diff options
Diffstat (limited to 'src/util')
| -rw-r--r-- | src/util/cli.js | 69 | ||||
| -rw-r--r-- | src/util/magic-constants.js | 1 | ||||
| -rw-r--r-- | src/util/transform-content.js | 453 | ||||
| -rw-r--r-- | src/util/urls.js | 118 |
4 files changed, 614 insertions, 27 deletions
diff --git a/src/util/cli.js b/src/util/cli.js index f1a31900..1ddc90e0 100644 --- a/src/util/cli.js +++ b/src/util/cli.js @@ -64,8 +64,10 @@ export async function parseOptions(options, optionDescriptorMap) { // options is the array of options you want to process; // optionDescriptorMap is a mapping of option names to objects that describe // the expected value for their corresponding options. - // Returned is a mapping of any specified option names to their values, or - // a process.exit(1) and error message if there were any issues. + // + // Returned is... + // - a mapping of any specified option names to their values + // - a process.exit(1) and error message if there were any issues // // Here are examples of optionDescriptorMap to cover all the things you can // do with it: @@ -95,11 +97,10 @@ export async function parseOptions(options, optionDescriptorMap) { // ['--directory', 'apple'] -> {'directory': 'apple'} // ['--directory', 'artichoke'] -> (error) // ['--files', 'a', 'b', 'c', ';'] -> {'files': ['a', 'b', 'c']} - // - // TODO: Be able to validate the values in a series option. const handleDashless = optionDescriptorMap[parseOptions.handleDashless]; const handleUnknown = optionDescriptorMap[parseOptions.handleUnknown]; + const result = Object.create(null); for (let i = 0; i < options.length; i++) { const option = options[i]; @@ -107,6 +108,7 @@ export async function parseOptions(options, optionDescriptorMap) { // --x can be a flag or expect a value or series of values let name = option.slice(2).split('=')[0]; // '--x'.split('=') = ['--x'] let descriptor = optionDescriptorMap[name]; + if (!descriptor) { if (handleUnknown) { handleUnknown(option); @@ -116,36 +118,49 @@ export async function parseOptions(options, optionDescriptorMap) { } continue; } + if (descriptor.alias) { name = descriptor.alias; descriptor = optionDescriptorMap[name]; } - if (descriptor.type === 'flag') { - result[name] = true; - } else if (descriptor.type === 'value') { - let value = option.slice(2).split('=')[1]; - if (!value) { - value = options[++i]; - if (!value || value.startsWith('-')) { - value = null; - } + + switch (descriptor.type) { + case 'flag': { + result[name] = true; + break; } - if (!value) { - console.error(`Expected a value for --${name}`); - process.exit(1); + + case 'value': { + let value = option.slice(2).split('=')[1]; + if (!value) { + value = options[++i]; + if (!value || value.startsWith('-')) { + value = null; + } + } + + if (!value) { + console.error(`Expected a value for --${name}`); + process.exit(1); + } + + result[name] = value; + break; } - result[name] = value; - } else if (descriptor.type === 'series') { - if (!options.slice(i).includes(';')) { - console.error( - `Expected a series of values concluding with ; (\\;) for --${name}` - ); - process.exit(1); + + case 'series': { + if (!options.slice(i).includes(';')) { + console.error(`Expected a series of values concluding with ; (\\;) for --${name}`); + process.exit(1); + } + + const endIndex = i + options.slice(i).indexOf(';'); + result[name] = options.slice(i + 1, endIndex); + i = endIndex; + break; } - const endIndex = i + options.slice(i).indexOf(';'); - result[name] = options.slice(i + 1, endIndex); - i = endIndex; } + if (descriptor.validate) { const validation = await descriptor.validate(result[name]); if (validation !== true) { @@ -167,10 +182,12 @@ export async function parseOptions(options, optionDescriptorMap) { } continue; } + if (descriptor.alias) { name = descriptor.alias; descriptor = optionDescriptorMap[name]; } + if (descriptor.type === 'flag') { result[name] = true; } else { diff --git a/src/util/magic-constants.js b/src/util/magic-constants.js index 73fdbc6d..83dd7db5 100644 --- a/src/util/magic-constants.js +++ b/src/util/magic-constants.js @@ -7,4 +7,3 @@ // (TM). export const OFFICIAL_GROUP_DIRECTORY = 'official'; -export const FANDOM_GROUP_DIRECTORY = 'fandom'; diff --git a/src/util/transform-content.js b/src/util/transform-content.js new file mode 100644 index 00000000..d1d0f51a --- /dev/null +++ b/src/util/transform-content.js @@ -0,0 +1,453 @@ +// See also replacer.js, which covers the actual syntax parser and node +// interpreter. This file works with replacer.js to provide higher-level +// interfaces for converting various content found in