mirror of
https://github.com/crunchy-labs/crunchy-cli.git
synced 2026-01-21 12:12:00 -06:00
Add options to specify audio & subtitle locales as IETF language tag and add --language_tagging flag for archive and download to modify the output file language tagging (#330)
This commit is contained in:
bytedream
parent
3f33db6728
commit
f1d266c940
5 changed files with 260 additions and 15 deletions
|
|
@ -12,7 +12,7 @@ use regex::Regex;
|
|||
use reqwest::Client;
|
||||
use std::borrow::Borrow;
|
||||
use std::cmp::Ordering;
|
||||
use std::collections::BTreeMap;
|
||||
use std::collections::{BTreeMap, HashMap};
|
||||
use std::io::Write;
|
||||
use std::path::{Path, PathBuf};
|
||||
use std::process::{Command, Stdio};
|
||||
|
|
@ -61,6 +61,8 @@ pub struct DownloadBuilder {
|
|||
no_closed_caption: bool,
|
||||
threads: usize,
|
||||
ffmpeg_threads: Option<usize>,
|
||||
audio_locale_output_map: HashMap<Locale, String>,
|
||||
subtitle_locale_output_map: HashMap<Locale, String>,
|
||||
}
|
||||
|
||||
impl DownloadBuilder {
|
||||
|
|
@ -78,6 +80,8 @@ impl DownloadBuilder {
|
|||
no_closed_caption: false,
|
||||
threads: num_cpus::get(),
|
||||
ffmpeg_threads: None,
|
||||
audio_locale_output_map: HashMap::new(),
|
||||
subtitle_locale_output_map: HashMap::new(),
|
||||
}
|
||||
}
|
||||
|
||||
|
|
@ -99,6 +103,9 @@ impl DownloadBuilder {
|
|||
ffmpeg_threads: self.ffmpeg_threads,
|
||||
|
||||
formats: vec![],
|
||||
|
||||
audio_locale_output_map: self.audio_locale_output_map,
|
||||
subtitle_locale_output_map: self.subtitle_locale_output_map,
|
||||
}
|
||||
}
|
||||
}
|
||||
|
|
@ -138,6 +145,9 @@ pub struct Downloader {
|
|||
ffmpeg_threads: Option<usize>,
|
||||
|
||||
formats: Vec<DownloadFormat>,
|
||||
|
||||
audio_locale_output_map: HashMap<Locale, String>,
|
||||
subtitle_locale_output_map: HashMap<Locale, String>,
|
||||
}
|
||||
|
||||
impl Downloader {
|
||||
|
|
@ -426,7 +436,12 @@ impl Downloader {
|
|||
maps.extend(["-map".to_string(), (i + videos.len()).to_string()]);
|
||||
metadata.extend([
|
||||
format!("-metadata:s:a:{}", i),
|
||||
format!("language={}", meta.language),
|
||||
format!(
|
||||
"language={}",
|
||||
self.audio_locale_output_map
|
||||
.get(&meta.language)
|
||||
.unwrap_or(&meta.language.to_string())
|
||||
),
|
||||
]);
|
||||
metadata.extend([
|
||||
format!("-metadata:s:a:{}", i),
|
||||
|
|
@ -457,7 +472,12 @@ impl Downloader {
|
|||
]);
|
||||
metadata.extend([
|
||||
format!("-metadata:s:s:{}", i),
|
||||
format!("language={}", meta.language),
|
||||
format!(
|
||||
"language={}",
|
||||
self.subtitle_locale_output_map
|
||||
.get(&meta.language)
|
||||
.unwrap_or(&meta.language.to_string())
|
||||
),
|
||||
]);
|
||||
metadata.extend([
|
||||
format!("-metadata:s:s:{}", i),
|
||||
|
|
|
|||
|
|
@ -1,4 +1,5 @@
|
|||
use crate::utils::filter::real_dedup_vec;
|
||||
use crate::utils::locale::LanguageTagging;
|
||||
use crate::utils::log::tab_info;
|
||||
use crate::utils::os::{is_special_file, sanitize};
|
||||
use anyhow::Result;
|
||||
|
|
@ -417,7 +418,12 @@ impl Format {
|
|||
}
|
||||
|
||||
/// Formats the given string if it has specific pattern in it. It also sanitizes the filename.
