Fix kepub issue, bump version and update deps

2024-04-26 08:32:34 +12:00 · 2018-05-17 11:36:04 +02:00 · 2018-05-17 11:36:04 +02:00 · ce3e7022b6
parent 3c18655eed
commit ce3e7022b6
3 changed files with 70 additions and 42 deletions
--- a/package.json
+++ b/package.json
@ -1,6 +1,6 @@
 {
  "name": "fimfic2epub",
-  "version": "1.7.32",
+  "version": "1.7.33",
  "description": "Tool to generate improved EPUB ebooks from Fimfiction stories",
  "author": "djazz",
  "license": "MIT",
@ -27,7 +27,7 @@
    "detect-node": "^2.0.3",
    "elementtree": "^0.1.7",
    "escape-string-regexp": "^1.0.5",
-    "file-type": "^7.7.1",
+    "file-type": "^8.0.0",
    "fonteditor-core": "^1.0.2",
    "html-entities": "^1.2.1",
    "html-to-text": "^4.0.0",
@ -39,7 +39,7 @@
    "mithril-node-render": "^2.3.0",
    "node-png": "^0.4.3",
    "pretty-data": "^0.40.0",
-    "request": "^2.85.0",
+    "request": "^2.86.0",
    "sanitize-filename": "^1.6.1",
    "syllable": "^3.0.0",
    "twemoji": "^2.5.1",
@ -51,7 +51,7 @@
    "autosize": "^4.0.2",
    "babel-core": "^6.26.3",
    "babel-loader": "^7.1.4",
-    "babel-preset-env": "^1.6.1",
+    "babel-preset-env": "^1.7.0",
    "babel-register": "^6.26.0",
    "binary-loader": "0.0.1",
    "del": "^3.0.0",
@ -66,7 +66,7 @@
    "gulp-chmod": "^2.0.0",
    "gulp-filter": "^5.1.0",
    "gulp-json-editor": "^2.3.0",
-    "gulp-rename": "^1.2.2",
+    "gulp-rename": "^1.2.3",
    "gulp-standard": "^11.0.0",
    "gulp-util": "^3.0.8",
    "gulp-watch": "^5.0.0",
@ -79,7 +79,7 @@
    "standard": "^11.0.1",
    "stylus": "^0.54.5",
    "stylus-loader": "^3.0.2",
-    "webpack": "^4.8.1",
+    "webpack": "^4.8.3",
    "webpack-node-externals": "^1.7.2"
  },
  "standard": {
--- a/src/kepubify.js
+++ b/src/kepubify.js
@ -8,7 +8,10 @@ export default function kepubify (html) {
  const body = tree.find('./body')
  addDivs(body)
  const state = {paragraph: 0, segment: 0}
-  body.getchildren().forEach((child) => addSpansToNode(child, body, state))
+  body.getchildren().forEach((child) => {
+    fixupTree(child, body)
+    addSpansToNode(child, body, state)
+  })
  return '<?xml version="1.0" encoding="utf-8"?>\n<!DOCTYPE html>\n' + tree.write({
    xml_declaration: false
  })
@ -33,7 +36,7 @@ function createSpan (paragraph, segment) {
  return span
 }

-function textToSpans (node, text, state) {
+function textToSentences (text) {
  const tokenSentences = text
    .replace('\0', '')
    .replace(/\s+/g, ' ') // Replace all whitespace (including newlines) with a single space
@ -52,45 +55,62 @@ function textToSpans (node, text, state) {
  }

  return tokenSentences.map((sentence, i) => {
-    const span = createSpan(state.paragraph, state.segment++)
-    span.text = sentence
-    return span
+    // const span = createSpan(state.paragraph, state.segment++)
+    // span.text = sentence
+    return sentence
+  })
+}
+
+// Makes text nodes of .text and .tail as children
+function fixupTree (node, parent) {
+  if (node.tag !== '#') {
+    if (node.text && !node.tag.match(specialTags)) {
+      let el = et.Element('#')
+      el.text = node.text
+      node._children.unshift(el)
+      delete node.text
+    }
+    if (node.tail) {
+      let el = et.Element('#')
+      el.text = node.tail
+      let pos = parent._children.indexOf(node) + 1
+      parent._children.splice(pos, 0, el)
+      delete node.tail
+    }
+  }
+  node._children.slice(0).forEach((child) => {
+    fixupTree(child, node)
  })
 }