wiki data to HTML for +// display on the site. + +import * as html from './html.js'; +export {transformInline} from './replacer.js'; + +export const replacerSpec = { + album: { + find: 'album', + link: 'album', + }, + 'album-commentary': { + find: 'album', + link: 'albumCommentary', + }, + 'album-gallery': { + find: 'album', + link: 'albumGallery', + }, + artist: { + find: 'artist', + link: 'artist', + }, + 'artist-gallery': { + find: 'artist', + link: 'artistGallery', + }, + 'commentary-index': { + find: null, + link: 'commentaryIndex', + }, + date: { + find: null, + value: (ref) => new Date(ref), + html: (date, {language}) => + html.tag('time', + {datetime: date.toString()}, + language.formatDate(date)), + }, + 'flash-index': { + find: null, + link: 'flashIndex', + }, + flash: { + find: 'flash', + link: 'flash', + transformName(name, node, input) { + const nextCharacter = input[node.iEnd]; + const lastCharacter = name[name.length - 1]; + if (![' ', '\n', '<'].includes(nextCharacter) && lastCharacter === '.') { + return name.slice(0, -1); + } else { + return name; + } + }, + }, + group: { + find: 'group', + link: 'groupInfo', + }, + 'group-gallery': { + find: 'group', + link: 'groupGallery', + }, + home: { + find: null, + link: 'home', + }, + 'listing-index': { + find: null, + link: 'listingIndex', + }, + listing: { + find: 'listing', + link: 'listing', + }, + media: { + find: null, + link: 'media', + }, + 'news-index': { + find: null, + link: 'newsIndex', + }, + 'news-entry': { + find: 'newsEntry', + link: 'newsEntry', + }, + root: { + find: null, + link: 'root', + }, + site: { + find: null, + link: 'site', + }, + static: { + find: 'staticPage', + link: 'staticPage', + }, + string: { + find: null, + value: (ref) => ref, + html: (ref, {language, args}) => language.$(ref, args), + }, + tag: { + find: 'artTag', + link: 'tag', + }, + track: { + find: 'track', + link: 'track', + }, +}; + +function splitLines(text) { + return text.split(/\r\n|\r|\n/); +} + +function joinLineBreaks(sourceLines) { + const outLines = []; + + let lineSoFar = ''; + for (let i = 0; i < sourceLines.length; i++) { + const line = sourceLines[i]; + lineSoFar += line; + if (!line.endsWith('<br>')) { + outLines.push(lineSoFar); + lineSoFar = ''; + } + } + + if (lineSoFar) { + outLines.push(lineSoFar); + } + + return outLines; +} + +function parseAttributes(string, {to}) { + const attributes = Object.create(null); + const skipWhitespace = (i) => { + const ws = /\s/; + if (ws.test(string[i])) { + const match = string.slice(i).match(/[^\s]/); + if (match) { + return i + match.index; + } else { + return string.length; + } + } else { + return i; + } + }; + + for (let i = 0; i < string.length; ) { + i = skipWhitespace(i); + const aStart = i; + const aEnd = i + string.slice(i).match(/[\s=]|$/).index; + const attribute = string.slice(aStart, aEnd); + i = skipWhitespace(aEnd); + if (string[i] === '=') { + i = skipWhitespace(i + 1); + let end, endOffset; + if (string[i] === '"' || string[i] === "'") { + end = string[i]; + endOffset = 1; + i++; + } else { + end = '\\s'; + endOffset = 0; + } + const vStart = i; + const vEnd = i + string.slice(i).match(new RegExp(`${end}|$`)).index; + const value = string.slice(vStart, vEnd); + i = vEnd + endOffset; + if (attribute === 'src' && value.startsWith('media/')) { + attributes[attribute] = to('media.path', value.slice('media/'.length)); + } else { + attributes[attribute] = value; + } + } else { + attributes[attribute] = attribute; + } + } + return Object.fromEntries( + Object.entries(attributes).map(([key, val]) => [ + key, + val === 'true' + ? true + : val === 'false' + ? false + : val === key + ? true + : val, + ]) + ); +} + +function unbound_transformMultiline(text, { + img, + to, + transformInline, + + thumb = null, +}) { + // Heck yes, HTML magics. + + text = transformInline(text.trim()); + + const outLines = []; + + const indentString = ' '.repeat(4); + + let levelIndents = []; + const openLevel = (indent) => { + // opening a sublist is a pain: to be semantically *and* visually + // correct, we have to append the <ul> at the end of the existing + // previous <li> + const previousLine = outLines[outLines.length - 1]; + if (previousLine?.endsWith('</li>')) { + // we will re-close the <li> later + outLines[outLines.length - 1] = previousLine.slice(0, -5) + ' <ul>'; + } else { + // if the previous line isn't a list item, this is the opening of + // the first list level, so no need for indent + outLines.push('<ul>'); + } + levelIndents.push(indent); + }; + const closeLevel = () => { + levelIndents.pop(); + if (levelIndents.length) { + // closing a sublist, so close the list item containing it too + outLines.push(indentString.repeat(levelIndents.length) + '</ul></li>'); + } else { + // closing the final list level! no need for indent here + outLines.push('</ul>'); + } + }; + + // okay yes we should support nested formatting, more than one blockquote + // layer, etc, but hear me out here: making all that work would basically + // be the same as implementing an entire markdown converter, which im not + // interested in doing lol. sorry!!! + let inBlockquote = false; + + let lines = splitLines(text); + lines = joinLineBreaks(lines); + for (let line of lines) { + const imageLine = line.startsWith('<img'); + line = line.replace(/<img (.*?)>/g, (match, attributes) => + img({ + lazy: true, + link: true, + thumb, + ...parseAttributes(attributes, {to}), + }) + ); + + let indentThisLine = 0; + let lineContent = line; + let lineTag = 'p'; + + const listMatch = line.match(/^( *)- *(.*)$/); + if (listMatch) { + // is a list item! + if (!levelIndents.length) { + // first level is always indent = 0, regardless of actual line + // content (this is to avoid going to a lesser indent than the + // initial level) + openLevel(0); + } else { + // find level corresponding to indent + const indent = listMatch[1].length; + let i; + for (i = levelIndents.length - 1; i >= 0; i--) { + if (levelIndents[i] <= indent) break; + } + // note: i cannot equal -1 because the first indentation level + // is always 0, and the minimum indentation is also 0 + if (levelIndents[i] === indent) { + // same indent! return to that level + while (levelIndents.length - 1 > i) closeLevel(); + // (if this is already the current level, the above loop + // will do nothing) + } else if (levelIndents[i] < indent) { + // lesser indent! branch based on index + if (i === levelIndents.length - 1) { + // top level is lesser: add a new level + openLevel(indent); + } else { + // lower level is lesser: return to that level + while (levelIndents.length - 1 > i) closeLevel(); + } + } + } + // finally, set variables for appending content line + indentThisLine = levelIndents.length; + lineContent = listMatch[2]; + lineTag = 'li'; + } else { + // not a list item! close any existing list levels + while (levelIndents.length) closeLevel(); + + // like i said, no nested shenanigans - quotes only appear outside + // of lists. sorry! + const quoteMatch = line.match(/^> *(.*)$/); + if (quoteMatch) { + // is a quote! open a blockquote tag if it doesnt already exist + if (!inBlockquote) { + inBlockquote = true; + outLines.push('<blockquote>'); + } + indentThisLine = 1; + lineContent = quoteMatch[1]; + } else if (inBlockquote) { + // not a quote! close a blockquote tag if it exists + inBlockquote = false; + outLines.push('</blockquote>'); + } + + // let some escaped symbols display as the normal symbol, since the + // point of escaping them is just to avoid having them be treated as + // syntax markers! + if (lineContent.match(/( *)\\-/)) { + lineContent = lineContent.replace('\\-', '-'); + } else if (lineContent.match(/( *)\\>/)) { + lineContent = lineContent.replace('\\>', '>'); + } + } + + if (lineTag === 'p') { + // certain inline element tags should still be postioned within a + // paragraph; other elements (e.g. headings) should be added as-is + const elementMatch = line.match(/^<(.*?)[ >]/); + if ( + elementMatch && + !imageLine && + ![ + 'a', + 'abbr', + 'b', + 'bdo', + 'br', + 'cite', + 'code', + 'data', + 'datalist', + 'del', + 'dfn', + 'em', + 'i', + 'img', + 'ins', + 'kbd', + 'mark', + 'output', + 'picture', + 'q', + 'ruby', + 'samp', + 'small', + 'span', + 'strong', + 'sub', + 'sup', + 'svg', + 'time', + 'var', + 'wbr', + ].includes(elementMatch[1]) + ) { + lineTag = ''; + } + + // for sticky headings! + if (elementMatch && elementMatch[1] === 'h2') { + lineContent = lineContent.replace(/<h2(.*?)>/g, (match, attributes) => { + const parsedAttributes = parseAttributes(attributes, {to}); + return `<h2 ${html.attributes({ + ...parsedAttributes, + class: [...parsedAttributes.class?.split(' ') ?? [], 'content-heading'], + })}>`; + }); + } + } + + let pushString = indentString.repeat(indentThisLine); + if (lineTag) { + pushString += `<${lineTag}>${lineContent}</${lineTag}>`; + } else { + pushString += lineContent; + } + outLines.push(pushString); + } + + // after processing all lines... + + // if still in a list, close all levels + while (levelIndents.length) closeLevel(); + + // if still in a blockquote, close its tag + if (inBlockquote) { + inBlockquote = false; + outLines.push('</blockquote>'); + } + + return