Merge pull request #11 from wscherphof/issue-#7---microdata

Issue #7   microdata

htmlparser.lua

@@ -16,14 +16,18 @@ local function parse(text)
 
     local tagst, apos = tag:gettext(), 1
     while true do
-      local start, k, quote, v
-      start, apos, k, quote = string.find(tagst, "%s+([^%s=]+)=(['\"]?)", apos)
+      local start, k, eq, quote, v
+      start, apos, k, eq, quote = string.find(tagst, "%s+([^%s=]+)(=?)(['\"]?)", apos)
       if not k then break end
-      local pattern = "=([^%s'\">]*)"
-      if quote ~= '' then
-        pattern = quote .. "([^" .. quote .. "]*)" .. quote
+      if eq == "" then
+        v = ""
+      else
+        local pattern = "=([^%s'\">]*)"
+        if quote ~= '' then
+          pattern = quote .. "([^" .. quote .. "]*)" .. quote
+        end
+        start, apos, v = string.find(tagst, pattern, apos)
       end
-      start, apos, v = string.find(tagst, pattern, apos)
       tag:addattribute(k, v)
     end
 
@@ -31,17 +35,16 @@ local function parse(text)
       descend = false
       tag:close()
     else
-      opentags[tag.name] = tag
+      opentags[tag.name] = opentags[tag.name] or {}
+      table.insert(opentags[tag.name], tag)
     end
 
     local closeend = tpos
     while true do
       local closestart, closing, closename
       closestart, closeend, closing, closename = string.find(root._text, "[^<]*<(/?)(%w+)", closeend)
-      closing = closing and closing ~= ''
-      if not closing then break end
-      tag = opentags[closename]
-      opentags[closename] = nil
+      if not closing or closing == "" then break end
+      tag = table.remove(opentags[closename])
       closestart = string.find(root._text, "<", closestart)
       tag:close(closestart, closeend + 1)
       node = tag.parent

test.html

@@ -39,5 +39,24 @@
       </a>
     </li>
   </ul>
+  <h1>About me</h1>
+  <section itemscope itemtype="http://schema.org/Person"> 
+    Hello, my name is 
+    <span itemprop="name">John Doe</span>,
+    I am a 
+    <span itemprop="jobTitle">graduate research assistant</span> 
+    at the 
+    <span itemprop="affiliation">University of Dreams</span>. 
+    My friends call me 
+    <span itemprop="additionalName">Johnny</span>. 
+    You can visit my homepage at 
+    <a href="http://www.JohnnyD.com" itemprop="url">www.JohnnyD.com</a>. 
+    <section itemprop="address" itemscope itemtype="http://schema.org/PostalAddress">
+      I live at 
+      <span itemprop="streetAddress">1234 Peach Drive</span>,
+      <span itemprop="addressLocality">Warner Robins</span>,
+      <span itemprop="addressRegion">Georgia</span>.
+    </section>
+  </section>
 </body>
 </html>

test.lua

@@ -51,22 +51,24 @@ select("ul > *")
 select("body [class]")
 select("body > [class]")
 
+print("\nchapters")
 local sel, chapters = root("ol.chapters > li"), {}
 for _,v in ipairs(sel.nodes) do
   table.insert(chapters, v:getcontent())
 end
-print("\nchapters")
+-- print
 for i,v in ipairs(chapters) do
   print(i, v)
 end
 
+print("\ncontacts")
 local sel, contacts = root("ul.contacts > li")("span[class]"), {}
 for _,v in ipairs(sel.nodes) do
   local id = v.parent.parent.id -- li > a > span
   contacts[id] = contacts[id] or {}
   contacts[id][v.classes[1]] = v:getcontent()
 end
-print("\ncontacts")
+-- print
 for k,v in pairs(contacts) do
   print(k)
   for fk,fv in pairs(v) do
@@ -74,5 +76,41 @@ for k,v in pairs(contacts) do
   end
 end
 
-
-
+print("\nmicrodata")
+local sel, scopes = root("[itemprop]"), {}
+for _,prop in ipairs(sel.nodes) do
+  if prop.attributes["itemscope"] then goto nextprop end
+  local descendantscopes, scope = {}, prop
+  while true do
+    repeat
+      scope = scope.parent
+    until scope.attributes["itemscope"]
+    if not scope.attributes["itemprop"] then break end
+    table.insert(descendantscopes, 1, scope)
+  end
+  scopes[scope] = scopes[scope] or {}
+  local entry = scopes[scope]
+  for _,v in ipairs(descendantscopes) do
+    entry[v] = entry[v] or {}
+    entry = entry[v]
+  end
+  local k, v = prop.attributes["itemprop"], prop:getcontent()
+  entry[k] = v
+  ::nextprop::
+end
+-- print
+local function printscope(node, table, level)
+  level = level or 1
+  local scopeprop = node.attributes["itemprop"] or ""
+  print(string.rep("  ", level - 1) .. node.attributes["itemtype"], scopeprop)
+  for prop,v in pairs(table) do
+    if type(prop) == "table" then
+      printscope(prop, v, level + 1)
+    else
+      print(string.rep("  ", level) .. prop .. "=[" .. v .. "]")
+    end
+  end
+end
+for node,table in pairs(scopes) do
+  printscope(node, table)
+end