feat: implement ffmpeg.rs, subtitles.rs, and fonts.rs

- ffmpeg.rs: discovery, duration extraction, metadata probing, download
- subtitles.rs: SRT-to-VTT conversion, sidecar discovery, storage, extraction
- fonts.rs: Google Fonts and Font Awesome local caching
This commit is contained in:
Your Name
2026-02-19 01:59:21 +02:00
parent 768ff1621f
commit 331020d936
4 changed files with 2075 additions and 0 deletions

614
src-tauri/src/fonts.rs Normal file
View File

@@ -0,0 +1,614 @@
use once_cell::sync::Lazy;
use regex::Regex;
use serde_json::json;
use sha2::{Digest, Sha256};
use std::collections::HashMap;
use std::fs;
use std::path::Path;
use std::time::SystemTime;
use crate::state::{atomic_write_json, load_json_with_fallbacks, BACKUP_COUNT};
// ---------------------------------------------------------------------------
// Constants
// ---------------------------------------------------------------------------
/// Current version for Google Fonts metadata (fonts_meta.json).
const GOOGLE_FONTS_META_VERSION: u64 = 7;
/// Current version for Font Awesome metadata (fa_meta.json).
const FA_META_VERSION: u64 = 3;
/// User-Agent header value for HTTP requests.
/// Google Fonts API returns different CSS based on User-Agent; we want woff2.
const USER_AGENT: &str = "Mozilla/5.0";
/// Google Fonts CSS URLs.
const GOOGLE_FONT_URLS: &[(&str, &str)] = &[
(
"Sora",
"https://fonts.googleapis.com/css2?family=Sora:wght@500;600;700;800&display=swap",
),
(
"Manrope",
"https://fonts.googleapis.com/css2?family=Manrope:wght@400;500;600;700;800&display=swap",
),
(
"IBM Plex Mono",
"https://fonts.googleapis.com/css2?family=IBM+Plex+Mono:wght@400;500;600&display=swap",
),
];
/// Font Awesome CSS URL.
const FA_CSS_URL: &str =
"https://cdnjs.cloudflare.com/ajax/libs/font-awesome/6.5.2/css/all.min.css";
/// Base URL for resolving relative Font Awesome webfont URLs.
const FA_WEBFONTS_BASE: &str =
"https://cdnjs.cloudflare.com/ajax/libs/font-awesome/6.5.2/webfonts/";
// ---------------------------------------------------------------------------
// Compiled regex patterns
// ---------------------------------------------------------------------------
/// Regex for extracting woff2 font URLs from Google Fonts CSS.
static GOOGLE_FONT_URL_RE: Lazy<Regex> =
Lazy::new(|| Regex::new(r#"url\(([^)]+)\)\s*format\(['"]woff2['"]\)"#).unwrap());
/// Regex for extracting all url(...) references from Font Awesome CSS.
static FA_URL_RE: Lazy<Regex> = Lazy::new(|| Regex::new(r"url\(([^)]+)\)").unwrap());
// ---------------------------------------------------------------------------
// 1. safe_filename_from_url
// ---------------------------------------------------------------------------
/// Generate a safe local filename from a URL using SHA-256 hash for uniqueness.
///
/// The filename is `{stem}-{hash}{suffix}` where `hash` is the first 10 hex
/// characters of the SHA-256 digest of the full URL. If the URL path has no
/// extension, `.woff2` is appended.
pub fn safe_filename_from_url(url: &str) -> String {
// Extract the last path component from the URL
let base = url
.split('?')
.next()
.unwrap_or(url)
.split('#')
.next()
.unwrap_or(url)
.rsplit('/')
.next()
.unwrap_or("font.woff2");
let base = if base.is_empty() { "font.woff2" } else { base };
// Ensure the base has an extension
let base = if !base.contains('.') {
format!("{}.woff2", base)
} else {
base.to_string()
};
// Split into stem and suffix
let (stem, suffix) = match base.rfind('.') {
Some(pos) => (&base[..pos], &base[pos..]),
None => (base.as_str(), ".woff2"),
};
// Compute SHA-256 hash of the full URL
let mut hasher = Sha256::new();
hasher.update(url.as_bytes());
let digest = format!("{:x}", hasher.finalize());
let url_hash = &digest[..10];
format!("{}-{}{}", stem, url_hash, suffix)
}
// ---------------------------------------------------------------------------
// 2. ensure_google_fonts_local
// ---------------------------------------------------------------------------
/// Download and cache Google Fonts (Sora, Manrope, IBM Plex Mono) locally.
