summaryrefslogtreecommitdiff
path: root/test
diff options
context:
space:
mode:
authorMaria Luiza Soares <[email protected]>2018-12-21 14:47:40 +0100
committerMalu Decks <[email protected]>2018-12-23 05:37:44 +0100
commitb63c65bdf86c241ae4e1dbe8978bcc341a1742fd (patch)
treeb7d5f3b3d797e79af678b4d722fd80e4fdcddebf /test
parent01f5f7c94e4c99643ed41ba26db6efcc018d4b07 (diff)
Update all tests to assert on getSiteName outcome
Diffstat (limited to 'test')
-rw-r--r--test/test-pages/001/expected-metadata.json5
-rw-r--r--test/test-pages/002/expected-metadata.json5
-rw-r--r--test/test-pages/003-metadata-preferred/expected-metadata.json5
-rw-r--r--test/test-pages/004-metadata-space-separated-properties/expected-metadata.json5
-rw-r--r--test/test-pages/aclu/expected-metadata.json5
-rw-r--r--test/test-pages/ars-1/expected-metadata.json5
-rw-r--r--test/test-pages/base-url-base-element-relative/expected-metadata.json5
-rw-r--r--test/test-pages/base-url-base-element/expected-metadata.json5
-rw-r--r--test/test-pages/base-url/expected-metadata.json5
-rw-r--r--test/test-pages/basic-tags-cleaning/expected-metadata.json5
-rw-r--r--test/test-pages/bbc-1/expected-metadata.json5
-rw-r--r--test/test-pages/blogger/expected-metadata.json5
-rw-r--r--test/test-pages/breitbart/expected-metadata.json5
-rw-r--r--test/test-pages/bug-1255978/expected-metadata.json5
-rw-r--r--test/test-pages/buzzfeed-1/expected-metadata.json5
-rw-r--r--test/test-pages/challenges/expected-metadata.json5
-rw-r--r--test/test-pages/clean-links/expected-metadata.json5
-rw-r--r--test/test-pages/cnet-svg-classes/expected-metadata.json5
-rw-r--r--test/test-pages/cnet/expected-metadata.json5
-rw-r--r--test/test-pages/cnn/expected-metadata.json5
-rw-r--r--test/test-pages/comment-inside-script-parsing/expected-metadata.json5
-rw-r--r--test/test-pages/daringfireball-1/expected-metadata.json5
-rw-r--r--test/test-pages/different-sources-for-images/expected-metadata.json5
-rw-r--r--test/test-pages/ehow-1/expected-metadata.json5
-rw-r--r--test/test-pages/ehow-2/expected-metadata.json5
-rw-r--r--test/test-pages/embedded-videos/expected-metadata.json5
-rw-r--r--test/test-pages/engadget/expected-metadata.json5
-rw-r--r--test/test-pages/fortressofsolitude/expected-metadata.json5
-rw-r--r--test/test-pages/gmw/expected-metadata.json5
-rw-r--r--test/test-pages/heise/expected-metadata.json5
-rw-r--r--test/test-pages/herald-sun-1/expected-metadata.json5
-rw-r--r--test/test-pages/hidden-nodes/expected-metadata.json5
-rw-r--r--test/test-pages/hukumusume/expected-metadata.json5
-rw-r--r--test/test-pages/iab-1/expected-metadata.json5
-rw-r--r--test/test-pages/ietf-1/expected-metadata.json5
-rw-r--r--test/test-pages/infobae/expected-metadata.json5
-rw-r--r--test/test-pages/keep-images/expected-metadata.json5
-rw-r--r--test/test-pages/la-nacion/expected-metadata.json5
-rw-r--r--test/test-pages/lemonde-1/expected-metadata.json5
-rw-r--r--test/test-pages/lemonde-2/expected-metadata.json5
-rw-r--r--test/test-pages/liberation-1/expected-metadata.json5
-rw-r--r--test/test-pages/lifehacker-post-comment-load/expected-metadata.json5
-rw-r--r--test/test-pages/lifehacker-working/expected-metadata.json5
-rw-r--r--test/test-pages/links-in-tables/expected-metadata.json5
-rw-r--r--test/test-pages/lwn-1/expected-metadata.json5
-rw-r--r--test/test-pages/medium-1/expected-metadata.json5
-rw-r--r--test/test-pages/medium-2/expected-metadata.json5
-rw-r--r--test/test-pages/medium-3/expected-metadata.json5
-rw-r--r--test/test-pages/missing-paragraphs/expected-metadata.json5
-rw-r--r--test/test-pages/mozilla-1/expected-metadata.json5
-rw-r--r--test/test-pages/mozilla-2/expected-metadata.json5
-rw-r--r--test/test-pages/msn/expected-metadata.json5
-rw-r--r--test/test-pages/nytimes-1/expected-metadata.json5
-rw-r--r--test/test-pages/nytimes-2/expected-metadata.json5
-rw-r--r--test/test-pages/pixnet/expected-metadata.json5
-rw-r--r--test/test-pages/remove-extra-brs/expected-metadata.json5
-rw-r--r--test/test-pages/remove-extra-paragraphs/expected-metadata.json5
-rw-r--r--test/test-pages/remove-script-tags/expected-metadata.json5
-rw-r--r--test/test-pages/reordering-paragraphs/expected-metadata.json5
-rw-r--r--test/test-pages/replace-brs/expected-metadata.json5
-rw-r--r--test/test-pages/replace-font-tags/expected-metadata.json5
-rw-r--r--test/test-pages/rtl-1/expected-metadata.json5
-rw-r--r--test/test-pages/rtl-2/expected-metadata.json5
-rw-r--r--test/test-pages/rtl-3/expected-metadata.json5
-rw-r--r--test/test-pages/rtl-4/expected-metadata.json5
-rw-r--r--test/test-pages/salon-1/expected-metadata.json5
-rw-r--r--test/test-pages/simplyfound-1/expected-metadata.json5
-rw-r--r--test/test-pages/social-buttons/expected-metadata.json5
-rw-r--r--test/test-pages/style-tags-removal/expected-metadata.json5
-rw-r--r--test/test-pages/svg-parsing/expected-metadata.json5
-rw-r--r--test/test-pages/table-style-attributes/expected-metadata.json5
-rw-r--r--test/test-pages/telegraph/expected-metadata.json5
-rw-r--r--test/test-pages/title-and-h1-discrepancy/expected-metadata.json5
-rw-r--r--test/test-pages/tmz-1/expected-metadata.json5
-rw-r--r--test/test-pages/tumblr/expected-metadata.json5
-rw-r--r--test/test-pages/wapo-1/expected-metadata.json5
-rw-r--r--test/test-pages/wapo-2/expected-metadata.json5
-rw-r--r--test/test-pages/webmd-1/expected-metadata.json5
-rw-r--r--test/test-pages/webmd-2/expected-metadata.json5
-rw-r--r--test/test-pages/wikia/expected-metadata.json5
-rw-r--r--test/test-pages/wikipedia/expected-metadata.json5
-rw-r--r--test/test-pages/wordpress/expected-metadata.json5
-rw-r--r--test/test-pages/yahoo-1/expected-metadata.json5
-rw-r--r--test/test-pages/yahoo-2/expected-metadata.json5
-rw-r--r--test/test-pages/yahoo-3/expected-metadata.json5
-rw-r--r--test/test-pages/yahoo-4/expected-metadata.json5
-rw-r--r--test/test-pages/youth/expected-metadata.json5
87 files changed, 261 insertions, 174 deletions
diff --git a/test/test-pages/001/expected-metadata.json b/test/test-pages/001/expected-metadata.json
index 2688eda..52dafa1 100644
--- a/test/test-pages/001/expected-metadata.json
+++ b/test/test-pages/001/expected-metadata.json
@@ -3,5 +3,6 @@
"Direction": null,
"Excerpt": "Nicolas Perriault's homepage.",
"Image": null,
- "Title": "Get your Frontend JavaScript Code Covered | Code"
-} \ No newline at end of file
+ "Title": "Get your Frontend JavaScript Code Covered | Code",
+ "SiteName": null
+}
diff --git a/test/test-pages/002/expected-metadata.json b/test/test-pages/002/expected-metadata.json
index b0a0a4d..74c3859 100644
--- a/test/test-pages/002/expected-metadata.json
+++ b/test/test-pages/002/expected-metadata.json
@@ -3,5 +3,6 @@
"Direction": null,
"Excerpt": "For more than a decade the Web has used XMLHttpRequest (XHR) to achieve asynchronous requests in JavaScript. While very useful, XHR is not a very ...",
"Image": null,
- "Title": "This API is so Fetching!"