 function addSpansToNode (node, parent, state) {
-  let nodePosition = parent.getchildren().indexOf(node)
+  // text node
+  if (node.tag === '#') {
+    state.segment++
+
+    let sentences = textToSentences(node.text)
+    let pos
+
+    sentences.forEach((sentence) => {
+      let span = createSpan(state.paragraph, state.segment++)
+      span.text = sentence
+
+      // insert the span before the text node
+      pos = parent._children.indexOf(node)
+      parent._children.splice(pos, 0, span)
+    })
+
+    // remove the text node
+    pos = parent._children.indexOf(node)
+    parent._children.splice(pos, 1)
+  }

  if (node.tag.match(paragraphTags)) {
-    state.paragraph++
    state.segment = 0
+    state.paragraph++
  }

-  if (node.tag.match(specialTags)) {
-    const span = createSpan(state.paragraph, state.segment++)
-    span.append(node)
-    parent.getchildren().splice(nodePosition, 1, span)
-  } else {
-    let prependNodes = []
-
-    if (node.text) {
-      prependNodes = textToSpans(node, node.text, state)
-      node.text = null
-    }
-
-    node.getchildren().forEach((child) => {
-      addSpansToNode(child, node, state)
-    })
-
-    prependNodes.forEach((span, i) => {
-      node.getchildren().splice(i, 0, span)
-    })
-  }
-  if (node.tail) {
-    nodePosition = parent.getchildren().indexOf(node)
-    textToSpans(node, node.tail, state).forEach((span, i) => {
-      parent.getchildren().splice(nodePosition + 1 + i, 0, span)
-    })
-    node.tail = null
-  }
+  node.getchildren().slice(0).forEach((child) => {
+    addSpansToNode(child, node, state)
+  })
 }
--- a/test/test-kepubify.js
+++ b/test/test-kepubify.js
@ -6,10 +6,18 @@ const kepubify = require('../src/kepubify').default

 console.log(1, kepubify(`<?xml version="1.0" encoding="utf-8"?>
 <!DOCTYPE html>
-<html><body>text <p>aaaa</p><p>Some text. Woo <!-- or --> not. Here is <img /> another sentence.</p><!-- comment --><p>More text <img/> tail</p> body tail</body> html tail</html>`) === `<?xml version="1.0" encoding="utf-8"?>\n<!DOCTYPE html>\n<html><body>text <div class="book-inner"><div class="book-columns"><p><span class="koboSpan" id="kobo.1.0">aaaa</span></p><p><span class="koboSpan" id="kobo.2.0">Some text.</span><span class="koboSpan" id="kobo.2.1"> Woo not.</span><span class="koboSpan" id="kobo.2.2"> Here is </span><span class="koboSpan" id="kobo.2.4"> another sentence.</span><span class="koboSpan" id="kobo.2.3"><img /></span></p><p><span class="koboSpan" id="kobo.3.0">More text </span><span class="koboSpan" id="kobo.3.2"> tail</span><span class="koboSpan" id="kobo.3.1"><img /></span></p><span class="koboSpan" id="kobo.3.3"> body tail</span></div></div></body></html>`)
+<html><body>text <p>aaaa</p><p>Some text. Woo <!-- or --> not. Here is <img /> another sentence.</p><!-- comment --><p>More text <img/> tail</p> body tail</body> html tail</html>`) === `<?xml version="1.0" encoding="utf-8"?>
+<!DOCTYPE html>
+<html><body>text <div class="book-inner"><div class="book-columns"><p><span class="koboSpan" id="kobo.1.1">aaaa</span></p><p><span class="koboSpan" id="kobo.2.1">Some text.</span><span class="koboSpan" id="kobo.2.2"> Woo not.</span><span class="koboSpan" id="kobo.2.3"> Here is </span><img /><span class="koboSpan" id="kobo.2.5"> another sentence.</span></p><p><span class="koboSpan" id="kobo.3.1">More text </span><img /><span class="koboSpan" id="kobo.3.3"> tail</span></p><span class="koboSpan" id="kobo.3.5"> body tail</span></div></div></body></html>`)

