Feed summaries

lib/Parser/XML/Construct.php

@@ -15,6 +15,10 @@ use MensBeam\Lax\Url;
 abstract class Construct {
     use \MensBeam\Lax\Parser\Construct;
 
+    protected const TEXT_LOOSE = "loose";
+    protected const TEXT_PLAIN = "plain";
+    protected const TEXT_HTML = "html";
+
     /** @var \DOMDocument */
     protected $document;
     /** @var \DOMXPath */
@@ -110,8 +114,6 @@ abstract class Construct {
                     $data = $this->trimText($data);
                 } elseif ($format === "html" || $format === "loose") {
                     $out->htmlBase = strlen($node->baseURI) ? $node->baseURI : null;
-                } elseif ($format === "xhtml") { // @codeCoverageIgnore
-                    $out->xhtmlBase = strlen($node->baseURI) ? $node->baseURI : null; // @codeCoverageIgnore
                 }
                 $out->$format = $data;
                 return $out;
@@ -289,19 +291,19 @@ abstract class Construct {
     }
 
     protected function getTitleRss1(): ?Text {
-        return $this->fetchText("rss1:title|rss0:title", "loose");
+        return $this->fetchText("rss1:title|rss0:title", self::TEXT_LOOSE);
     }
 
     protected function getTitleRss2(): ?Text {
-        return $this->fetchText("title", "loose");
+        return $this->fetchText("title", self::TEXT_LOOSE);
     }
 
     protected function getTitleDC(): ?Text {
-        return $this->fetchText("dc:title", "plain");
+        return $this->fetchText("dc:title", self::TEXT_PLAIN);
     }
 
     protected function getTitlePod(): ?Text {
-        return $this->fetchText("apple:title", "plain");
+        return $this->fetchText("apple:title", self::TEXT_PLAIN);
     }
     
 }

lib/Parser/XML/Feed.php

@@ -87,7 +87,7 @@ class Feed extends Construct implements \MensBeam\Lax\Parser\Feed {
         $feed->url = $this->getUrl();
         $feed->link = $this->getLink();
         $feed->title = $this->getTitle();
-        //$feed->summary = $this->getSummary();
+        $feed->summary = $this->getSummary();
         //$feed->dateModified = $this->getDateModified();
         //$feed->icon = $this->getIcon();
         //$feed->image = $this->getImage();
@@ -132,8 +132,7 @@ class Feed extends Construct implements \MensBeam\Lax\Parser\Feed {
     }
 
     public function getSummary(): ?Text {
-        // unlike most other data, Atom is not preferred, because Atom doesn't really have feed summaries
-        return $this->getSummaryDC() ?? $this->getSummaryRss1() ?? $this->getSummaryRss2() ?? $this->getSummaryPod() ?? $this->getSummaryAtom();
+        return $this->getSummaryAtom() ?? $this->getSummaryDC() ?? $this->getSummaryRss1() ?? $this->getSummaryRss2() ?? $this->getSummaryPod();
     }
 
     public function getCategories(): CategoryCollection {
@@ -266,4 +265,24 @@ class Feed extends Construct implements \MensBeam\Lax\Parser\Feed {
     protected function getUrlPod(): ?Url {
         return $this->fetchUrl("apple:new-feed-url");
     }
+
+    protected function getSummaryAtom(): ?Text {
+        return $this->fetchAtomText("atom:summary") ?? $this->fetchAtomText("atom:subtitle"); 
+    }
+
+    protected function getSummaryRss2(): ?Text {
+        return $this->fetchText("description", self::TEXT_LOOSE);
+    }
+
+    protected function getSummaryRss1(): ?Text {
+        return $this->fetchText("rss1:description|rss0:description", self::TEXT_LOOSE);
+    }
+
+    protected function getSummaryDC(): ?Text {
+        return $this->fetchText("dc:description", self::TEXT_PLAIN);
+    }
+
+    protected function getSummaryPod(): ?Text {
+        return $this->fetchText("apple:summary|gplay:description", self::TEXT_PLAIN) ?? $this->fetchText("apple:subtitle", self::TEXT_PLAIN);
+    }
 }

tests/cases/XML/feed-atom.yaml

@@ -28,7 +28,6 @@ Atom ID with whitespace:
         id: 'http://example.com/'
 