///
/// The `fonts_dir` is the directory where `fonts.css`, `fonts_meta.json`, and
/// individual `.woff2` files are stored.
///
/// If already cached (version matches, ok=true, CSS file exists), this is a
/// no-op. Otherwise, downloads each font family's CSS from the Google Fonts
/// API, extracts woff2 URLs, downloads each font file, rewrites the CSS to
/// use local paths, and writes the combined CSS and metadata.
pub async fn ensure_google_fonts_local(fonts_dir: &Path) -> Result<(), String> {
fs::create_dir_all(fonts_dir).map_err(|e| format!("Failed to create fonts dir: {}", e))?;
let meta_path = fonts_dir.join("fonts_meta.json");
let css_path = fonts_dir.join("fonts.css");
// Check if already cached
if let Some(meta) = load_json_with_fallbacks(&meta_path, BACKUP_COUNT) {
if let Some(obj) = meta.as_object() {
let version_ok = obj
.get("version")
.and_then(|v| v.as_u64())
.map(|v| v == GOOGLE_FONTS_META_VERSION)
.unwrap_or(false);
let ok_flag = obj
.get("ok")
.and_then(|v| v.as_bool())
.unwrap_or(false);
if version_ok && ok_flag && css_path.exists() {
return Ok(());
}
}
}
let client = reqwest::Client::builder()
.user_agent(USER_AGENT)
.build()
.map_err(|e| format!("Failed to build HTTP client: {}", e))?;
let mut all_css_parts: Vec<String> = Vec::new();
let mut downloaded_files: Vec<String> = Vec::new();
let mut errors: Vec<String> = Vec::new();
for (family, css_url) in GOOGLE_FONT_URLS {
// Download the CSS for this font family
let css_text = match client.get(*css_url).send().await {
Ok(resp) => match resp.text().await {
Ok(text) => text,
Err(e) => {
errors.push(format!("Failed to read CSS for {}: {}", family, e));
continue;
}
},
Err(e) => {
errors.push(format!("Failed to download CSS for {}: {}", family, e));
continue;
}
};
// Find all woff2 url(...) references and download each font file
let mut rewritten_css = css_text.clone();
let mut replacements: Vec<(String, String)> = Vec::new();
for cap in GOOGLE_FONT_URL_RE.captures_iter(&css_text) {
let raw_url = cap[1].trim().trim_matches('\'').trim_matches('"');
let safe_name = safe_filename_from_url(raw_url);
let local_path = fonts_dir.join(&safe_name);
// Download the font file
match client.get(raw_url).send().await {
Ok(resp) => match resp.bytes().await {
Ok(bytes) => {
if let Err(e) = fs::write(&local_path, &bytes) {
errors.push(format!("Failed to write {}: {}", safe_name, e));
continue;
}
downloaded_files.push(safe_name.clone());
}
Err(e) => {
errors.push(format!("Failed to read bytes for {}: {}", safe_name, e));
continue;
}
},
Err(e) => {
errors.push(format!("Failed to download {}: {}", raw_url, e));
continue;
}
}
// Record the replacement: original url(...) content -> local path
let replacement_url = format!("/fonts/{}", safe_name);
replacements.push((cap[1].to_string(), replacement_url));
}
// Apply all URL replacements to the CSS
for (original, replacement) in &replacements {
let old = format!("url({}) format", original);
let new = format!("url({}) format", replacement);
rewritten_css = rewritten_css.replace(&old, &new);
}
all_css_parts.push(rewritten_css);
}
// Write combined CSS
let combined_css = all_css_parts.join("\n");
fs::write(&css_path, &combined_css)
.map_err(|e| format!("Failed to write fonts.css: {}", e))?;
// Write metadata
let timestamp = SystemTime::now()
.duration_since(SystemTime::UNIX_EPOCH)
.map(|d| d.as_secs())
.unwrap_or(0);
let ok = errors.is_empty();
let meta = json!({
"version": GOOGLE_FONTS_META_VERSION,
"ok": ok,
"timestamp": timestamp,
"downloaded": downloaded_files,
"errors": errors,
});
atomic_write_json(&meta_path, &meta, BACKUP_COUNT);
if ok {
Ok(())
} else {
Err(format!(
"Google Fonts download completed with errors: {}",
errors.join("; ")
))
}
}
// ---------------------------------------------------------------------------
// 3. ensure_fontawesome_local
// ---------------------------------------------------------------------------
/// Clean and resolve a Font Awesome URL reference.