-} \ No newline at end of file
+ "Title": "This API is so Fetching!",
+ "SiteName": "Mozilla Hacks – the Web developer blog"
+}
diff --git a/test/test-pages/003-metadata-preferred/expected-metadata.json b/test/test-pages/003-metadata-preferred/expected-metadata.json
index c4d29fe..7c31c22 100644
--- a/test/test-pages/003-metadata-preferred/expected-metadata.json
+++ b/test/test-pages/003-metadata-preferred/expected-metadata.json
@@ -3,5 +3,6 @@
"Direction": null,
"Excerpt": "Dublin Core property description",
"Image": null,
- "Title": "Dublin Core property title"
-} \ No newline at end of file
+ "Title": "Dublin Core property title",
+ "SiteName": null
+}
diff --git a/test/test-pages/004-metadata-space-separated-properties/expected-metadata.json b/test/test-pages/004-metadata-space-separated-properties/expected-metadata.json
index 25d52a6..a30d097 100644
--- a/test/test-pages/004-metadata-space-separated-properties/expected-metadata.json
+++ b/test/test-pages/004-metadata-space-separated-properties/expected-metadata.json
@@ -3,5 +3,6 @@
"Direction": null,
"Excerpt": "Preferred description",
"Image": "http:\/\/fakehost.com\/image.jpg",
- "Title": "Preferred title"
-} \ No newline at end of file
+ "Title": "Preferred title",
+ "SiteName": null
+}
diff --git a/test/test-pages/aclu/expected-metadata.json b/test/test-pages/aclu/expected-metadata.json
index 30cb9a3..9b20703 100644
--- a/test/test-pages/aclu/expected-metadata.json
+++ b/test/test-pages/aclu/expected-metadata.json
@@ -3,5 +3,6 @@
"Direction": "ltr",
"Excerpt": "I don't use Facebook. I'm not technophobic — I'm a geek. I've been using email since the early 1990s, I have accounts on hundreds of services around the net, and I do software development and internet protocol design both for work and for fun. I believe that a globe-spanning communications network like the internet can be a positive social force, and I publish much of my own work on the open web.",
"Image": "https:\/\/www.aclu.org\/sites\/default\/files\/styles\/metatag_og_image_1200x630\/public\/field_share_image\/web18-facebook-socialshare-1200x628-v02.png?itok=p77cQjOm",
- "Title": "Facebook Is Tracking Me Even Though I’m Not on Facebook"
-} \ No newline at end of file
+ "Title": "Facebook Is Tracking Me Even Though I’m Not on Facebook",
+ "SiteName": "American Civil Liberties Union"
+}
diff --git a/test/test-pages/ars-1/expected-metadata.json b/test/test-pages/ars-1/expected-metadata.json
index ad58a9f..0594bf2 100644
--- a/test/test-pages/ars-1/expected-metadata.json
+++ b/test/test-pages/ars-1/expected-metadata.json
@@ -3,5 +3,6 @@
"Direction": null,
"Excerpt": "Two-year-old bug exposes thousands of servers to crippling attack.",
"Image": "http:\/\/cdn.arstechnica.net\/wp-content\/uploads\/2015\/04\/server-crash-640x426.jpg",
- "Title": "Just-released Minecraft exploit makes it easy to crash game servers"
-} \ No newline at end of file
+ "Title": "Just-released Minecraft exploit makes it easy to crash game servers",
+ "SiteName": "Ars Technica"
+}
diff --git a/test/test-pages/base-url-base-element-relative/expected-metadata.json b/test/test-pages/base-url-base-element-relative/expected-metadata.json
index 127ccba..e58519c 100644
--- a/test/test-pages/base-url-base-element-relative/expected-metadata.json
+++ b/test/test-pages/base-url-base-element-relative/expected-metadata.json
@@ -3,5 +3,6 @@
"Direction": null,
"Excerpt": "Lorem ipsum dolor sit amet, consectetur adipisicing elit, sed do eiusmod\n tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam,\n quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo\n consequat. Duis aute irure dolor in reprehenderit in voluptate velit esse\n cillum dolore eu fugiat nulla pariatur. Excepteur sint occaecat cupidatat non\n proident, sunt in culpa qui officia deserunt mollit anim id est laborum.",
"Image": null,
- "Title": "Base URL with base relative test"
-} \ No newline at end of file
+ "Title": "Base URL with base relative test",
+ "SiteName": null
+}
diff --git a/test/test-pages/base-url-base-element/expected-metadata.json b/test/test-pages/base-url-base-element/expected-metadata.json
index d6a0d98..e7a86a7 100644
--- a/test/test-pages/base-url-base-element/expected-metadata.json
+++ b/test/test-pages/base-url-base-element/expected-metadata.json
@@ -3,5 +3,6 @@
"Direction": null,
"Excerpt": "Lorem ipsum dolor sit amet, consectetur adipisicing elit, sed do eiusmod\n tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam,\n quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo\n consequat. Duis aute irure dolor in reprehenderit in voluptate velit esse\n cillum dolore eu fugiat nulla pariatur. Excepteur sint occaecat cupidatat non\n proident, sunt in culpa qui officia deserunt mollit anim id est laborum.",
"Image": null,
- "Title": "Base URL with base test"
-} \ No newline at end of file
+ "Title": "Base URL with base test",
+ "SiteName": null
+}
diff --git a/test/test-pages/base-url/expected-metadata.json b/test/test-pages/base-url/expected-metadata.json
index f4c8b4b..291ff94 100644
--- a/test/test-pages/base-url/expected-metadata.json
+++ b/test/test-pages/base-url/expected-metadata.json
@@ -3,5 +3,6 @@
"Direction": null,
"Excerpt": "Lorem ipsum dolor sit amet, consectetur adipisicing elit, sed do eiusmod\n tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam,\n quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo\n consequat. Duis aute irure dolor in reprehenderit in voluptate velit esse\n cillum dolore eu fugiat nulla pariatur. Excepteur sint occaecat cupidatat non\n proident, sunt in culpa qui officia deserunt mollit anim id est laborum.",
"Image": null,
- "Title": "Base URL test"
-} \ No newline at end of file
+ "Title": "Base URL test",
+ "SiteName": null
+}
diff --git a/test/test-pages/basic-tags-cleaning/expected-metadata.json b/test/test-pages/basic-tags-cleaning/expected-metadata.json
index 4b13220..a31ae0b 100644
--- a/test/test-pages/basic-tags-cleaning/expected-metadata.json
+++ b/test/test-pages/basic-tags-cleaning/expected-metadata.json
@@ -3,5 +3,6 @@
"Direction": null,
"Excerpt": "Lorem ipsum dolor sit amet, consectetur adipisicing elit, sed do eiusmod\n tempor incididunt ut labore et dolore magna aliqua.",
"Image": null,
- "Title": "Basic tag cleaning test"
-} \ No newline at end of file
+ "Title": "Basic tag cleaning test",
+ "SiteName": null
+}
diff --git a/test/test-pages/bbc-1/expected-metadata.json b/test/test-pages/bbc-1/expected-metadata.json
index 6c25fe8..6af682f 100644
--- a/test/test-pages/bbc-1/expected-metadata.json
+++ b/test/test-pages/bbc-1/expected-metadata.json
@@ -3,5 +3,6 @@
"Direction": null,
"Excerpt": "President Barack Obama tells the BBC his failure to pass \"common sense gun safety laws\" is the greatest frustration of his presidency.",
"Image": "http:\/\/ichef.bbci.co.uk\/news\/1024\/cpsprodpb\/3D8B\/production\/_84455751_84455749.jpg",
- "Title": "Obama admits US gun laws are his 'biggest frustration' - BBC News"
-} \ No newline at end of file
+ "Title": "Obama admits US gun laws are his 'biggest frustration' - BBC News",
+ "SiteName": "BBC News"
+}
diff --git a/test/test-pages/blogger/expected-metadata.json b/test/test-pages/blogger/expected-metadata.json
index 43a34d6..a081c88 100644
--- a/test/test-pages/blogger/expected-metadata.json
+++ b/test/test-pages/blogger/expected-metadata.json
@@ -3,5 +3,6 @@
"Direction": null,
"Excerpt": "I've written a couple of posts in the past few months but they were all for the blog at work so I figured I'm long overdue for one on Silic...",
"Image": "https:\/\/1.bp.blogspot.com\/-YIPC5jkXkDE\/Vy7YPSqFKWI\/AAAAAAAAAxI\/a7D6Ji2GxoUvcrwUkI4RLZcr2LFQEJCTACLcB\/w1200-h630-p-nu\/block-diagram.png",
- "Title": "Open Verilog flow for Silego GreenPak4 programmable logic devices"
-} \ No newline at end of file
+ "Title": "Open Verilog flow for Silego GreenPak4 programmable logic devices",
+ "SiteName": null
+}
diff --git a/test/test-pages/breitbart/expected-metadata.json b/test/test-pages/breitbart/expected-metadata.json
index c1deac9..f044513 100644