|
||||
pub fn format_path(&self, path: PathBuf, universal: bool) -> PathBuf {
|
||||
pub fn format_path(
|
||||
&self,
|
||||
path: PathBuf,
|
||||
universal: bool,
|
||||
language_tagging: Option<&LanguageTagging>,
|
||||
) -> PathBuf {
|
||||
let path = path
|
||||
.to_string_lossy()
|
||||
.to_string()
|
||||
|
|
@ -427,7 +433,7 @@ impl Format {
|
|||
&sanitize(
|
||||
self.locales
|
||||
.iter()
|
||||
.map(|(a, _)| a.to_string())
|
||||
.map(|(a, _)| language_tagging.map_or(a.to_string(), |t| t.for_locale(a)))
|
||||
.collect::<Vec<String>>()
|
||||
.join(
|
||||
&env::var("CRUNCHY_CLI_FORMAT_DELIMITER")
|
||||
|
|
|
|||
|
|
@ -1,4 +1,124 @@
|
|||
use crunchyroll_rs::Locale;
|
||||
use log::warn;
|
||||
|
||||
#[derive(Clone, Debug)]
|
||||
#[allow(clippy::upper_case_acronyms)]
|
||||
pub enum LanguageTagging {
|
||||
Default,
|
||||
IETF,
|
||||
}
|
||||
|
||||
impl LanguageTagging {
|
||||
pub fn parse(s: &str) -> Result<Self, String> {
|
||||
Ok(match s.to_lowercase().as_str() {
|
||||
"default" => Self::Default,
|
||||
"ietf" => Self::IETF,
|
||||
_ => return Err(format!("'{}' is not a valid language tagging", s)),
|
||||
})
|
||||
}
|
||||
|
||||
pub fn convert_locales(&self, locales: &[Locale]) -> Vec<String> {
|
||||
let ietf_language_codes = ietf_language_codes();
|
||||
let mut converted = vec![];
|
||||
|
||||
match &self {
|
||||
LanguageTagging::Default => {
|
||||
for locale in locales {
|
||||
let Some((_, available)) =
|
||||
ietf_language_codes.iter().find(|(_, l)| l.contains(locale))
|
||||
else {
|
||||
// if no matching IETF language code was found, just pass it as it is
|
||||
converted.push(locale.to_string());
|
||||
continue;
|
||||
};
|
||||
converted.push(available.first().unwrap().to_string())
|
||||
}
|
||||
}
|
||||
LanguageTagging::IETF => {
|
||||
for locale in locales {
|
||||
let Some((tag, _)) =
|
||||
ietf_language_codes.iter().find(|(_, l)| l.contains(locale))
|
||||
else {
|
||||
// if no matching IETF language code was found, just pass it as it is
|
||||
converted.push(locale.to_string());
|
||||
continue;
|
||||
};
|
||||
converted.push(tag.to_string())
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
converted
|
||||
}
|
||||
|
||||
pub fn for_locale(&self, locale: &Locale) -> String {
|
||||
match &self {
|
||||
LanguageTagging::Default => ietf_language_codes()
|
||||
.iter()
|
||||
.find(|(_, l)| l.contains(locale))
|
||||
.map_or(locale.to_string(), |(_, l)| l[0].to_string()),
|
||||
LanguageTagging::IETF => ietf_language_codes()
|
||||
.iter()
|
||||
.find(|(_, l)| l.contains(locale))
|
||||
.map_or(locale.to_string(), |(tag, _)| tag.to_string()),
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
pub fn resolve_locales(locales: &[Locale]) -> Vec<Locale> {
|
||||
let ietf_language_codes = ietf_language_codes();
|
||||
let all_locales = Locale::all();
|
||||
|
||||
let mut resolved = vec![];
|
||||
for locale in locales {
|
||||
if all_locales.contains(locale) {
|
||||
resolved.push(locale.clone())
|
||||
} else if let Some((_, resolved_locales)) = ietf_language_codes
|
||||
.iter()
|
||||
.find(|(tag, _)| tag == &locale.to_string().as_str())
|
||||
{
|
||||
let (first, alternatives) = resolved_locales.split_first().unwrap();
|
||||
|
||||
resolved.push(first.clone());
|
||||
// ignoring `Locale::en_IN` because I think the majority of users which want english
|
||||
// audio / subs want the "actual" english version and not the hindi accent dub
|
||||
if !alternatives.is_empty() && resolved_locales.first().unwrap() != &Locale::en_IN {
|
||||
warn!("Resolving locale '{}' to '{}', but there are some alternatives: {}. If you an alternative instead, please write it completely out instead of '{}'", locale, first, alternatives.iter().map(|l| format!("'{l}'")).collect::<Vec<String>>().join(", "), locale)
|
||||
}
|
||||
} else {
|
||||
resolved.push(locale.clone());
|
||||
warn!("Unknown locale '{}'", locale)
|
||||
}
|
||||
}
|
||||
|
||||
resolved
|
||||
}
|
||||
|
||||
fn ietf_language_codes<'a>() -> Vec<(&'a str, Vec<Locale>)> {
|
||||
vec![
|
||||
("ar", vec![Locale::ar_ME, Locale::ar_SA]),
|
||||
("ca", vec![Locale::ca_ES]),
|
||||
("de", vec![Locale::de_DE]),
|
||||
("en", vec![Locale::en_US, Locale::hi_IN]),
|
||||
("es", vec![Locale::es_ES, Locale::es_419, Locale::es_LA]),
|
||||
("fr", vec![Locale::fr_FR]),
|
||||
("hi", vec![Locale::hi_IN]),
|
||||
("id", vec![Locale::id_ID]),
|
||||
("it", vec![Locale::it_IT]),
|
||||
("ja", vec![Locale::ja_JP]),
|
||||
("ko", vec![Locale::ko_KR]),
|
||||
("ms", vec![Locale::ms_MY]),
|
||||
("pl", vec![Locale::pl_PL]),
|
||||
("pt", vec![Locale::pt_PT, Locale::pt_BR]),
|
||||
("ru", vec![Locale::ru_RU]),
|
||||
("ta", vec![Locale::ta_IN]),
|
||||
("te", vec![Locale::te_IN]),
|
||||
("th", vec![Locale::th_TH]),
|
||||
("tr", vec![Locale::tr_TR]),
|
||||
("vi", vec![Locale::vi_VN]),
|
||||
("zh", vec![Locale::zh_CN, Locale::zh_HK, Locale::zh_TW]),
|
||||
]
|
||||
}
|
||||
|
||||
/// Return the locale of the system.
|
||||
pub fn system_locale() -> Locale {
|
||||
|
|
|
|||
Loading…
Add table
Add a link
Reference in a new issue