///
/// Strips whitespace and quotes, resolves relative URLs against the FA
/// webfonts base URL. Returns the URL unchanged if it is a `data:` URI.
fn clean_fa_url(u: &str) -> String {
let u = u.trim().trim_matches('\'').trim_matches('"');
if u.starts_with("data:") {
return u.to_string();
}
if u.starts_with("//") {
return format!("https:{}", u);
}
if u.starts_with("http://") || u.starts_with("https://") {
return u.to_string();
}
// Relative URL: strip leading "./" and "../" then join with base
let cleaned = u
.trim_start_matches("./")
.replace("../", "");
format!("{}{}", FA_WEBFONTS_BASE, cleaned)
}
/// Download and cache Font Awesome 6.5.2 locally.
///
/// The `fa_dir` is the directory where `fa.css` and `fa_meta.json` live.
/// The `fa_dir/webfonts/` subdirectory holds individual webfont files.
///
/// If already cached (version matches, ok=true, CSS file exists), this is a
/// no-op. Otherwise, downloads the Font Awesome CSS, extracts all `url(...)`
/// references, downloads each font file (skipping `data:` URIs), rewrites
/// the CSS to use local paths, and writes the CSS and metadata.
pub async fn ensure_fontawesome_local(fa_dir: &Path) -> Result<(), String> {
fs::create_dir_all(fa_dir).map_err(|e| format!("Failed to create fa dir: {}", e))?;
let webfonts_dir = fa_dir.join("webfonts");
fs::create_dir_all(&webfonts_dir)
.map_err(|e| format!("Failed to create webfonts dir: {}", e))?;
let meta_path = fa_dir.join("fa_meta.json");
let css_path = fa_dir.join("fa.css");
// Check if already cached
if let Some(meta) = load_json_with_fallbacks(&meta_path, BACKUP_COUNT) {
if let Some(obj) = meta.as_object() {
let version_ok = obj
.get("version")
.and_then(|v| v.as_u64())
.map(|v| v == FA_META_VERSION)
.unwrap_or(false);
let ok_flag = obj
.get("ok")
.and_then(|v| v.as_bool())
.unwrap_or(false);
if version_ok && ok_flag && css_path.exists() {
return Ok(());
}
}
}
let client = reqwest::Client::builder()
.user_agent(USER_AGENT)
.build()
.map_err(|e| format!("Failed to build HTTP client: {}", e))?;
// Download the Font Awesome CSS
let css_text = client
.get(FA_CSS_URL)
.send()
.await
.map_err(|e| format!("Failed to download FA CSS: {}", e))?
.text()
.await
.map_err(|e| format!("Failed to read FA CSS: {}", e))?;
let mut downloaded_files: Vec<String> = Vec::new();
let mut errors: Vec<String> = Vec::new();
let mut replacements: HashMap<String, String> = HashMap::new();
for cap in FA_URL_RE.captures_iter(&css_text) {
let raw_url = &cap[1];
let resolved = clean_fa_url(raw_url);
// Skip data: URIs
if resolved.starts_with("data:") {
continue;
}
// Determine the filename from the resolved URL
let filename = resolved
.split('?')
