyrics.apply_delta(-500)
# 使用 << / >> 运算符
shifted = lyrics >> 500 # 延后 500ms
shifted = lyrics << 500 # 提前 500ms
# shifted 的时间戳已被整体偏移, 原始 lyrics 不受影响
如果偏移会导致时间戳变为负数, 将抛出 TimestampUnderflowError.
如需在序列化前偏移时间戳, 请先调用 apply_delta() 再序列化返回的副本.
如果你的偏移量来自歌词文件元数据, 你可能还需要记得手动清理 lyrics.metadata 中的偏移值.
Subtitle Conversion (SRT / WebVTT)
Lyrics 可与常见的简单字幕格式互相转换, 便于把歌词用于视频字幕制作,
或把已有字幕导入为歌词:
from lemony_lrc_parser import Lyrics
from lemony_lrc_parser.models import SubtitleOptions
lyrics = Lyrics.loads("[00:01.000]Hello\n[00:03.000]World\n")
# LRC → SRT / WebVTT
srt_text = lyrics.to_srt()
vtt_text = lyrics.to_webvtt()
# SRT / WebVTT → LRC
lyrics2 = Lyrics.from_srt(srt_text)
lyrics3 = Lyrics.from_webvtt(vtt_text)
# 顶层便捷函数也可用
import lemony_lrc_parser as llp
srt_text = llp.dump_srt(lyrics)
vtt_text = llp.dump_webvtt(lyrics)
lyrics2 = llp.parse_srt(srt_text)
lyrics3 = llp.parse_webvtt(vtt_text)
字幕以 [start, end] 时间区间为单位, 与 LRC 存在语义差异, 转换时的行为可通过
SubtitleOptions 控制:
options = SubtitleOptions(
fill_end_from_next=True, # 缺少行尾时间时, 用下一行的 start 补齐
default_duration_ms=5000, # 无法推断时长时使用的默认时长 (也用于修正非法区间)
include_reference_lines=True, # 是否把参考行 (翻译/音译) 作为 cue 附加文本输出
)
srt_text = lyrics.to_srt(options=options)
转换注意事项:
- LRC 的逐字标签与 metadata 不会写入字幕 (会被拍平/丢弃).
- 导出时若某行缺少
end, 会依次尝试用下一行start或default_duration_ms补齐;end <= start的非法区间也会被修正. - 一条字幕 cue 可含多行文本: 导出时主行在前、参考行在后; 解析时 cue 首行作为主行, 其余行作为参考行.
- 解析会自动跳过 WebVTT 的
WEBVTT头部以及NOTE/STYLE/REGION块.
References
UwU?
UwU!
License
MIT License
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
Filter files by name, interpreter, ABI, and platform.
If you're not sure about the file name format, learn more about wheel file names.
Copy a direct link to the current filters
File details
Details for the file lemony_lrc_parser-0.4.0b2.tar.gz.
File metadata
- Download URL: lemony_lrc_parser-0.4.0b2.tar.gz
- Upload date:
- Size: 134.3 kB
- Tags: Source
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/6.2.0 CPython/3.14.3
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
e2db8a8f4a22f86cf51a4a375367dc22cb3c480fe9e6340a4c03e9365dc78859
|
|
| MD5 |
33271502ba80566c163839cca5a1071d
|
|
| BLAKE2b-256 |
073f44eb2723b99f6d72ce9e0f0141df2084436eeffafc335c2e52f9374ff8c5
|
File details
Details for the file lemony_lrc_parser-0.4.0b2-py3-none-any.whl.
File metadata
- Download URL: lemony_lrc_parser-0.4.0b2-py3-none-any.whl
- Upload date:
- Size: 29.5 kB
- Tags: Python 3
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/6.2.0 CPython/3.14.3
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
67bc3749d582deb2ce874135f111dd3ca3d8f5bfae7192bbb4b5b7dc5885386c
|
|
| MD5 |
54a4f609cb26921c4827070af4626092
|
|
| BLAKE2b-256 |
028d1f9bb0cbbf1e6d280350b280cf54762219bd4d27bc442e0a61d0440a7da5
|