Start generalizing to work for Ward too

* Updates the command-line options in various ways, notably changing --start-url to --book
* Renames the "book" directory to "staging"
* Abstracts out most book-specific data into books.js (but not yet cover or substitutions.json)
* Makes per-book subfolders under cache and staging directories

.gitignore

@@ -2,5 +2,5 @@
 /npm-debug.log
 
 cache/
-book/
+staging/
-Worm.epub
+*.epub

lib/books.js

@@ -0,0 +1,41 @@
+"use strict";
+
+// TODO cover
+
+exports.worm = {
+  startURL: "https://parahumans.wordpress.com/2011/06/11/1-1/",
+  title: "Worm",
+  author: "wildbow",
+  id: "e7f3532d-8db6-4888-be80-1976166b7059",
+
+  // First paragraph of https://parahumans.wordpress.com/about/
+  description: `
+An introverted teenage girl with an unconventional superpower, Taylor goes out in costume to find escape from a deeply
+unhappy and frustrated civilian life. Her first attempt at taking down a supervillain sees her mistaken for one,
+thrusting her into the midst of the local ‘cape’ scene’s politics, unwritten rules, and ambiguous morals. As she risks
+life and limb, Taylor faces the dilemma of having to do the wrong things for the right reasons.
+`.trim()
+};
+
+exports.ward = {
+  startURL: "https://www.parahumans.net/2017/09/11/daybreak-1-1/",
+  title: "Ward",
+  author: "wildbow",
+  id: "a6b6b156-2f17-43c0-8bb1-bfa91f3ef62a",
+
+  // Synposis from https://www.parahumans.net/
+  description: `
+The unwritten rules that govern the fights and outright wars between ‘capes’ have been amended: everyone gets their
+second chance. It’s an uneasy thing to come to terms with when notorious supervillains and even monsters are playing at
+being hero. The world ended two years ago, and as humanity straddles the old world and the new, there aren’t records,
+witnesses, or facilities to answer the villains’ past actions in the present. One of many compromises, uneasy truces and
+deceptions that are starting to splinter as humanity rebuilds.
+
+None feel the injustice of this new status quo or the lack of established footing more than the past residents of the
+parahuman asylums. The facilities hosted parahumans and their victims, but the facilities are ruined or gone; one of
+many fragile ex-patients is left to find a place in a fractured world. She’s perhaps the person least suited to have
+anything to do with this tenuous peace or to stand alongside these false heroes. She’s put in a position to make the
+decision: will she compromise to help forge what they call, with dark sentiment, a second golden age? Or will she stand
+tall as a gilded dark age dawns?
+`.trim()
+};

lib/scaffold.js

@@ -3,17 +3,7 @@ const fs = require("fs").promises;
 const path = require("path");
 const cpr = require("util").promisify(require("cpr"));
 
-const BOOK_TITLE = "Worm";
-const BOOK_AUTHOR = "wildbow";
 const BOOK_PUBLISHER = "Domenic Denicola";
-const BOOK_ID = "urn:uuid:e7f3532d-8db6-4888-be80-1976166b7059";
-
-// First paragraph of https://parahumans.wordpress.com/about/
-const BOOK_DESCRIPTION = `
-An introverted teenage girl with an unconventional superpower, Taylor goes out in costume to find escape from a deeply
-unhappy and frustrated civilian life. Her first attempt at taking down a supervillain sees her mistaken for one,
-thrusting her into the midst of the local ‘cape’ scene’s politics, unwritten rules, and ambiguous morals. As she risks
-life and limb, Taylor faces the dilemma of having to do the wrong things for the right reasons.`;
 
 const NCX_FILENAME = "toc.ncx";
 
@@ -21,13 +11,13 @@ const COVER_IMG_FILENAME = "cover.png";
 const COVER_XHTML_FILENAME = "cover.xhtml";
 const COVER_MIMETYPE = "image/png";
 