.next()
.unwrap_or(&resolved)
.split('#')
.next()
.unwrap_or(&resolved)
.rsplit('/')
.next()
.unwrap_or("font.woff2")
.to_string();
if filename.is_empty() {
continue;
}
let local_path = webfonts_dir.join(&filename);
// Only download each file once
if !replacements.contains_key(raw_url) {
match client.get(&resolved).send().await {
Ok(resp) => match resp.bytes().await {
Ok(bytes) => {
if let Err(e) = fs::write(&local_path, &bytes) {
errors.push(format!("Failed to write {}: {}", filename, e));
continue;
}
downloaded_files.push(filename.clone());
}
Err(e) => {
errors.push(format!("Failed to read bytes for {}: {}", filename, e));
continue;
}
},
Err(e) => {
errors.push(format!("Failed to download {}: {}", resolved, e));
continue;
}
}
let replacement = format!("/fa/webfonts/{}", filename);
replacements.insert(raw_url.to_string(), replacement);
}
}
// Rewrite CSS with local paths
let mut rewritten_css = css_text.clone();
for (original, replacement) in &replacements {
let old = format!("url({})", original);
let new = format!("url({})", replacement);
rewritten_css = rewritten_css.replace(&old, &new);
}
// Write rewritten CSS
fs::write(&css_path, &rewritten_css)
.map_err(|e| format!("Failed to write fa.css: {}", e))?;
// Write metadata
let timestamp = SystemTime::now()
.duration_since(SystemTime::UNIX_EPOCH)
.map(|d| d.as_secs())
.unwrap_or(0);
let ok = errors.is_empty();
let meta = json!({
"version": FA_META_VERSION,
"ok": ok,
"timestamp": timestamp,
"downloaded": downloaded_files,
"errors": errors,
});
atomic_write_json(&meta_path, &meta, BACKUP_COUNT);
if ok {
Ok(())
} else {
Err(format!(
"Font Awesome download completed with errors: {}",
errors.join("; ")
))
}
}
// ===========================================================================
// Tests
// ===========================================================================
#[cfg(test)]
mod tests {
use super::*;
// -- safe_filename_from_url -----------------------------------------------
#[test]
fn test_safe_filename_from_url_basic() {
let url = "https://fonts.gstatic.com/s/sora/v12/abc123.woff2";
let result = safe_filename_from_url(url);
// Should contain the original stem
assert!(result.starts_with("abc123-"));
// Should end with .woff2
assert!(result.ends_with(".woff2"));
// Should contain a 10-char hash between stem and extension
let parts: Vec<&str> = result.rsplitn(2, '.').collect();
let before_ext = parts[1]; // "abc123-{hash}"
let hash_part = before_ext.rsplit('-').next().unwrap();
assert_eq!(hash_part.len(), 10);
}
#[test]
fn test_safe_filename_from_url_no_extension() {
let url = "https://example.com/fontfile";
let result = safe_filename_from_url(url);
// Should have .woff2 appended
assert!(result.ends_with(".woff2"));
assert!(result.starts_with("fontfile-"));
}
#[test]
fn test_safe_filename_from_url_deterministic() {
let url = "https://fonts.gstatic.com/s/sora/v12/abc.woff2";
let result1 = safe_filename_from_url(url);
let result2 = safe_filename_from_url(url);
assert_eq!(result1, result2);
}
#[test]
fn test_safe_filename_different_urls() {
let url1 = "https://fonts.gstatic.com/s/sora/v12/abc.woff2";
let url2 = "https://fonts.gstatic.com/s/manrope/v14/def.woff2";
let result1 = safe_filename_from_url(url1);
let result2 = safe_filename_from_url(url2);
assert_ne!(result1, result2);
}
// -- clean_fa_url ---------------------------------------------------------
#[test]
fn test_clean_fa_url_data() {
let result = clean_fa_url("data:font/woff2;base64,abc");
assert_eq!(result, "data:font/woff2;base64,abc");
}
#[test]
fn test_clean_fa_url_protocol_relative() {