--- a/test/test-pages/breitbart/expected-metadata.json
+++ b/test/test-pages/breitbart/expected-metadata.json
@@ -3,5 +3,6 @@
"Direction": null,
"Excerpt": "Snopes fact checker and staff writer David Emery posted to Twitter asking if there were “any un-angry Trump supporters?”",
"Image": "http:\/\/media.breitbart.com\/media\/2016\/11\/GettyImages-621866810.jpg",
- "Title": "'Neutral' Snopes Fact-Checker David Emery: 'Are There Any Un-Angry Trump Supporters?' - Breitbart"
-} \ No newline at end of file
+ "Title": "'Neutral' Snopes Fact-Checker David Emery: 'Are There Any Un-Angry Trump Supporters?' - Breitbart",
+ "SiteName": "Breitbart"
+}
diff --git a/test/test-pages/bug-1255978/expected-metadata.json b/test/test-pages/bug-1255978/expected-metadata.json
index 3085f86..b2b0d86 100644
--- a/test/test-pages/bug-1255978/expected-metadata.json
+++ b/test/test-pages/bug-1255978/expected-metadata.json
@@ -3,5 +3,6 @@
"Direction": null,
"Excerpt": "Most people go to hotels for the pleasure of sleeping in a giant bed with clean white sheets and waking up to fresh towels in the morning. But those towels and sheets might not be as clean as they look, according to the hotel bosses that responded to an online thread about the things hotel owners don’t want you to know.",
"Image": "https:\/\/static.independent.co.uk\/s3fs-public\/thumbnails\/image\/2015\/12\/06\/10\/bed-hotel-room.jpg",
- "Title": "Seven secrets that hotel owners don't want you to know"
-} \ No newline at end of file
+ "Title": "Seven secrets that hotel owners don't want you to know",
+ "SiteName": "The Independent"
+}
diff --git a/test/test-pages/buzzfeed-1/expected-metadata.json b/test/test-pages/buzzfeed-1/expected-metadata.json
index bfff1bd..c332ef4 100644
--- a/test/test-pages/buzzfeed-1/expected-metadata.json
+++ b/test/test-pages/buzzfeed-1/expected-metadata.json
@@ -3,5 +3,6 @@
"Direction": null,
"Excerpt": "An inquest into Eloise Parry's death has been adjourned until July.",
"Image": "http:\/\/s3-static-ak.buzzfed.com\/static\/2015-04\/22\/5\/campaign_images\/webdr03\/student-dies-after-diet-pills-she-bought-online-b-2-28712-1429696299-24_dblbig.jpg",
- "Title": "Student Dies After Diet Pills She Bought Online \"Burned Her Up From Within\""
-} \ No newline at end of file
+ "Title": "Student Dies After Diet Pills She Bought Online \"Burned Her Up From Within\"",
+ "SiteName": "BuzzFeed"
+}
diff --git a/test/test-pages/challenges/expected-metadata.json b/test/test-pages/challenges/expected-metadata.json
index d073d00..3631109 100644
--- a/test/test-pages/challenges/expected-metadata.json
+++ b/test/test-pages/challenges/expected-metadata.json
@@ -3,5 +3,6 @@
"Direction": null,
"Excerpt": "par Alexandria Sage et Lisa Girion",
"Image": "http:\/\/fakehost\/img\/cha\/placeholders\/placeholder_1000x750.jpg",
- "Title": "Les motivations du tueur de Las Vegas demeurent floues"
-} \ No newline at end of file
+ "Title": "Les motivations du tueur de Las Vegas demeurent floues",
+ "SiteName": "Challenges"
+}
diff --git a/test/test-pages/clean-links/expected-metadata.json b/test/test-pages/clean-links/expected-metadata.json
index 28aa9f8..c5166f5 100644
--- a/test/test-pages/clean-links/expected-metadata.json
+++ b/test/test-pages/clean-links/expected-metadata.json
@@ -3,5 +3,6 @@
"Direction": null,
"Excerpt": "Ere introducing the scrivener, as he first appeared to me, it is fit \n I make some mention of myself, my employees, my business, my chambers, \n and general surroundings; because some such description is indispensable \n to an adequate understanding of the chief character about to be presented.",
"Image": null,
- "Title": "Bartleby the Scrivener Web Study Text"
-} \ No newline at end of file
+ "Title": "Bartleby the Scrivener Web Study Text",
+ "SiteName": null
+}
diff --git a/test/test-pages/cnet-svg-classes/expected-metadata.json b/test/test-pages/cnet-svg-classes/expected-metadata.json
index d4c5947..be68e31 100644
--- a/test/test-pages/cnet-svg-classes/expected-metadata.json
+++ b/test/test-pages/cnet-svg-classes/expected-metadata.json
@@ -3,5 +3,6 @@
"Direction": null,
"Excerpt": "Twitter Lite llega a 11 países de América Latina, para ayudar a los usuarios con mala señal de sus redes móviles.",
"Image": "https:\/\/cdn1.cnet.com\/img\/JumVcu1ZSLtPP8ui0UWaSlgi5RU=\/670x503\/2017\/12\/01\/b36ce794-e0b8-495c-a198-184923a8f4e9\/twitter-lite.jpg",
- "Title": "Twitter Lite se estrena en México, Venezuela y otros nueve países"
-} \ No newline at end of file
+ "Title": "Twitter Lite se estrena en México, Venezuela y otros nueve países",
+ "SiteName": "CNET en Español"
+}
diff --git a/test/test-pages/cnet/expected-metadata.json b/test/test-pages/cnet/expected-metadata.json
index 5893635..773cbb9 100644
--- a/test/test-pages/cnet/expected-metadata.json
+++ b/test/test-pages/cnet/expected-metadata.json
@@ -3,5 +3,6 @@
"Direction": null,
"Excerpt": "Facebook CEO says be a friend and have a shared vision, but scare them when you have to and move fast.",
"Image": "https:\/\/cnet3.cbsistatic.com\/img\/1JaRRjqhoGxDVkFxTRRWkZgyK2Q=\/670x503\/2014\/03\/21\/863df5d9-e8b8-4b38-851b-5e3f77f2cf0e\/mark-zuckerberg-facebook-home-10671610x407.jpg",
- "Title": "Zuckerberg offers peek at Facebook's acquisition strategies"
-} \ No newline at end of file
+ "Title": "Zuckerberg offers peek at Facebook's acquisition strategies",
+ "SiteName": "CNET"
+}
diff --git a/test/test-pages/cnn/expected-metadata.json b/test/test-pages/cnn/expected-metadata.json
index 2b13919..46a71fa 100644
--- a/test/test-pages/cnn/expected-metadata.json
+++ b/test/test-pages/cnn/expected-metadata.json
@@ -3,5 +3,6 @@
"Direction": null,
"Excerpt": "A recently-released report on poverty and inequality found that the U.S. ranks the lowest among countries with welfare states.",
"Image": "http:\/\/i2.cdn.turner.com\/money\/dam\/assets\/141103182938-income-inequality-780x439.png",
- "Title": "The 'birth lottery' and economic mobility"
-} \ No newline at end of file
+ "Title": "The 'birth lottery' and economic mobility",
+ "SiteName": "CNNMoney"
+}
diff --git a/test/test-pages/comment-inside-script-parsing/expected-metadata.json b/test/test-pages/comment-inside-script-parsing/expected-metadata.json
index 09d330b..b1c1035 100644
--- a/test/test-pages/comment-inside-script-parsing/expected-metadata.json
+++ b/test/test-pages/comment-inside-script-parsing/expected-metadata.json
@@ -3,5 +3,6 @@
"Direction": null,
"Excerpt": "Lorem ipsum dolor sit amet, consectetur adipisicing elit, sed do eiusmod\n tempor incididunt ut labore et dolore magna aliqua.",
"Image": null,
- "Title": "Test script parsing"
-} \ No newline at end of file
+ "Title": "Test script parsing",
+ "SiteName": null
+}
diff --git a/test/test-pages/daringfireball-1/expected-metadata.json b/test/test-pages/daringfireball-1/expected-metadata.json
index 22b578d..1dbda03 100644
--- a/test/test-pages/daringfireball-1/expected-metadata.json
+++ b/test/test-pages/daringfireball-1/expected-metadata.json
@@ -3,5 +3,6 @@
"Direction": null,
"Excerpt": "Daring Fireball is written and produced by John Gruber.",
"Image": null,
- "Title": "Daring Fireball: Colophon"
-} \ No newline at end of file
+ "Title": "Daring Fireball: Colophon",
+ "SiteName": null
+}
diff --git a/test/test-pages/different-sources-for-images/expected-metadata.json b/test/test-pages/different-sources-for-images/expected-metadata.json
index 227cc96..c442d50 100644
--- a/test/test-pages/different-sources-for-images/expected-metadata.json
+++ b/test/test-pages/different-sources-for-images/expected-metadata.json
@@ -3,5 +3,6 @@
"Direction": null,
"Excerpt": "Lorem ipsum dolor sit amet, consectetur adipisicing elit, sed do eiusmod\n tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam,\n quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo\n consequat. Duis aute irure dolor in reprehenderit in voluptate velit esse\n cillum dolore eu fugiat nulla pariatur. Excepteur sint occaecat cupidatat non\n proident, sunt in culpa qui officia deserunt mollit anim id est laborum.",
"Image": null,
- "Title": "Image sources test"