outLines.join('\n'); +} + +function unbound_transformLyrics(text, { + transformInline, + transformMultiline, +}) { + // Different from transformMultiline 'cuz it joins multiple lines together + // with line 8reaks (<br>); transformMultiline treats each line as its own + // complete paragraph (or list, etc). + + // If it looks like old data, then like, oh god. + // Use the normal transformMultiline tool. + if (text.includes('<br')) { + return transformMultiline(text); + } + + text = transformInline(text.trim()); + + let buildLine = ''; + const addLine = () => outLines.push(`<p>${buildLine}</p>`); + const outLines = []; + for (const line of text.split('\n')) { + if (line.length) { + if (buildLine.length) { + buildLine += '<br>'; + } + buildLine += line; + } else if (buildLine.length) { + addLine(); + buildLine = ''; + } + } + if (buildLine.length) { + addLine(); + } + return outLines.join('\n'); +} + +export { + unbound_transformLyrics as transformLyrics, + unbound_transformMultiline as transformMultiline +} diff --git a/src/util/urls.js b/src/util/urls.js index 1f9cd9c0..c2119b8d 100644 --- a/src/util/urls.js +++ b/src/util/urls.js @@ -136,3 +136,121 @@ export const thumb = { medium: thumbnailHelper('.medium'), small: thumbnailHelper('.small'), }; + +// Makes the generally-used and wiki-specialized "to" page utility. +// "to" returns a relative path from the current page to the target. +export function getURLsFrom({ + baseDirectory, + pagePath, + urls, +}) { + const pageSubKey = pagePath[0]; + const subdirectoryPrefix = getPageSubdirectoryPrefix({pagePath}); + + return (targetFullKey, ...args) => { + const [groupKey, subKey] = targetFullKey.split('.'); + let from, to; + + // When linking to *outside* the localized area of the site, we need to + // make sure the result is correctly relative to the 8ase directory. + if ( + groupKey !== 'localized' && + groupKey !== 'localizedDefaultLanguage' && + baseDirectory + ) { + from = 'localizedWithBaseDirectory.' + pageSubKey; + to = targetFullKey; + } else if (groupKey === 'localizedDefaultLanguage' && baseDirectory) { + // Special case for specifically linking *from* a page with base + // directory *to* a page without! Used for the language switcher and + // hopefully nothing else oh god. + from = 'localizedWithBaseDirectory.' + pageSubKey; + to = 'localized.' + subKey; + } else if (groupKey === 'localizedDefaultLanguage') { + // Linking to the default, except surprise, we're already IN the default + // (no baseDirectory set). + from = 'localized.' + pageSubKey; + to = 'localized.' + subKey; + } else { + // If we're linking inside the localized area (or there just is no + // 8ase directory), the 8ase directory doesn't matter. + from = 'localized.' + pageSubKey; + to = targetFullKey; + } + + return ( + subdirectoryPrefix + + urls.from(from).to(to, ...args)); + }; +} + +// Makes the generally-used and wiki-specialized "absoluteTo" page utility. +// "absoluteTo" returns an absolute path, starting at site root (/) leading +// to the target. +export function getURLsFromRoot({ + baseDirectory, + urls, +}) { + const {to} = urls.from('shared.root'); + + return (targetFullKey, ...args) => { + const [groupKey, subKey] = targetFullKey.split('.'); + return ( + '/' + + (groupKey === 'localized' && baseDirectory + ? to( + 'localizedWithBaseDirectory.' + subKey, + baseDirectory, + ...args + ) + : to(targetFullKey, ...args)) + ); + }; +} + +export function getPagePathname({ + baseDirectory, + device = false, + pagePath, + urls, +}) { + const {[device ? 'toDevice' : 'to']: to} = urls.from('shared.root'); + + return (baseDirectory + ? to('localizedWithBaseDirectory.' + pagePath[0], baseDirectory, ...pagePath.slice(1)) + : to('localized.' + pagePath[0], ...pagePath.slice(1))); +} + +export function getPagePathnameAcrossLanguages({ + defaultLanguage, + languages, + pagePath, + urls, +}) { + return withEntries(languages, entries => entries + .filter(([key, language]) => key !== 'default' && !language.hidden) + .map(([_key, language]) => [ + language.code, + getPagePathname({ + baseDirectory: + (language === defaultLanguage + ? '' + : language.code), + pagePath, + urls, + }), + ])); +} + +// Needed for the rare path arguments which themselves contains one or more +// slashes, e.g. for listings, with arguments like 'albums/by-name'. +export function getPageSubdirectoryPrefix({ + pagePath, +}) { + const timesNestedDeeply = (pagePath + .slice(1) // skip URL key, only check arguments + .join('/') + .split('/') + .length - 1); + return '../'.repeat(timesNestedDeeply); +} |