MystiquePlus/.forgejo/scripts/cultural-groups.js
2025-03-01 20:51:53 -03:00

214 lines
No EOL
7.1 KiB
JavaScript

const fs = require('fs');
const path = require('path');
// Define cultural groups by country with more precise matching
const culturalGroups = {
anglosphere: [
'united kingdom', 'uk', 'great britain', 'britain',
'england', 'scotland', 'wales', 'northern ireland',
'^united states$', '^usa$', '^us$', '^america$',
'^canada$',
'^australia$',
'^new zealand$',
'^ireland$'
],
francophone: [
'^france$', '^belgium$', '^switzerland$', '^quebec$', '^monaco$',
'^luxembourg$', '^haiti$', '^ivory coast$', '^senegal$', '^cameroon$'
],
hispanic: [
'^spain$', '^mexico$', '^argentina$', '^chile$', '^colombia$', '^peru$',
'^venezuela$', '^ecuador$', '^guatemala$', '^cuba$', '^dominican republic$',
'^honduras$', '^el salvador$', '^nicaragua$', '^costa rica$', '^panama$',
'^bolivia$', '^paraguay$', '^uruguay$', 'latin america'
],
lusophone: [
'^portugal$', '^brazil$', '^angola$', '^mozambique$',
'^cape verde$', '^guinea-bissau$', '^sao tome and principe$'
],
arabic: [
'saudi arabia', 'egypt', 'uae', 'united arab emirates', 'qatar',
'kuwait', 'oman', 'bahrain', 'yemen', 'iraq', 'syria',
'jordan', 'lebanon', 'palestine', 'libya', 'tunisia',
'algeria', 'morocco', 'sudan'
],
germanosphere: [
'^germany$', '^austria$', '^switzerland$', '^luxembourg$', '^liechtenstein$'
],
slavic: [
'^russia$', '^ukraine$', '^belarus$', '^poland$', '^czech republic$',
'^slovakia$', '^serbia$', '^croatia$', '^bosnia$', '^montenegro$',
'^slovenia$', '^bulgaria$', '^north macedonia$'
],
sinosphere: [
'^china$', 'hong kong', '^taiwan$', '^singapore$', '^macau$'
],
indosphere: [
'^india$', '^pakistan$', '^bangladesh$', '^nepal$', '^sri lanka$',
'^bhutan$', '^maldives$'
],
turkic: [
'^turkey$', '^azerbaijan$', '^uzbekistan$', '^kazakhstan$',
'^kyrgyzstan$', '^turkmenistan$'
],
nordic: [
'^sweden$', '^norway$', '^denmark$', '^finland$', '^iceland$',
'faroe islands', '^greenland$'
],
baltic: [
'^estonia$', '^latvia$', '^lithuania$'
],
hellenic: [
'^greece$', '^cyprus$'
],
benelux: [
'^netherlands$', '^belgium$', '^luxembourg$'
],
persian: [
'^iran$', '^afghanistan$', '^tajikistan$'
],
malaysphere: [
'^malaysia$', '^brunei$', '^indonesia$'
],
korean: [
'south korea', 'korea', 'north korea'
],
japanese: [
'^japan$'
],
vietnamese: [
'^vietnam$'
],
thai: [
'^thailand$'
]
};
function getCulturalGroup(channelInfo) {
// Look specifically for group-title (country information)
const groupMatch = channelInfo.match(/group-title="([^"]*)"/) || [];
const groupTitle = (groupMatch[1] || '').toLowerCase();
// Check if the country belongs to any cultural group
for (const [group, countries] of Object.entries(culturalGroups)) {
// Use exact matching with RegExp
if (countries.some(country => {
// If the country pattern starts with ^, use it as a RegExp
if (country.startsWith('^')) {
const regex = new RegExp(country, 'i');
return regex.test(groupTitle);
}
// Otherwise, use includes for flexible matching (for multi-word countries)
return groupTitle.includes(country);
})) {
return group;
}
}
return null;
}
function splitByCulturalGroup(filePath) {
const content = fs.readFileSync(filePath, 'utf8');
const lines = content.split('\n');
// Create cultural groups directory
const groupsDir = path.join(path.dirname(filePath), 'cultural-groups');
if (!fs.existsSync(groupsDir)) {
fs.mkdirSync(groupsDir);
}
// Get existing cultural group files
const existingFiles = fs.readdirSync(groupsDir)
.filter(file => file.endsWith('.m3u'))
.map(file => file.toLowerCase());
const groups = {};
// Verify M3U header
const header = lines[0];
if (!header.startsWith('#EXTM3U')) {
throw new Error('Invalid M3U file: Missing #EXTM3U header');
}
// Process the file line by line to handle multi-line entries
let currentCulturalGroup = null;
let currentEntry = [];
let isInEntry = false;
for (let i = 1; i < lines.length; i++) {
const line = lines[i].trim();
if (!line) continue;
if (line.startsWith('#EXTINF')) {
// Start of a new entry
isInEntry = true;
currentEntry = [line];
// Determine cultural group
currentCulturalGroup = getCulturalGroup(line);
} else if (isInEntry) {
// Add the line to the current entry
currentEntry.push(line);
// If this is a URL line (doesn't start with #), this completes the entry
if (!line.startsWith('#')) {
// Add all lines of the entry to the appropriate cultural group
if (currentCulturalGroup) {
if (!groups[currentCulturalGroup]) {
groups[currentCulturalGroup] = ['#EXTM3U'];
}
groups[currentCulturalGroup].push(...currentEntry);
}
// Reset for the next entry
isInEntry = false;
currentEntry = [];
currentCulturalGroup = null;
}
}
}
// Get list of current cultural group files
const currentGroupFiles = Object.keys(groups).map(groupTitle =>
`${groupTitle.toLowerCase()}.m3u`
);
// Remove obsolete files
existingFiles.forEach(existingFile => {
if (!currentGroupFiles.includes(existingFile)) {
const fileToRemove = path.join(groupsDir, existingFile);
fs.unlinkSync(fileToRemove);
console.log(`Removed obsolete cultural group playlist: ${existingFile}`);
}
});
// Write cultural group files
Object.entries(groups).forEach(([groupTitle, groupLines]) => {
const groupFilePath = path.join(groupsDir, `${groupTitle.toLowerCase()}.m3u`);
fs.writeFileSync(groupFilePath, groupLines.join('\n') + '\n');
console.log(`Created/updated cultural group playlist: ${groupFilePath}`);
});
// Generate summary - count entries properly by counting #EXTINF lines
const summary = Object.entries(groups).map(([group, lines]) => {
const channelCount = lines.filter(line => line.startsWith('#EXTINF')).length;
return `${group}: ${channelCount} channels`;
});
console.log('\nCultural group split summary:');
console.log(summary.join('\n'));
}
const filePath = process.argv[2];
if (!filePath) {
console.error('Please provide the path to the M3U file');
process.exit(1);
}
try {
splitByCulturalGroup(filePath);
} catch (error) {
console.error('Error splitting M3U file:', error.message);
process.exit(1);
}