-} \ No newline at end of file
+ "Title": "Image sources test",
+ "SiteName": null
+}
diff --git a/test/test-pages/ehow-1/expected-metadata.json b/test/test-pages/ehow-1/expected-metadata.json
index 48bc2dc..4de1448 100644
--- a/test/test-pages/ehow-1/expected-metadata.json
+++ b/test/test-pages/ehow-1/expected-metadata.json
@@ -3,5 +3,6 @@
"Direction": null,
"Excerpt": "Glass cloche terrariums are not only appealing to the eye, but they also preserve a bit of nature in your home and serve as a simple, yet beautiful, piece of art. Closed terrariums are easy to care for, as they retain much of their own moisture and provide a warm environment with a consistent level of humidity. You won’t have to water the...",
"Image": "http:\/\/img-aws.ehowcdn.com\/200x200\/cme\/photography.prod.demandstudios.com\/16149374-814f-40bc-baf3-ca20f149f0ba.jpg",
- "Title": "How to Build a Terrarium | eHow"
-} \ No newline at end of file
+ "Title": "How to Build a Terrarium | eHow",
+ "SiteName": "eHow"
+}
diff --git a/test/test-pages/ehow-2/expected-metadata.json b/test/test-pages/ehow-2/expected-metadata.json
index 6b3ccac..756b8c1 100644
--- a/test/test-pages/ehow-2/expected-metadata.json
+++ b/test/test-pages/ehow-2/expected-metadata.json
@@ -3,5 +3,6 @@
"Direction": null,
"Excerpt": "Graduation parties are a great way to commemorate the years of hard work teens and college co-eds devote to education. They’re also costly for mom and dad.The average cost of a graduation party in 2013 was a whopping $1,200, according to Graduationparty.com; $700 of that was allocated for food. However that budget was based on Midwestern...",
"Image": "http:\/\/img-aws.ehowcdn.com\/200x200\/cme\/cme_public_images\/www_ehow_com\/cdn-write.demandstudios.com\/upload\/image\/2F\/86\/5547EF62-EAF5-4256-945D-0496F61C862F\/5547EF62-EAF5-4256-945D-0496F61C862F.jpg",
- "Title": "How to Throw a Graduation Party on a Budget | eHow"
-} \ No newline at end of file
+ "Title": "How to Throw a Graduation Party on a Budget | eHow",
+ "SiteName": "eHow"
+}
diff --git a/test/test-pages/embedded-videos/expected-metadata.json b/test/test-pages/embedded-videos/expected-metadata.json
index 00239bc..e0e8859 100644
--- a/test/test-pages/embedded-videos/expected-metadata.json
+++ b/test/test-pages/embedded-videos/expected-metadata.json
@@ -3,5 +3,6 @@
"Direction": null,
"Excerpt": "Lorem ipsum dolor sit amet, consectetur adipisicing elit, sed do eiusmod\n tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam,\n quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo\n consequat. Duis aute irure dolor in reprehenderit in voluptate velit esse\n cillum dolore eu fugiat nulla pariatur. Excepteur sint occaecat cupidatat non\n proident, sunt in culpa qui officia deserunt mollit anim id est laborum.",
"Image": null,
- "Title": "Embedded videos test"
-} \ No newline at end of file
+ "Title": "Embedded videos test",
+ "SiteName": null
+}
diff --git a/test/test-pages/engadget/expected-metadata.json b/test/test-pages/engadget/expected-metadata.json
index 9d4c031..827dd63 100644
--- a/test/test-pages/engadget/expected-metadata.json
+++ b/test/test-pages/engadget/expected-metadata.json
@@ -3,5 +3,6 @@
"Direction": null,
"Excerpt": "The Xbox One X is the most powerful gaming console ever, but it's not for everyone yet.",
"Image": "https:\/\/o.aolcdn.com\/images\/dims?thumbnail=1200%2C630&quality=80&image_uri=https%3A%2F%2Fo.aolcdn.com%2Fimages%2Fdims%3Fcrop%3D1600%252C943%252C0%252C0%26quality%3D85%26format%3Djpg%26resize%3D1600%252C943%26image_uri%3Dhttp%253A%252F%252Fo.aolcdn.com%252Fhss%252Fstorage%252Fmidas%252Fd457f269d1400106a402302a310de800%252F205826071%252FXbox%252BOne%252BX%252Breview%252Bgallery%252B1.jpg%26client%3Da1acac3e1b3290917d92%26signature%3D9fa7b29452763464190c6edff6822489df516d1b&client=cbc79c14efcebee57402&signature=ab988e814a2686e0d93dd7ae306d241356fb7b9c",
- "Title": "Xbox One X review: A console that keeps up with gaming PCs"
-} \ No newline at end of file
+ "Title": "Xbox One X review: A console that keeps up with gaming PCs",
+ "SiteName": "Engadget"
+}
diff --git a/test/test-pages/fortressofsolitude/expected-metadata.json b/test/test-pages/fortressofsolitude/expected-metadata.json
index c166d47..8405f67 100644
--- a/test/test-pages/fortressofsolitude/expected-metadata.json
+++ b/test/test-pages/fortressofsolitude/expected-metadata.json
@@ -3,5 +3,6 @@
"Direction": null,
"Excerpt": "Just when you think the marketing behind the film couldn't get any more amazing, they go and release a music video from the Deadpool 2 soundtrack starring Canadian singer Céline Dion.",
"Image": "http:\/\/www.fortressofsolitude.co.za\/wp-content\/uploads\/2018\/05\/Céline-Dion-Belches-Her-Way-Onto-Deadpool-2s-Soundtrack.jpg",
- "Title": "Céline Dion Belches Her Way Onto Deadpool 2's Soundtrack"
-} \ No newline at end of file
+ "Title": "Céline Dion Belches Her Way Onto Deadpool 2's Soundtrack",
+ "SiteName": "Fortress of Solitude"
+}
diff --git a/test/test-pages/gmw/expected-metadata.json b/test/test-pages/gmw/expected-metadata.json
index b6f43bf..65cd6ab 100644
--- a/test/test-pages/gmw/expected-metadata.json
+++ b/test/test-pages/gmw/expected-metadata.json
@@ -3,5 +3,6 @@
"Direction": null,
"Excerpt": "不幸的是,对于希望能喝上一杯的太空探险者,那些将他们送上太空的政府机构普遍禁止他们染指包括酒在内的含酒精饮料。",
"Image": null,
- "Title": "宇航员在太空中喝酒会怎么样?后果很严重 _探索者 _光明网"
-} \ No newline at end of file
+ "Title": "宇航员在太空中喝酒会怎么样?后果很严重 _探索者 _光明网",
+ "SiteName": null
+}
diff --git a/test/test-pages/heise/expected-metadata.json b/test/test-pages/heise/expected-metadata.json
index e914ef1..bb8e4c4 100644
--- a/test/test-pages/heise/expected-metadata.json
+++ b/test/test-pages/heise/expected-metadata.json
@@ -3,5 +3,6 @@
"Direction": null,
"Excerpt": "Das in der iOS-Version bereits enthaltene TOTP-Feature ist nun auch für OS X 10.10 verfügbar. Zudem gibt es neue Zusatzfelder in der Datenbank und weitere Verbesserungen.",
"Image": "http:\/\/www.heise.de\/imgs\/18\/1\/4\/6\/2\/3\/5\/1\/Barcode-Scanner-With-Border-f0c62350bd8d9d96.jpeg",
- "Title": "1Password für Mac generiert Einmal-Passwörter"
-} \ No newline at end of file
+ "Title": "1Password für Mac generiert Einmal-Passwörter",
+ "SiteName": "Mac & i"
+}
diff --git a/test/test-pages/herald-sun-1/expected-metadata.json b/test/test-pages/herald-sun-1/expected-metadata.json
index 8067601..1024132 100644
--- a/test/test-pages/herald-sun-1/expected-metadata.json
+++ b/test/test-pages/herald-sun-1/expected-metadata.json
@@ -3,5 +3,6 @@
"Direction": null,
"Excerpt": "A HIGH-powered federal government team has been doing the rounds of media organisations in the past few days in an attempt to allay concerns about the impact of new surveillance legislation on press freedom. It failed.",
"Image": "http:\/\/api.news.com.au\/content\/1.0\/heraldsun\/images\/1227261885862?format=jpg&group=iphone&size=medium",
- "Title": "Angry media won’t buckle over new surveillance laws"
-} \ No newline at end of file
+ "Title": "Angry media won’t buckle over new surveillance laws",
+ "SiteName": "HeraldSun"
+}
diff --git a/test/test-pages/hidden-nodes/expected-metadata.json b/test/test-pages/hidden-nodes/expected-metadata.json
index a6d990d..332533a 100644
--- a/test/test-pages/hidden-nodes/expected-metadata.json
+++ b/test/test-pages/hidden-nodes/expected-metadata.json
@@ -3,5 +3,6 @@
"Direction": null,
"Excerpt": "Lorem ipsum dolor sit amet, consetetur sadipscing elitr, sed diam nonumy eirmod tempor invidunt ut labore et dolore magna aliquyam erat, sed diam voluptua. At vero eos et accusam et justo duo dolores et ea rebum. Stet clita kasd gubergren, no sea takimata sanctus est Lorem ipsum dolor sit amet. Lorem ipsum dolor sit amet, consetetur sadipscing elitr, sed diam nonumy eirmod tempor invidunt ut labore et dolore magna aliquyam erat, sed diam voluptua. At vero eos et accusam et justo duo dolores et ea rebum. Stet clita kasd gubergren, no sea takimata sanctus est Lorem ipsum dolor sit amet. Lorem ipsum dolor sit amet, consetetur sadipscing elitr, sed diam nonumy eirmod tempor invidunt ut labore et dolore magna aliquyam erat, sed diam voluptua. At vero eos et accusam et justo duo dolores et ea rebum. Stet clita kasd gubergren, no sea takimata sanctus est Lorem ipsum dolor sit amet.",
