Parsing empty parts

3 files changed, 115 insertions, 15 deletions

diff --git a/packages/gitbook-html/lib/summary.js b/packages/gitbook-html/lib/summary.js
index e55719b..0a9ddb5 100755
--- a/packages/gitbook-html/lib/summary.js
+++ b/packages/gitbook-html/lib/summary.js
@@ -3,6 +3,8 @@ var dom = require('./dom');
 
 var SELECTOR_LIST = 'ol, ul';
 var SELECTOR_LINK = '> a, p > a';
+var SELECTOR_PART = 'h2, h3, h4';
+
 var BL = '\n';
 
 /**
@@ -15,7 +17,7 @@ function findList($parent) {
     var $container = $parent.children('.olist');
     if ($container.length > 0) $parent = $container.first();
 
-    return $parent.children('ul, ol');
+    return $parent.children(SELECTOR_LIST);
 }
 
 /**
@@ -40,7 +42,7 @@ function parseList($ul, $) {
         var $a = $li.find(SELECTOR_LINK);
         if ($a.length > 0) {
             article.title = $a.first().text();
-            article.ref = $a.attr('href').replace(/\\/g, '/').replace(/^\/+/, '')
+            article.ref = $a.attr('href').replace(/\\/g, '/').replace(/^\/+/, '');
         }
 
         // Sub articles
@@ -55,6 +57,71 @@ function parseList($ul, $) {
 }
 
 /**
+     Find all parts and their corresponding lists
+
+     @param {cheerio.Node}
+     @param {cheerio.DOM}
+     @return {Array<{title: String, list: cheerio.Node}>}
+ */
+function findParts($parent, $) {
+    // Find parts and lists
+    // TODO asciidoc compatibility
+    var partsAndLists = $parent.children(SELECTOR_LIST + ', ' + SELECTOR_PART);
+
+    // Group each part with the list after
+    var parts = [];
+    var previousPart = null;
+
+    partsAndLists.each(function (i, el) {
+        if (isPartNode(el)) {
+            if (previousPart !== null) {
+                // The previous part was empty
+                parts.push(previousPart);
+            }
+            previousPart = {
+                title: getPartTitle(el, $),
+                list: null
+            };
+
+        } else { // It is a list
+            if (previousPart !== null) {
+                previousPart.list = el;
+            } else {
+                previousPart = {
+                    title: '',
+                    list: el
+                };
+            }
+            parts.push(previousPart);
+            previousPart = null;
+        }
+    });
+
+    return parts;
+}
+
+/**
+    True if the element is a part
+
+    @param el
+    @return {Boolean}
+ */
+function isPartNode(el) {
+    return SELECTOR_PART.indexOf(el.name) !== -1;
+}
+
+/**
+    Parse the title of a part element
+
+    @param el
+    @param {cheerio.DOM} $
+    @return {String}
+ */
+function getPartTitle(el, $) {
+    return $(el).text().trim();
+}
+
+/**
     Parse an HTML content into a tree of articles/parts
 
     @param {String} html
@@ -64,21 +131,22 @@ function parseSummary(html) {
     var $ = dom.parse(html);
     var $root = dom.cleanup(dom.root($), $);
 
-    var $lists = findList($root);
-    var parts = [];
-
-    $lists.each(function() {
-        var $list = $(this);
-        var $title = $list.prevUntil(SELECTOR_LIST, 'h2, h3, h4').first();
-
-        parts.push({
-            title: $title.text().trim(),
-            articles: parseList($list, $)
+    var parts = findParts($root, $);
+
+    // Parse each list
+    debugger;
+    var parsedParts = [];
+    var part;
+    for (var i = 0; i < parts.length; ++i) {
+        part = parts[i];
+        parsedParts.push({
+            title: part.title,
+            articles: parseList($(part.list), $)
         });
-    });
+    }
 
     return {
-        parts: parts
+        parts: parsedParts
     };
 }
 
diff --git a/packages/gitbook-html/test/fixtures/SUMMARY-EMPTY.html b/packages/gitbook-html/test/fixtures/SUMMARY-EMPTY.html
new file mode 100644
index 0000000..03579a1
--- /dev/null
+++ b/packages/gitbook-html/test/fixtures/SUMMARY-EMPTY.html
@@ -0,0 +1,21 @@
+<h1>Summary</h1>
+
+<h2>Part 1</h2>
+
+<ul>
+    <li><a href="chapter-1/README.md">Chapter 1</a></li>
+    <li><a href="chapter-2/README.md">Chapter 2</a></li>
+    <li><a href="chapter-3/README.md">Chapter 3</a></li>
+</ul>
+
+<ul>
+    <li><a href="chapter-1/README.md">Chapter for untitled part</a></li>
+</ul>
+
+<h2>Empty part</h2>
+
+<h2>Part 2</h2>
+
+<ul>
+    <li><a href="chapter-1/README.md">Chapter for Part 2</a></li>
+</ul>
diff --git a/packages/gitbook-html/test/summary.js b/packages/gitbook-html/test/summary.js
index 24d22c0..9c064f9 100755
--- a/packages/gitbook-html/test/summary.js
+++ b/packages/gitbook-html/test/summary.js
@@ -8,9 +8,11 @@ describe('Summary parsing', function () {
     var LEXED, PART;
 
     before(function() {
-        var CONTENT = fs.readFileSync(path.join(__dirname, './fixtures/SUMMARY.html'), 'utf8');
+        var CONTENT = fs.readFileSync(
+            path.join(__dirname, './fixtures/SUMMARY.html'), 'utf8');
         LEXED = summary(CONTENT);
         PART = LEXED.parts[0];
+
     });
 
     describe('Parts', function() {
@@ -23,6 +25,15 @@ describe('Summary parsing', function () {
             assert.equal(LEXED.parts[1].title, 'Part 2');
             assert.equal(LEXED.parts[2].title, '');
         });
+
+        it('should detect empty parts', function() {
+            var CONTENT_EMPTY = fs.readFileSync(
+                path.join(__dirname, './fixtures/SUMMARY-EMPTY.html'), 'utf8');
+            var LEXED_EMPTY = summary(CONTENT_EMPTY);
+
+            assert.equal(LEXED_EMPTY.parts.length, 4);
+            assert.equal(LEXED_EMPTY.parts[2].title, 'Empty part');
+        });
     });
 
     it('should detect chapters', function() {