-module.exports = async (scaffoldingPath, bookPath, contentPath, chaptersPath, manifestPath) => {
+module.exports = async (scaffoldingPath, bookPath, contentPath, chaptersPath, manifestPath, bookInfo) => {
   await Promise.all([
     cpr(scaffoldingPath, bookPath, { overwrite: true, confirm: true, filter: noThumbs }),
     getChapters(contentPath, chaptersPath, manifestPath).then(chapters => {
       return Promise.all([
-        writeOPF(chapters, contentPath),
+        writeOPF(chapters, contentPath, bookInfo),
-        writeNcx(chapters, contentPath)
+        writeNcx(chapters, contentPath, bookInfo)
       ]);
     })
   ]);
@@ -38,7 +28,7 @@ function noThumbs(filePath) {
   return path.basename(filePath) !== "Thumbs.db";
 }
 
-function writeOPF(chapters, contentPath) {
+function writeOPF(chapters, contentPath, bookInfo) {
   const manifestChapters = chapters.map(c => {
     return `<item id="${c.id}" href="${c.href}" media-type="application/xhtml+xml"/>`;
   }).join("\n");
@@ -51,12 +41,12 @@ function writeOPF(chapters, contentPath) {
 <package version="2.0" xmlns="http://www.idpf.org/2007/opf" unique-identifier="BookId">
 
   <metadata xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:opf="http://www.idpf.org/2007/opf">
-    <dc:title>${BOOK_TITLE}</dc:title>
+    <dc:title>${bookInfo.title}</dc:title>
     <dc:language>en</dc:language>
-    <dc:identifier id="BookId" opf:scheme="UUID">${BOOK_ID}</dc:identifier>
+    <dc:identifier id="BookId" opf:scheme="UUID">urn:uuid:${bookInfo.id}</dc:identifier>
-    <dc:creator opf:file-as="${BOOK_AUTHOR}" opf:role="aut">${BOOK_AUTHOR}</dc:creator>
+    <dc:creator opf:file-as="${bookInfo.author}" opf:role="aut">${bookInfo.author}</dc:creator>
     <dc:publisher>${BOOK_PUBLISHER}</dc:publisher>
-    <dc:description>${BOOK_DESCRIPTION}</dc:description>
+    <dc:description>${bookInfo.description}</dc:description>
     <meta name="cover" content="cover-image"/>
   </metadata>
 
@@ -80,7 +70,7 @@ ${spineChapters}
   return fs.writeFile(path.resolve(contentPath, "content.opf"), contents);
 }
 
-function writeNcx(chapters, contentPath) {
+function writeNcx(chapters, contentPath, bookInfo) {
   const navPoints = chapters.map((c, i) => {
     return `<navPoint class="chapter" id="${c.id}" playOrder="${i + 1}">
   <navLabel><text>${c.title}</text></navLabel>
@@ -92,18 +82,18 @@ function writeNcx(chapters, contentPath) {
 <!DOCTYPE ncx PUBLIC "-//NISO//DTD ncx 2005-1//EN" "http://www.daisy.org/z3986/2005/ncx-2005-1.dtd">
  <ncx version="2005-1" xml:lang="en" xmlns="http://www.daisy.org/z3986/2005/ncx/">
   <head>
-    <meta name="dtb:uid" content="${BOOK_ID}"/>
+    <meta name="dtb:uid" content="urn:uuid:${bookInfo.id}"/>
     <meta name="dtb:depth" content="1"/>
     <meta name="dtb:totalPageCount" content="0"/>
     <meta name="dtb:maxPageNumber" content="0"/>
   </head>
 
   <docTitle>
-    <text>${BOOK_TITLE}</text>
+    <text>${bookInfo.title}</text>
   </docTitle>
 
   <docAuthor>
-    <text>${BOOK_AUTHOR}</text>
+    <text>${bookInfo.author}</text>
   </docAuthor>
 
   <navMap>

lib/worm-scraper.js

@@ -6,51 +6,54 @@ const fs = require("fs").promises;
 const yargs = require("yargs");
 
 const packageJson = require("../package.json");
+const books = require("./books.js");
 const download = require("./download.js");
 const convert = require("./convert.js");
 const scaffold = require("./scaffold.js");
 const zip = require("./zip.js");
 