"Image": null,
- "Title": "Lorem ipsum dolor sit amet, consetetur sadipscing elitr, sed diam nonumy eirmod tempor invidunt"
-} \ No newline at end of file
+ "Title": "Lorem ipsum dolor sit amet, consetetur sadipscing elitr, sed diam nonumy eirmod tempor invidunt",
+ "SiteName": null
+}
diff --git a/test/test-pages/hukumusume/expected-metadata.json b/test/test-pages/hukumusume/expected-metadata.json
index 14eb945..b16149a 100644
--- a/test/test-pages/hukumusume/expected-metadata.json
+++ b/test/test-pages/hukumusume/expected-metadata.json
@@ -3,5 +3,6 @@
"Direction": null,
"Excerpt": "福娘童話集 > きょうのイソップ童話 > 1月のイソップ童話 > 欲張りなイヌ",
"Image": null,
- "Title": "欲張りなイヌ <福娘童話集 きょうのイソップ童話>"
-} \ No newline at end of file
+ "Title": "欲張りなイヌ <福娘童話集 きょうのイソップ童話>",
+ "SiteName": null
+}
diff --git a/test/test-pages/iab-1/expected-metadata.json b/test/test-pages/iab-1/expected-metadata.json
index b013131..a8d6766 100644
--- a/test/test-pages/iab-1/expected-metadata.json
+++ b/test/test-pages/iab-1/expected-metadata.json
@@ -3,5 +3,6 @@
"Direction": null,
"Excerpt": "We messed up. As technologists, tasked with delivering content and services to users, we lost track of the user experience. Twenty years ago we saw an explosion of websites, built by developers around the world, providing all forms of content. This was the beginning of an age of enlightenment, the intersection of content and technology. … Continued",
"Image": "http:\/\/www.iab.com\/wp-content\/uploads\/2015\/10\/getting-lean-with-digital-ad-ux-300x250.jpg",
- "Title": "Getting LEAN with Digital Ad UX | IAB"
-} \ No newline at end of file
+ "Title": "Getting LEAN with Digital Ad UX | IAB",
+ "SiteName": "IAB"
+}
diff --git a/test/test-pages/ietf-1/expected-metadata.json b/test/test-pages/ietf-1/expected-metadata.json
index 184dc78..ddf0094 100644
--- a/test/test-pages/ietf-1/expected-metadata.json
+++ b/test/test-pages/ietf-1/expected-metadata.json
@@ -3,5 +3,6 @@
"Direction": null,
"Excerpt": null,
"Image": null,
- "Title": "remoteStorage"
-} \ No newline at end of file
+ "Title": "remoteStorage",
+ "SiteName": null
+}
diff --git a/test/test-pages/infobae/expected-metadata.json b/test/test-pages/infobae/expected-metadata.json
index 6a85498..b6d2597 100644
--- a/test/test-pages/infobae/expected-metadata.json
+++ b/test/test-pages/infobae/expected-metadata.json
@@ -3,5 +3,6 @@
"Direction": null,
"Excerpt": "Jonathan Castellari fue salvajemente golpeado en un local de comidas rápidas, mientras esperaba el desayuno junto a un amigo. Podría perder un ojo",
"Image": "https:\/\/www.infobae.com\/new-resizer\/shi8E0F5ldd8a7vtvFaJomRURqU=\/1200x0\/filters:quality(100)\/s3.amazonaws.com\/arc-wordpress-client-uploads\/infobae-wp\/wp-content\/uploads\/2017\/12\/01154429\/Rugbier-golpeado-1920.jpg",
- "Title": "Una patota de varones atacó a un rugbier gay: \"Te vamos a matar por puto\""
-} \ No newline at end of file
+ "Title": "Una patota de varones atacó a un rugbier gay: \"Te vamos a matar por puto\"",
+ "SiteName": "Infobae"
+}
diff --git a/test/test-pages/keep-images/expected-metadata.json b/test/test-pages/keep-images/expected-metadata.json
index 4db5427..3b705f9 100644
--- a/test/test-pages/keep-images/expected-metadata.json
+++ b/test/test-pages/keep-images/expected-metadata.json
@@ -3,5 +3,6 @@
"Direction": null,
"Excerpt": "Welcome to DoctorX’s Barcelona lab, where the drugs you bought online are tested for safety and purity. No questions ask…",
"Image": "https:\/\/d262ilb51hltx0.cloudfront.net\/max\/800\/1*sLDnS1UWEFIS33uLMxq3cw.jpeg",
- "Title": "Inside the Deep Web Drug Lab"
-} \ No newline at end of file
+ "Title": "Inside the Deep Web Drug Lab",
+ "SiteName": "Medium"
+}
diff --git a/test/test-pages/la-nacion/expected-metadata.json b/test/test-pages/la-nacion/expected-metadata.json
index a33a01d..d2e56b0 100644
--- a/test/test-pages/la-nacion/expected-metadata.json
+++ b/test/test-pages/la-nacion/expected-metadata.json
@@ -3,5 +3,6 @@
"Direction": null,
"Excerpt": "Los pueblos indígenas reclaman por derechos que permanecen incumplidos, por eso es más eficiente canalizar la protesta que reprimirla",
"Image": "http:\/\/bucket.glanacion.com\/anexos\/fotos\/77\/2585177.jpg",
- "Title": "Una solución no violenta para la cuestión mapuche"
-} \ No newline at end of file
+ "Title": "Una solución no violenta para la cuestión mapuche",
+ "SiteName": null
+}
diff --git a/test/test-pages/lemonde-1/expected-metadata.json b/test/test-pages/lemonde-1/expected-metadata.json
index 0b5313c..a45fcaf 100644
--- a/test/test-pages/lemonde-1/expected-metadata.json
+++ b/test/test-pages/lemonde-1/expected-metadata.json
@@ -3,5 +3,6 @@
"Direction": null,
"Excerpt": "Largement approuvé par les députés, le texte sera désormais examiné par le Sénat, puis le Conseil constitutionnel.",
"Image": "http:\/\/s1.lemde.fr\/image\/2015\/05\/05\/600x315\/4628128_3_47fc_projet-de-loi-renseignement_aeba800424730d672d1bd08faf203438.jpg",
- "Title": "Le projet de loi sur le renseignement massivement approuvé à l'Assemblée"
-} \ No newline at end of file
+ "Title": "Le projet de loi sur le renseignement massivement approuvé à l'Assemblée",
+ "SiteName": "Le Monde.fr"
+}
diff --git a/test/test-pages/lemonde-2/expected-metadata.json b/test/test-pages/lemonde-2/expected-metadata.json
index c7056f1..c402d3e 100644
--- a/test/test-pages/lemonde-2/expected-metadata.json
+++ b/test/test-pages/lemonde-2/expected-metadata.json
@@ -3,5 +3,6 @@
"Direction": null,
"Excerpt": "Le retour à la stabilité coûtera des milliards d'euros, mais l'Union européenne en vaut la peine, estime un collectif d'industriels.",
"Image": "http:\/\/s1.lemde.fr\/medias\/web\/1.2.705\/img\/placeholder\/default.png",
- "Title": "La crise européenne est une crise de la dette, pas de l'euro"
-} \ No newline at end of file
+ "Title": "La crise européenne est une crise de la dette, pas de l'euro",
+ "SiteName": "Le Monde.fr"
+}
diff --git a/test/test-pages/liberation-1/expected-metadata.json b/test/test-pages/liberation-1/expected-metadata.json
index 32108e3..5f2b7b5 100644
--- a/test/test-pages/liberation-1/expected-metadata.json
+++ b/test/test-pages/liberation-1/expected-metadata.json
@@ -3,5 +3,6 @@
"Direction": null,
"Excerpt": "Laurent Fabius a accueilli jeudi matin à Roissy un premier avion spécial ramenant des rescapés.",
"Image": "http:\/\/md1.libe.com\/photo\/755923-000_hkg10175905.jpg?modified_at=1430371146&amp;width=750",
- "Title": "Un troisième Français mort dans le séisme au Népal"
-} \ No newline at end of file
+ "Title": "Un troisième Français mort dans le séisme au Népal",
+ "SiteName": "Libération.fr"
+}
diff --git a/test/test-pages/lifehacker-post-comment-load/expected-metadata.json b/test/test-pages/lifehacker-post-comment-load/expected-metadata.json
index 31ab106..fa13f9e 100644
--- a/test/test-pages/lifehacker-post-comment-load/expected-metadata.json
+++ b/test/test-pages/lifehacker-post-comment-load/expected-metadata.json
@@ -3,5 +3,6 @@
"Direction": null,
"Excerpt": "We all buy things from time to time that we don't really need. It's okay to appeal to your wants every once in a while, as long as you're in control. If you struggle with clutter, impulse buys, and buyer's remorse, here's how to put your mind in the right place before you even set foot in a store.",
"Image": "http:\/\/i.kinja-img.com\/gawker-media\/image\/upload\/s--hqqO9fze--\/n1s6c2m6kc07iqdyllj6.jpg",
- "Title": "How to Program Your Mind to Stop Buying Crap You Don’t Need"
-} \ No newline at end of file
+ "Title": "How to Program Your Mind to Stop Buying Crap You Don’t Need",
+ "SiteName": "Lifehacker"
+}
diff --git a/test/test-pages/lifehacker-working/expected-metadata.json b/test/test-pages/lifehacker-working/expected-metadata.json
index 31ab106..fa13f9e 100644
--- a/test/test-pages/lifehacker-working/expected-metadata.json
+++ b/test/test-pages/lifehacker-working/expected-metadata.json