let result = clean_fa_url("//example.com/font.woff2");
assert_eq!(result, "https://example.com/font.woff2");
}
#[test]
fn test_clean_fa_url_absolute() {
let result = clean_fa_url("https://example.com/font.woff2");
assert_eq!(result, "https://example.com/font.woff2");
}
#[test]
fn test_clean_fa_url_relative() {
let result = clean_fa_url("../webfonts/fa-solid-900.woff2");
assert_eq!(
result,
"https://cdnjs.cloudflare.com/ajax/libs/font-awesome/6.5.2/webfonts/webfonts/fa-solid-900.woff2"
);
}
#[test]
fn test_clean_fa_url_relative_dot_slash() {
let result = clean_fa_url("./webfonts/fa-solid-900.woff2");
assert_eq!(
result,
"https://cdnjs.cloudflare.com/ajax/libs/font-awesome/6.5.2/webfonts/webfonts/fa-solid-900.woff2"
);
}
#[test]
fn test_clean_fa_url_strips_quotes() {
let result = clean_fa_url("'https://example.com/font.woff2'");
assert_eq!(result, "https://example.com/font.woff2");
}
// -- Integration tests (require network) ----------------------------------
#[tokio::test]
#[ignore]
async fn test_google_fonts_download() {
let dir = tempfile::tempdir().unwrap();
let fonts_dir = dir.path().join("fonts");
let result = ensure_google_fonts_local(&fonts_dir).await;
assert!(result.is_ok(), "Google Fonts download failed: {:?}", result);
// Verify fonts.css was created
let css_path = fonts_dir.join("fonts.css");
assert!(css_path.exists(), "fonts.css should exist");
let css_content = fs::read_to_string(&css_path).unwrap();
assert!(!css_content.is_empty(), "fonts.css should not be empty");
// CSS should contain rewritten local paths
assert!(
css_content.contains("/fonts/"),
"CSS should contain /fonts/ local paths"
);
// Verify metadata was created
let meta_path = fonts_dir.join("fonts_meta.json");
assert!(meta_path.exists(), "fonts_meta.json should exist");
let meta = load_json_with_fallbacks(&meta_path, BACKUP_COUNT).unwrap();
assert_eq!(meta["version"], GOOGLE_FONTS_META_VERSION);
assert_eq!(meta["ok"], true);
assert!(
meta["downloaded"].as_array().unwrap().len() > 0,
"Should have downloaded at least one font file"
);
// Second call should be a no-op (cached)
let result2 = ensure_google_fonts_local(&fonts_dir).await;
assert!(result2.is_ok());
}
#[tokio::test]
#[ignore]
async fn test_fontawesome_download() {
let dir = tempfile::tempdir().unwrap();
let fa_dir = dir.path().join("fa");
let result = ensure_fontawesome_local(&fa_dir).await;
assert!(
result.is_ok(),
"Font Awesome download failed: {:?}",
result
);
// Verify fa.css was created
let css_path = fa_dir.join("fa.css");
assert!(css_path.exists(), "fa.css should exist");
let css_content = fs::read_to_string(&css_path).unwrap();
assert!(!css_content.is_empty(), "fa.css should not be empty");
// CSS should contain rewritten local paths
assert!(
css_content.contains("/fa/webfonts/"),
"CSS should contain /fa/webfonts/ local paths"
);
// Verify webfonts directory has files
let webfonts_dir = fa_dir.join("webfonts");
assert!(webfonts_dir.exists(), "webfonts dir should exist");
let webfont_files: Vec<_> = fs::read_dir(&webfonts_dir)
.unwrap()
.filter_map(|e| e.ok())
.collect();
assert!(
!webfont_files.is_empty(),
"Should have downloaded at least one webfont file"
);
// Verify metadata was created
let meta_path = fa_dir.join("fa_meta.json");
assert!(meta_path.exists(), "fa_meta.json should exist");
let meta = load_json_with_fallbacks(&meta_path, BACKUP_COUNT).unwrap();
assert_eq!(meta["version"], FA_META_VERSION);
assert_eq!(meta["ok"], true);
// Second call should be a no-op (cached)
let result2 = ensure_fontawesome_local(&fa_dir).await;
assert!(result2.is_ok());
}
}