+const OUTPUT_DEFAULT = "(Book name).epub";
+
 const argv = yargs
   .usage(`${packageJson.description}\n\n${packageJson.name} [<command1> [<command2> [<command3> ...]]]\n\n` +
          "Each command will fail if the previously-listed one has not yet been run (with matching options).")
-  .command("download", "download all chapters by crawling parahumans.wordpress.com")
+  .command("download", "download all chapters into the cache")
-  .command("convert", "convert the raw chapter HTML files into cleaned-up ebook chapters")
+  .command("convert", "convert the raw HTML into cleaned-up ebook chapters")
-  .command("scaffold", "assemble the table of contents, etc. to complete the EPUB")
+  .command("scaffold", "assemble the table of contents, etc.")
-  .command("zip", "zip up the EPUB files into a .epub output")
+  .command("zip", "zip up the created files into a .epub output")
-  .option("s", {
+  .option("b", {
-    alias: "start-url",
+    alias: "book",
-    default: "https://parahumans.wordpress.com/2011/06/11/1-1/",
+    default: Object.keys(books)[0],
-    describe: "the URL from which to start crawling, for the download command",
+    describe: "the book to operate on",
+    choices: Object.keys(books),
     requiresArg: true,
     global: true
   })
   .option("c", {
-    alias: "cache-directory",
+    alias: "cache",
     default: "cache",
-    describe: "cache directory, for the download and convert commands",
+    describe: "cache directory for downloaded raw chapters",
     requiresArg: true,
     global: true
   })
-  .option("b", {
+  .option("s", {
-    alias: "book-directory",
+    alias: "staging",
-    default: "book",
+    default: "staging",
-    describe: "directory in which to assemble the EPUB files before zipping, for the convert, scaffold, and zip " +
+    describe: "directory in which to assemble the EPUB files",
-              "commands",
     requiresArg: true,
     global: true
   })
   .option("o", {
     alias: "out",
-    default: "Worm.epub",
+    default: OUTPUT_DEFAULT,
-    describe: "output file destination, for the zip command",
+    describe: "output file destination",
     requiresArg: true,
     global: true
   })
   .option("j", {
     alias: "jobs",
     default: 10,
-    describe: "Number of concurrent read/write jobs to run, for the convert command",
+    describe: "Number of concurrent read/write conversion jobs",
     requiresArg: true,
     global: true
   })
@@ -60,18 +63,21 @@ const argv = yargs
   .version()
   .argv;
 
-const cachePath = path.resolve(argv.cacheDirectory);
+const outputFilename = argv.out === OUTPUT_DEFAULT ? `${books[argv.book].title}.epub` : argv.out;
+
+const cachePath = path.resolve(argv.cache, argv.book);
 const manifestPath = path.resolve(cachePath, "manifest.json");
 
 const scaffoldingPath = path.resolve(__dirname, "../scaffolding");
-const bookPath = path.resolve(argv.bookDirectory);
+const stagingPath = path.resolve(argv.staging, argv.book);
-const contentPath = path.resolve(bookPath, "OEBPS");
+const contentPath = path.resolve(stagingPath, "OEBPS");
 const chaptersPath = path.resolve(contentPath, "chapters");
 
 const commands = [];
 
 if (argv._.includes("download")) {
-  commands.push(() => download(argv.startUrl, cachePath, manifestPath));
+  const startURL = books[argv.book].startURL;
+  commands.push(() => download(startURL, cachePath, manifestPath));
 }
 
 if (argv._.includes("convert")) {
@@ -83,11 +89,12 @@ if (argv._.includes("convert")) {
 }
 
 if (argv._.includes("scaffold")) {
-  commands.push(() => scaffold(scaffoldingPath, bookPath, contentPath, chaptersPath, manifestPath));
+  const bookInfo = books[argv.book];
+  commands.push(() => scaffold(scaffoldingPath, stagingPath, contentPath, chaptersPath, manifestPath, bookInfo));
 }
 
 if (argv._.includes("zip")) {
-  commands.push(() => zip(bookPath, contentPath, path.resolve(argv.out)));
+  commands.push(() => zip(stagingPath, contentPath, path.resolve(outputFilename)));
 }
 
 (async () => {

scaffolding/OEBPS/cover.xhtml

@@ -15,6 +15,6 @@
     </style>
   </head>
   <body>
-    <img src="cover.png" alt="Worm"/>
+    <img src="cover.png" alt=""/>
   </body>
 </html>