@@ -3,5 +3,6 @@
"Direction": null,
"Excerpt": "We all buy things from time to time that we don't really need. It's okay to appeal to your wants every once in a while, as long as you're in control. If you struggle with clutter, impulse buys, and buyer's remorse, here's how to put your mind in the right place before you even set foot in a store.",
"Image": "http:\/\/i.kinja-img.com\/gawker-media\/image\/upload\/s--hqqO9fze--\/n1s6c2m6kc07iqdyllj6.jpg",
- "Title": "How to Program Your Mind to Stop Buying Crap You Don’t Need"
-} \ No newline at end of file
+ "Title": "How to Program Your Mind to Stop Buying Crap You Don’t Need",
+ "SiteName": "Lifehacker"
+}
diff --git a/test/test-pages/links-in-tables/expected-metadata.json b/test/test-pages/links-in-tables/expected-metadata.json
index f0b9794..d6b1330 100644
--- a/test/test-pages/links-in-tables/expected-metadata.json
+++ b/test/test-pages/links-in-tables/expected-metadata.json
@@ -3,5 +3,6 @@
"Direction": null,
"Excerpt": "Posted by Andrew Hayden, Software Engineer on Google Play Android users are downloading tens of billions of apps and games on Google Pla...",
"Image": "https:\/\/2.bp.blogspot.com\/-chCZZinlUTg\/WEcxvJo9gdI\/AAAAAAAADnk\/3ND_BspqN6Y2j5xxkLFW3RyS2Ig0NHZpQCLcB\/w1200-h630-p-k-nu\/ipsum-opsum.gif",
- "Title": "Saving Data: Reducing the size of App Updates by 65%"
-} \ No newline at end of file
+ "Title": "Saving Data: Reducing the size of App Updates by 65%",
+ "SiteName": null
+}
diff --git a/test/test-pages/lwn-1/expected-metadata.json b/test/test-pages/lwn-1/expected-metadata.json
index 2a94d7a..584887a 100644
--- a/test/test-pages/lwn-1/expected-metadata.json
+++ b/test/test-pages/lwn-1/expected-metadata.json
@@ -3,5 +3,6 @@
"Direction": null,
"Excerpt": "The Arduino has been one of the biggest success stories of the open-hardware movement, but that success does not protect it from internal conflict. In recent months, two of the project's founders have come into conflict about the direction of future efforts—and that conflict has turned into a legal dispute about who owns the rights to the Arduino trademark.",
"Image": null,
- "Title": "LWN.net Weekly Edition for March 26, 2015 [LWN.net]"
-} \ No newline at end of file
+ "Title": "LWN.net Weekly Edition for March 26, 2015 [LWN.net]",
+ "SiteName": null
+}
diff --git a/test/test-pages/medium-1/expected-metadata.json b/test/test-pages/medium-1/expected-metadata.json
index 2d57836..a974b43 100644
--- a/test/test-pages/medium-1/expected-metadata.json
+++ b/test/test-pages/medium-1/expected-metadata.json
@@ -3,5 +3,6 @@
"Direction": null,
"Excerpt": "We pushed out the first version of the Open Journalism site in January. Here’s what we’ve learned about student journali…",
"Image": "https:\/\/d262ilb51hltx0.cloudfront.net\/max\/800\/1*oBWUXtszDsiv_-Qq2bFLTQ.png",
- "Title": "The Open Journalism Project: Better Student Journalism"
-} \ No newline at end of file
+ "Title": "The Open Journalism Project: Better Student Journalism",
+ "SiteName": "Medium"
+}
diff --git a/test/test-pages/medium-2/expected-metadata.json b/test/test-pages/medium-2/expected-metadata.json
index 960f17f..71804a0 100644
--- a/test/test-pages/medium-2/expected-metadata.json
+++ b/test/test-pages/medium-2/expected-metadata.json
@@ -3,5 +3,6 @@
"Direction": null,
"Excerpt": "In defense of the word “literally” and why you or someone you know should stop misusing the word, lest they drive us fig…",
"Image": "https:\/\/d262ilb51hltx0.cloudfront.net\/max\/1600\/1*eR_J8DurqygbhrwDg-WPnQ.png",
- "Title": "On Behalf of “Literally”"
-} \ No newline at end of file
+ "Title": "On Behalf of “Literally”",
+ "SiteName": "Medium"
+}
diff --git a/test/test-pages/medium-3/expected-metadata.json b/test/test-pages/medium-3/expected-metadata.json
index 5035ee8..59ef204 100644
--- a/test/test-pages/medium-3/expected-metadata.json
+++ b/test/test-pages/medium-3/expected-metadata.json
@@ -3,5 +3,6 @@
"Direction": null,
"Excerpt": "How to get shanked doing what people say they want",
"Image": "https:\/\/cdn-images-1.medium.com\/max\/1200\/1*kbPh7V97eyRodSOw2-ALDw.png",
- "Title": "Samantha and The Great Big Lie"
-} \ No newline at end of file
+ "Title": "Samantha and The Great Big Lie",
+ "SiteName": "Medium"
+}
diff --git a/test/test-pages/missing-paragraphs/expected-metadata.json b/test/test-pages/missing-paragraphs/expected-metadata.json
index e12c231..b0ef946 100644
--- a/test/test-pages/missing-paragraphs/expected-metadata.json
+++ b/test/test-pages/missing-paragraphs/expected-metadata.json
@@ -3,5 +3,6 @@
"Direction": null,
"Excerpt": "Lorem ipsum dolor sit amet, consetetur sadipscing elitr, sed diam nonumy\n eirmod tempor invidunt ut labore et dolore magna aliquyam erat, sed diam\n voluptua. At vero eos et accusam et justo duo dolores et ea rebum. Stet\n clita kasd gubergren, no sea takimata sanctus est Lorem ipsum dolor sit\n amet. Lorem ipsum dolor sit amet, consetetur sadipscing elitr, sed diam\n nonumy eirmod tempor invidunt ut labore et dolore magna aliquyam erat,\n sed diam voluptua. At vero eos et accusam et justo duo dolores et ea rebum.\n Stet clita kasd gubergren, no sea takimata sanctus est Lorem ipsum dolor\n sit amet. Lorem ipsum dolor sit amet, consetetur sadipscing elitr, sed\n diam nonumy eirmod tempor invidunt ut labore et dolore magna aliquyam erat,\n sed diam voluptua. At vero eos et accusam et justo duo dolores et ea rebum.\n Stet clita kasd gubergren, no sea takimata sanctus est Lorem ipsum dolor\n sit amet.",
"Image": null,
- "Title": "Lorem ipsum dolor sit amet, consetetur sadipscing elitr, sed diam nonumy\n eirmod tempor invidunt"
-} \ No newline at end of file
+ "Title": "Lorem ipsum dolor sit amet, consetetur sadipscing elitr, sed diam nonumy\n eirmod tempor invidunt",
+ "SiteName": null
+}
diff --git a/test/test-pages/mozilla-1/expected-metadata.json b/test/test-pages/mozilla-1/expected-metadata.json
index bb03f81..1b17f9b 100644
--- a/test/test-pages/mozilla-1/expected-metadata.json
+++ b/test/test-pages/mozilla-1/expected-metadata.json
@@ -3,5 +3,6 @@
"Direction": "ltr",
"Excerpt": "It’s easier than ever to personalize Firefox and make it work the way\n you do.\n No other browser gives you so much choice and flexibility.",
"Image": "https:\/\/mozorg.cdn.mozilla.net\/media\/img\/firefox\/template\/page-image.af8027a425de.png",
- "Title": "Firefox — Customize and make it your own — The most flexible browser on the Web"
-} \ No newline at end of file
+ "Title": "Firefox — Customize and make it your own — The most flexible browser on the Web",
+ "SiteName": "Mozilla"
+}
diff --git a/test/test-pages/mozilla-2/expected-metadata.json b/test/test-pages/mozilla-2/expected-metadata.json
index 49e7ffd..7336ada 100644
--- a/test/test-pages/mozilla-2/expected-metadata.json
+++ b/test/test-pages/mozilla-2/expected-metadata.json
@@ -3,5 +3,6 @@
"Direction": "ltr",
"Excerpt": "Built for those who build the Web. Introducing the only browser made for developers.",
"Image": "https:\/\/mozorg.cdn.mozilla.net\/media\/img\/firefox\/developer\/page-image.03bbe7da3199.png",
- "Title": "Welcome to Firefox Developer Edition"
-} \ No newline at end of file
+ "Title": "Welcome to Firefox Developer Edition",
+ "SiteName": "Mozilla"
+}
diff --git a/test/test-pages/msn/expected-metadata.json b/test/test-pages/msn/expected-metadata.json
index a11c62a..673d995 100644
--- a/test/test-pages/msn/expected-metadata.json
+++ b/test/test-pages/msn/expected-metadata.json
@@ -3,5 +3,6 @@
"Direction": null,
"Excerpt": "Nintendo and Apple shocked the world earlier this year by announcing \"Super Mario Run,\" the legendary gaming company's first foray into mobile gaming.&nbsp;",
"Image": "http:\/\/img-s-msn-com.akamaized.net\/tenant\/amp\/entityid\/AAkk5fh.img",
- "Title": "Nintendo's first iPhone game will launch in December for $10"
-} \ No newline at end of file
+ "Title": "Nintendo's first iPhone game will launch in December for $10",
+ "SiteName": "MSN"
+}
diff --git a/test/test-pages/nytimes-1/expected-metadata.json b/test/test-pages/nytimes-1/expected-metadata.json