 console.log(2, kepubify(`<?xml version="1.0" encoding="utf-8"?>
 <!DOCTYPE html>
 <html><body><p>Dated: June 5th. Wohoo</p></body></html>`) === `<?xml version="1.0" encoding="utf-8"?>
 <!DOCTYPE html>
-<html><body><div class="book-inner"><div class="book-columns"><p><span class="koboSpan" id="kobo.1.0">Dated:</span><span class="koboSpan" id="kobo.1.1"> June 5th.</span><span class="koboSpan" id="kobo.1.2"> Wohoo</span></p></div></div></body></html>`)
+<html><body><div class="book-inner"><div class="book-columns"><p><span class="koboSpan" id="kobo.1.1">Dated:</span><span class="koboSpan" id="kobo.1.2"> June 5th.</span><span class="koboSpan" id="kobo.1.3"> Wohoo</span></p></div></div></body></html>`)
+
+console.log(3, kepubify(`<?xml version="1.0" encoding="utf-8"?>
+<!DOCTYPE html>
+<html><body><pre>hello</pre><p>“Well, you know, Water Bearer and all,” she laughed weakly. “Don’t kid yourself, though. You’re <i>strong,</i> Daphne, and I’m not just saying that. You saved us, and you’ve come all this way through all these trials. If you get a little heartbroken now and again, well, you’re entitled to it. You’re like your mother in that.” She smiled brightly. “The Seer told me a bit about that, you know. The whole… thing passed down the female line, on and on for ages, all of them determined and gifted. <i>Apparently</i> that’s why things here resemble stuff on our earth so well—because the Everfree Ways have been following your maternal line all across Europe and the Americas, and apparently a brief stop at New Zealand. They were super adventurous.”</p></body></html>`) === `<?xml version="1.0" encoding="utf-8"?>
+<!DOCTYPE html>
+<html><body><div class="book-inner"><div class="book-columns"><pre>hello</pre><p><span class="koboSpan" id="kobo.1.1">“Well, you know, Water Bearer and all,” she laughed weakly.</span><span class="koboSpan" id="kobo.1.2"> “Don’t kid yourself, though.</span><span class="koboSpan" id="kobo.1.3"> You’re </span><i><span class="koboSpan" id="kobo.1.5">strong,</span></i><span class="koboSpan" id="kobo.1.7"> Daphne, and I’m not just saying that.</span><span class="koboSpan" id="kobo.1.8"> You saved us, and you’ve come all this way through all these trials.</span><span class="koboSpan" id="kobo.1.9"> If you get a little heartbroken now and again, well, you’re entitled to it.</span><span class="koboSpan" id="kobo.1.10"> You’re like your mother in that.”</span><span class="koboSpan" id="kobo.1.11"> She smiled brightly.</span><span class="koboSpan" id="kobo.1.12"> “The Seer told me a bit about that, you know.</span><span class="koboSpan" id="kobo.1.13"> The whole… thing passed down the female line, on and on for ages, all of them determined and gifted. </span><i><span class="koboSpan" id="kobo.1.15">Apparently</span></i><span class="koboSpan" id="kobo.1.17"> that’s why things here resemble stuff on our earth so well—because the Everfree Ways have been following your maternal line all across Europe and the Americas, and apparently a brief stop at New Zealand.</span><span class="koboSpan" id="kobo.1.18"> They were super adventurous.”</span></p></div></div></body></html>`)