recognize pattern of h-entry + h-card

* a single h-entry and h-card, where the h-entry has no URL, will result in a permalink page with that h-entry
* multiple h-entrys followed by an h-card is a feed

lib/XRay/Formats/Mf2.php

@@ -22,8 +22,26 @@ class Mf2 extends Format {
       return self::parseAsHFeed($mf2, $http, $url);
     }
 
-    // If there is only one item on the page, just use that
-    if(count($mf2['items']) == 1) {
+    // Remove h-breadcrumb since we never use it and it causes problems determining
+    // whether a page is a feed or permalink
+    $mf2['items'] = array_values(array_filter($mf2['items'], function($item){
+      return !in_array('h-breadcrumb', $item['type']);
+    }));
+
+    $items = $mf2['items'];
+
+    // If there is more than one item on the page, it may be a feed.
+    // Remove an h-card if there is one that doesn't match the page URL, then try again.
+    // (Don't modify the actual tree, but compare on the modified tree)
+    if(count($items) > 1) {
+      $tmpmf2 = array_filter($items, function($item) use($url){
+        return !(in_array('h-card', $item['type']) && isset($item['properties']['url'][0]) && $item['properties']['url'][0] != $url);
+      });
+      $items = array_values($tmpmf2);
+    }
+
+    // If there is only one item left on the page, it's a permalink, and just use that
+    if(count($items) == 1) {
       $item = $mf2['items'][0];
       if(in_array('h-entry', $item['type']) || in_array('h-cite', $item['type'])) {
         #Parse::debug("mf2:0: Recognized $url as an h-entry it is the only item on the page");
@@ -130,18 +148,18 @@ class Mf2 extends Format {
       }
     }
 
-    // If there was more than one h-entry on the page, treat the whole page as a feed
-    if(count($mf2['items']) > 1) {
-      if(count(array_filter($mf2['items'], function($item){
+    // At this point, if there are any h-entrys left on the page, it's probably a feed.
+    if(count($items) > 0) {
+      if(count(array_filter($items, function($item){
         return in_array('h-entry', $item['type']);
-      })) > 1) {
+      })) > 0) {
         #Parse::debug("mf2:2: Recognized $url as an h-feed because there are more than one object on the page");
         return self::parseAsHFeed($mf2, $http, $url);
       }
     }
 
     // If the first item is an h-feed, parse as a feed
-    $first = $mf2['items'][0];
+    $first = $items[0];
     if(in_array('h-feed', $first['type'])) {
       #Parse::debug("mf2:3: Recognized $url as an h-feed because the first item is an h-feed");
       return self::parseAsHFeed($mf2, $http, $url);

tests/FeedTest.php

@@ -61,6 +61,48 @@ class FeedTest extends PHPUnit_Framework_TestCase {
     $this->assertEquals('Author Name', $data->items[3]->author->name);
   }
 
+  public function testListOfHEntrysWithHCardNoExpect() {
+    $url = 'http://feed.example.com/list-of-hentrys-with-h-card';
+    $response = $this->parse(['url' => $url]);
+
+    $body = $response->getContent();
+    $this->assertEquals(200, $response->getStatusCode());
+    $result = json_decode($body);
+    $this->assertEquals('mf2+html', $result->{'source-format'});
+    $data = $result->data;
+
+    $this->assertEquals('feed', $data->type);
+    $this->assertEquals(4, count($data->items));
+    $this->assertEquals('One', $data->items[0]->name);
+    $this->assertEquals('article', $data->items[0]->{'post-type'});
+    $this->assertEquals('Two', $data->items[1]->name);
+    $this->assertEquals('Three', $data->items[2]->name);
+    $this->assertEquals('Four', $data->items[3]->name);
+
+    // Check that the author h-card was matched up with each h-entry
+    $this->assertEquals('Author Name', $data->items[0]->author->name);
+    $this->assertEquals('Author Name', $data->items[1]->author->name);
+    $this->assertEquals('Author Name', $data->items[2]->author->name);
+    $this->assertEquals('Author Name', $data->items[3]->author->name);
+  }
+
+  public function testShortListOfHEntrysWithHCardNoExpect() {
+    $url = 'http://feed.example.com/short-list-of-hentrys-with-h-card';
+    $response = $this->parse(['url' => $url]);
+
+    $body = $response->getContent();
+    $this->assertEquals(200, $response->getStatusCode());
+    $result = json_decode($body);
+    $this->assertEquals('mf2+html', $result->{'source-format'});
+    $data = $result->data;
+
+    // In this case, this looks like a page permalink
+    $this->assertEquals('entry', $data->type);
+    // This test should find the h-entry rather than the h-card, because the h-card does not contain the page URL
+    $this->assertEquals('http://feed.example.com/1', $data->url);
+    $this->assertEquals('Author', $data->author->name);
+  }
+
   public function testShortListOfHEntrysWithHCard() {
     $url = 'http://feed.example.com/short-list-of-hentrys-with-h-card';
     $response = $this->parse(['url' => $url, 'expect' => 'feed']);

tests/data/source.example.com/rel-alternate-mf2-json

@@ -256,7 +256,7 @@ Connection: keep-alive
       #<a href="/tag/indieauth" class="p-category">indieauth</a>
   </div>
     <div class="metaline pad">
-      <a href="https://aaronparecki.com/2018/07/12/10/indieauth" class="u-url">
+      <a href="http://source.example.com/rel-alternate-mf2-json" class="u-url">
       <time class="dt-published" datetime="2018-07-12T13:02:04-07:00">
         Thu, Jul 12, 2018 1:02pm -07:00
       </time>

tests/data/source.example.com/rel-alternate-mf2-json.json

@@ -18,7 +18,7 @@ Connection: keep-alive
                     "indieauth"
                 ],
                 "url": [
-                    "https://aaronparecki.com/2018/07/12/10/indieauth"
+                    "http://source.example.com/rel-alternate-mf2-json"
                 ],
                 "syndication": [
                     "https://twitter.com/aaronpk/status/1017500609631567872"