index c84bcce..0d1c802 100644
--- a/test/test-pages/nytimes-1/expected-metadata.json
+++ b/test/test-pages/nytimes-1/expected-metadata.json
@@ -3,5 +3,6 @@
"Direction": null,
"Excerpt": "For the first time since the 1990s, the country will be able to trade extensively with the United States.",
"Image": "https:\/\/static01.nyt.com\/images\/2017\/01\/14\/world\/13SUDAN-1\/13SUDAN-1-facebookJumbo.jpg",
- "Title": "United States to Lift Sudan Sanctions"
-} \ No newline at end of file
+ "Title": "United States to Lift Sudan Sanctions",
+ "SiteName": null
+}
diff --git a/test/test-pages/nytimes-2/expected-metadata.json b/test/test-pages/nytimes-2/expected-metadata.json
index 1ac8a4e..392535f 100644
--- a/test/test-pages/nytimes-2/expected-metadata.json
+++ b/test/test-pages/nytimes-2/expected-metadata.json
@@ -3,5 +3,6 @@
"Direction": null,
"Excerpt": "The internet giant’s decision to sell its business is plagued with challenges that reveal how unusual deal structures can affect shareholders.",
"Image": "https:\/\/static01.nyt.com\/images\/2016\/07\/30\/business\/db-dealprof\/db-dealprof-facebookJumbo.jpg",
- "Title": "Yahoo’s Sale to Verizon Leaves Shareholders With Little Say"
-} \ No newline at end of file
+ "Title": "Yahoo’s Sale to Verizon Leaves Shareholders With Little Say",
+ "SiteName": null
+}
diff --git a/test/test-pages/pixnet/expected-metadata.json b/test/test-pages/pixnet/expected-metadata.json
index 7fde5d0..dafb56b 100644
--- a/test/test-pages/pixnet/expected-metadata.json
+++ b/test/test-pages/pixnet/expected-metadata.json
@@ -3,5 +3,6 @@
"Direction": null,
"Excerpt": "一波波接續性低溫寒流報到 已將新竹尖石鄉後山一帶層層山巒披上嫣紅的彩衣 玉峰道路一路上雲氣山嵐滯留山頭 順路下切蜿蜒道路後不久即抵達來到&quot;玉峰國小&quot; &quot;美樹&quot;美",
"Image": "http:\/\/pic.pimg.tw\/stevenhgm\/1387895093-631461272.jpg",
- "Title": "新竹尖石_美樹營地賞楓 (2) @ 史蒂文的家_藍天 :: 痞客邦 PIXNET ::"
-} \ No newline at end of file
+ "Title": "新竹尖石_美樹營地賞楓 (2) @ 史蒂文的家_藍天 :: 痞客邦 PIXNET ::",
+ "SiteName": "史蒂文的家_藍天"
+}
diff --git a/test/test-pages/remove-extra-brs/expected-metadata.json b/test/test-pages/remove-extra-brs/expected-metadata.json
index 29c7130..9a7e3d1 100644
--- a/test/test-pages/remove-extra-brs/expected-metadata.json
+++ b/test/test-pages/remove-extra-brs/expected-metadata.json
@@ -3,5 +3,6 @@
"Direction": null,
"Excerpt": "Lorem ipsum dolor sit amet, consectetur adipisicing elit, sed do eiusmod\n tempor incididunt ut labore et dolore magna aliqua.",
"Image": null,
- "Title": "Remove trailing brs test"
-} \ No newline at end of file
+ "Title": "Remove trailing brs test",
+ "SiteName": null
+}
diff --git a/test/test-pages/remove-extra-paragraphs/expected-metadata.json b/test/test-pages/remove-extra-paragraphs/expected-metadata.json
index 3754222..2e6f722 100644
--- a/test/test-pages/remove-extra-paragraphs/expected-metadata.json
+++ b/test/test-pages/remove-extra-paragraphs/expected-metadata.json
@@ -3,5 +3,6 @@
"Direction": null,
"Excerpt": "Lorem ipsum dolor sit amet, consectetur adipisicing elit, sed do eiusmod\n tempor incididunt ut labore et dolore magna aliqua.",
"Image": null,
- "Title": "Replace font tags test"
-} \ No newline at end of file
+ "Title": "Replace font tags test",
+ "SiteName": null
+}
diff --git a/test/test-pages/remove-script-tags/expected-metadata.json b/test/test-pages/remove-script-tags/expected-metadata.json
index ff392e7..13cd351 100644
--- a/test/test-pages/remove-script-tags/expected-metadata.json
+++ b/test/test-pages/remove-script-tags/expected-metadata.json
@@ -3,5 +3,6 @@
"Direction": null,
"Excerpt": "Lorem ipsum dolor sit amet, consectetur adipisicing elit, sed do eiusmod\n tempor incididunt ut labore et dolore magna aliqua.",
"Image": null,
- "Title": "Remove script tags test"
-} \ No newline at end of file
+ "Title": "Remove script tags test",
+ "SiteName": null
+}
diff --git a/test/test-pages/reordering-paragraphs/expected-metadata.json b/test/test-pages/reordering-paragraphs/expected-metadata.json
index da9fc37..4878f5d 100644
--- a/test/test-pages/reordering-paragraphs/expected-metadata.json
+++ b/test/test-pages/reordering-paragraphs/expected-metadata.json
@@ -3,5 +3,6 @@
"Direction": null,
"Excerpt": "Regarding item# 11111, under sufficiently extreme conditions, quarks may\n become deconfined and exist as free particles. In the course of asymptotic\n freedom, the strong interaction becomes weaker at higher temperatures.\n Eventually, color confinement would be lost and an extremely hot plasma\n of freely moving quarks and gluons would be formed. This theoretical phase\n of matter is called quark-gluon plasma.[81] The exact conditions needed\n to give rise to this state are unknown and have been the subject of a great\n deal of speculation and experimentation.",
"Image": null,
- "Title": null
-} \ No newline at end of file
+ "Title": null,
+ "SiteName": null
+}
diff --git a/test/test-pages/replace-brs/expected-metadata.json b/test/test-pages/replace-brs/expected-metadata.json
index b8e513b..2f31b06 100644
--- a/test/test-pages/replace-brs/expected-metadata.json
+++ b/test/test-pages/replace-brs/expected-metadata.json
@@ -3,5 +3,6 @@
"Direction": null,
"Excerpt": "Lorem ipsumdolor sit",
"Image": null,
- "Title": "Replace brs test"
-} \ No newline at end of file
+ "Title": "Replace brs test",
+ "SiteName": null
+}
diff --git a/test/test-pages/replace-font-tags/expected-metadata.json b/test/test-pages/replace-font-tags/expected-metadata.json
index 8f62e52..18f46dd 100644
--- a/test/test-pages/replace-font-tags/expected-metadata.json
+++ b/test/test-pages/replace-font-tags/expected-metadata.json
@@ -3,5 +3,6 @@
"Direction": null,
"Excerpt": "Lorem ipsum dolor sit amet, consectetur adipisicing elit, sed do eiusmod\n tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam,\n quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo\n consequat. Duis aute irure dolor in reprehenderit in voluptate velit esse\n cillum dolore eu fugiat nulla pariatur. Excepteur sint occaecat cupidatat non\n proident, sunt in culpa qui officia deserunt mollit anim id est laborum.",
"Image": null,
- "Title": "Replace font tags test"
-} \ No newline at end of file
+ "Title": "Replace font tags test",
+ "SiteName": null
+}
diff --git a/test/test-pages/rtl-1/expected-metadata.json b/test/test-pages/rtl-1/expected-metadata.json
index e3d3341..f9613b2 100644
--- a/test/test-pages/rtl-1/expected-metadata.json
+++ b/test/test-pages/rtl-1/expected-metadata.json
@@ -3,5 +3,6 @@
"Direction": "rtl",
"Excerpt": "Lorem ipsum dolor sit amet.",
"Image": null,
- "Title": "RTL Test"
-} \ No newline at end of file
+ "Title": "RTL Test",
+ "SiteName": null
+}
diff --git a/test/test-pages/rtl-2/expected-metadata.json b/test/test-pages/rtl-2/expected-metadata.json
index e3d3341..f9613b2 100644
--- a/test/test-pages/rtl-2/expected-metadata.json
+++ b/test/test-pages/rtl-2/expected-metadata.json
@@ -3,5 +3,6 @@
"Direction": "rtl",
"Excerpt": "Lorem ipsum dolor sit amet.",
"Image": null,
- "Title": "RTL Test"
-} \ No newline at end of file
+ "Title": "RTL Test",
+ "SiteName": null
+}
diff --git a/test/test-pages/rtl-3/expected-metadata.json b/test/test-pages/rtl-3/expected-metadata.json
index e3d3341..f9613b2 100644
--- a/test/test-pages/rtl-3/expected-metadata.json
+++ b/test/test-pages/rtl-3/expected-metadata.json
@@ -3,5 +3,6 @@
"Direction": "rtl",
"Excerpt": "Lorem ipsum dolor sit amet.",
"Image": null,
- "Title": "RTL Test"
-} \ No newline at end of file
+ "Title": "RTL Test",
+ "SiteName": null
+}
diff --git a/test/test-pages/rtl-4/expected-metadata.json b/test/test-pages/rtl-4/expected-metadata.json
index 4846771..c8358d5 100644
--- a/test/test-pages/rtl-4/expected-metadata.json
+++ b/test/test-pages/rtl-4/expected-metadata.json
@@ -3,5 +3,6 @@
"Direction": null,
"Excerpt": "Lorem ipsum dolor sit amet.",
"Image": null,
- "Title": "RTL Test"
-} \ No newline at end of file
+ "Title": "RTL Test",
+ "SiteName": null
+}
diff --git a/test/test-pages/salon-1/expected-metadata.json b/test/test-pages/salon-1/expected-metadata.json
index 45af008..63cf935 100644
--- a/test/test-pages/salon-1/expected-metadata.json