 
-
 Bogus ID before good:
     input: >
         <feed xmlns="http://www.w3.org/2005/Atom">
@@ -310,6 +309,30 @@ Feed title in multiple formats:
             xhtml: '<div xmlns="http://www.w3.org/1999/xhtml"> There  &amp; Then   </div>'
             xhtmlBase: 'http://example.com/'
 
+Feed summary 1:
+    input: >
+        <feed xmlns="http://www.w3.org/2005/Atom">
+            <subtitle type="html">Subtitle text</subtitle>
+            <summary type="html">Summary text</summary>
+        </feed>
+    output:
+        format: atom
+        version: '1.0'
+        summary:
+            html: 'Summary text'
+
+Feed summary 2:
+    input: >
+        <feed xmlns="http://www.w3.org/2005/Atom">
+            <subtitle type="html">Subtitle text</subtitle>
+        </feed>
+    output:
+        format: atom
+        version: '1.0'
+        summary:
+            html: 'Subtitle text'
+
+
 Ignored text constructs:
     input: >
         <feed xmlns="http://www.w3.org/2005/Atom">

tests/cases/XML/feed-other.yaml

@@ -59,3 +59,27 @@ iPod title:
         format: rss
         title:
             plain: 'Plain text'
+
+iPod summary: # Apple's own documentation doesn't mention a namespaced summary element, but Google's does
+    input: >
+        <rss xmlns:itunes="http://www.itunes.com/dtds/podcast-1.0.dtd">
+            <channel>
+                <itunes:summary> Plain  text</itunes:summary>
+            </channel>
+        </rss>
+    output:
+        format: rss
+        summary:
+            plain: 'Plain text'
+
+Google Play summary:
+    input: >
+        <rss xmlns:play="http://www.google.com/schemas/play-podcasts/1.0">
+            <channel>
+                <play:description> Plain  text</play:description>
+            </channel>
+        </rss>
+    output:
+        format: rss
+        summary:
+            plain: 'Plain text'

tests/cases/XML/feed-rss0.yaml

@@ -95,3 +95,16 @@ Feed title 2:
         title:
             loose: 'Loose  text'
             htmlBase: 'https://example.com/'
+
+Feed summary:
+    input: >
+        <rdf:RDF xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#" xmlns="http://channel.netscape.com/rdf/simple/0.9/">
+            <channel>
+              <description>Loose text</description>
+            </channel>
+        </rdf:RDF>
+    output:
+        format: rdf
+        version: '0.90'
+        summary:
+            loose: 'Loose text'

tests/cases/XML/feed-rss1.yaml

@@ -243,3 +243,29 @@ DC title:
         version: '1.0'
         title:
             plain: 'Plain text'
+
+Feed summary:
+    input: >
+        <rdf:RDF xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#" xmlns="http://purl.org/rss/1.0/">
+            <channel>
+              <description>Loose text</description>
+            </channel>
+        </rdf:RDF>
+    output:
+        format: rdf
+        version: '1.0'
+        summary:
+            loose: 'Loose text'
+
+DC summary:
+    input: >
+        <rdf:RDF xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#" xmlns="http://purl.org/rss/1.0/" xmlns:dc="http://purl.org/dc/elements/1.1/">
+            <channel>
+              <dc:description>Plain text</dc:description>
+            </channel>
+        </rdf:RDF>
+    output:
+        format: rdf
+        version: '1.0'
+        summary:
+            plain: 'Plain text'

tests/cases/XML/feed-rss2.yaml

@@ -320,3 +320,13 @@ Feed title 2:
         title:
             loose: 'Loose  text'
             htmlBase: 'https://example.com/'
+
+Feed summary:
+    input: >
+        <rss><channel>
+            <description>Loose text</description>
+        </channel></rss>
+    output:
+        format: rss
+        summary:
+            loose: 'Loose text'