+++ b/test/test-pages/salon-1/expected-metadata.json
@@ -3,5 +3,6 @@
"Direction": null,
"Excerpt": "Disruptive companies talk a good game about sharing. Uber's really just an under-regulated company making riches",
"Image": "http:\/\/media.salon.com\/2014\/12\/uber_rand_paul.jpg",
- "Title": "The sharing economy is a lie: Uber, Ayn Rand and the truth about tech and libertarians"
-} \ No newline at end of file
+ "Title": "The sharing economy is a lie: Uber, Ayn Rand and the truth about tech and libertarians",
+ "SiteName": null
+}
diff --git a/test/test-pages/simplyfound-1/expected-metadata.json b/test/test-pages/simplyfound-1/expected-metadata.json
index 62494e2..b024aa8 100644
--- a/test/test-pages/simplyfound-1/expected-metadata.json
+++ b/test/test-pages/simplyfound-1/expected-metadata.json
@@ -3,5 +3,6 @@
"Direction": null,
"Excerpt": "The Raspberry Pi Foundation started by a handful of volunteers in 2012 when they released the original Raspberry Pi 256MB Model B without knowing what to expect. In a short four-year period they have grown to over sixty full-time employees and ha...",
"Image": "https:\/\/d34hb2g9mvfppu.cloudfront.net\/m\/images\/cache\/images\/2016\/02\/29\/apcnews2012raspberry_pi_logo_mainimage8_jpg8_322_27630a8388eb_lg.jpg",
- "Title": "Raspberry Pi 3 - The credit card sized PC that cost only $35 - All-time bestselling computer in UK"
-} \ No newline at end of file
+ "Title": "Raspberry Pi 3 - The credit card sized PC that cost only $35 - All-time bestselling computer in UK",
+ "SiteName": "SIMPLYFOUND.COM | BY: JOE WEE"
+}
diff --git a/test/test-pages/social-buttons/expected-metadata.json b/test/test-pages/social-buttons/expected-metadata.json
index 6505a03..7a80654 100644
--- a/test/test-pages/social-buttons/expected-metadata.json
+++ b/test/test-pages/social-buttons/expected-metadata.json
@@ -3,5 +3,6 @@
"Direction": null,
"Excerpt": "Lorem ipsum dolor sit amet, consectetur adipisicing elit, sed do eiusmod\n tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam,\n quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo\n consequat. Duis aute irure dolor in reprehenderit in voluptate velit esse\n cillum dolore eu fugiat nulla pariatur. Excepteur sint occaecat cupidatat non\n proident, sunt in culpa qui officia deserunt mollit anim id est laborum.",
"Image": null,
- "Title": "Share buttons removal test"
-} \ No newline at end of file
+ "Title": "Share buttons removal test",
+ "SiteName": null
+}
diff --git a/test/test-pages/style-tags-removal/expected-metadata.json b/test/test-pages/style-tags-removal/expected-metadata.json
index b014ce3..47039a9 100644
--- a/test/test-pages/style-tags-removal/expected-metadata.json
+++ b/test/test-pages/style-tags-removal/expected-metadata.json
@@ -3,5 +3,6 @@
"Direction": null,
"Excerpt": "Lorem ipsum dolor sit amet, consectetur adipisicing elit, sed do eiusmod\n tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam,\n quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo\n consequat. Duis aute irure dolor in reprehenderit in voluptate velit esse\n cillum dolore eu fugiat nulla pariatur. Excepteur sint occaecat cupidatat non\n proident, sunt in culpa qui officia deserunt mollit anim id est laborum.",
"Image": null,
- "Title": "Style tags removal"
-} \ No newline at end of file
+ "Title": "Style tags removal",
+ "SiteName": null
+}
diff --git a/test/test-pages/svg-parsing/expected-metadata.json b/test/test-pages/svg-parsing/expected-metadata.json
index 809c999..321414f 100644
--- a/test/test-pages/svg-parsing/expected-metadata.json
+++ b/test/test-pages/svg-parsing/expected-metadata.json
@@ -3,5 +3,6 @@
"Direction": null,
"Excerpt": "Lorem ipsum dolor sit amet, consectetur adipisicing elit, sed do eiusmod\ntempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam,\nquis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo\nconsequat. Duis aute irure dolor in reprehenderit in voluptate velit esse\ncillum dolore eu fugiat nulla pariatur. Excepteur sint occaecat cupidatat non\nproident, sunt in culpa qui officia deserunt mollit anim id est laborum.",
"Image": null,
- "Title": "SVG parsing"
-} \ No newline at end of file
+ "Title": "SVG parsing",
+ "SiteName": null
+}
diff --git a/test/test-pages/table-style-attributes/expected-metadata.json b/test/test-pages/table-style-attributes/expected-metadata.json
index 65d07c6..cdb8fc3 100644
--- a/test/test-pages/table-style-attributes/expected-metadata.json
+++ b/test/test-pages/table-style-attributes/expected-metadata.json
@@ -3,5 +3,6 @@
"Direction": null,
"Excerpt": "linux usability\n ...or, why do I bother. © 2002, 2003\n Jamie Zawinski",
"Image": null,
- "Title": "linux video"
-} \ No newline at end of file
+ "Title": "linux video",
+ "SiteName": null
+}
diff --git a/test/test-pages/telegraph/expected-metadata.json b/test/test-pages/telegraph/expected-metadata.json
index f2303e2..be19af1 100644
--- a/test/test-pages/telegraph/expected-metadata.json
+++ b/test/test-pages/telegraph/expected-metadata.json
@@ -3,5 +3,6 @@
"Direction": null,
"Excerpt": "Zimbabwe President Robert Mugabe, his wife Grace and two key figures from her G40 political faction are under house arrest at Mugabe's &quot;Blue House&quot; compound in Harare and are insisting the 93 year-old finishes his presidential term, a source said.",
"Image": "http:\/\/www.telegraph.co.uk\/content\/dam\/news\/2017\/11\/16\/TELEMMGLPICT000146889449-xlarge_trans_NvBQzQNjv4BqySoB6nTCgtc7U4LQ_FPO4hKi2sT3vi7ux2-RDZwC4QA.jpeg",
- "Title": "Zimbabwe coup: Robert Mugabe and wife Grace 'insisting he finishes his term', as priest steps in to mediate"
-} \ No newline at end of file
+ "Title": "Zimbabwe coup: Robert Mugabe and wife Grace 'insisting he finishes his term', as priest steps in to mediate",
+ "SiteName": "The Telegraph"
+}
diff --git a/test/test-pages/title-and-h1-discrepancy/expected-metadata.json b/test/test-pages/title-and-h1-discrepancy/expected-metadata.json
index 691b4d6..07338d0 100644
--- a/test/test-pages/title-and-h1-discrepancy/expected-metadata.json
+++ b/test/test-pages/title-and-h1-discrepancy/expected-metadata.json
@@ -3,5 +3,6 @@
"Direction": null,
"Excerpt": "Lorem\n ipsum dolor sit amet, consectetur adipisicing elit, sed do eiusmod\n incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam,\n quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo\n consequat. Duis aute irure dolor in reprehenderit in voluptate velit esse\n cillum dolore eu fugiat nulla pariatur. Excepteur sint occaecat cupidatat non\n proident, sunt in culpa qui officia deserunt mollit anim id est laborum.",
"Image": null,
- "Title": "This is a long title with a colon: Hello there"
-} \ No newline at end of file
+ "Title": "This is a long title with a colon: Hello there",
+ "SiteName": null
+}
diff --git a/test/test-pages/tmz-1/expected-metadata.json b/test/test-pages/tmz-1/expected-metadata.json
index 2614e96..63d6796 100644
--- a/test/test-pages/tmz-1/expected-metadata.json
+++ b/test/test-pages/tmz-1/expected-metadata.json
@@ -3,5 +3,6 @@
"Direction": null,
"Excerpt": "Lupita Nyong'o's now-famous Oscar dress -- adorned in pearls -- was stolen right out of her hotel room ... TMZ has learned. Law enforcement sources tell…",
"Image": "http:\/\/ll-media.tmz.com\/2015\/02\/26\/0225-lupita-nyongo-getty-01-1200x630.jpg",
- "Title": "Lupita Nyong'o's $150K Pearl Oscar Dress -- STOLEN!!!"
-} \ No newline at end of file
+ "Title": "Lupita Nyong'o's $150K Pearl Oscar Dress -- STOLEN!!!",
+ "SiteName": "http://www.tmz.com"
+}
diff --git a/test/test-pages/tumblr/expected-metadata.json b/test/test-pages/tumblr/expected-metadata.json
index 97b3aee..225236c 100644
--- a/test/test-pages/tumblr/expected-metadata.json
+++ b/test/test-pages/tumblr/expected-metadata.json
@@ -3,5 +3,6 @@
"Direction": null,
"Excerpt": "+ Added Granite, Andesite, and Diorite stone blocks, with smooth versions\n+ Added Slime Block\n+ Added Iron Trapdoor\n+ Added Prismarine and Sea Lantern blocks\n+ Added the Ocean Monument\n+ Added Red...",
"Image": "http:\/\/assets.tumblr.com\/images\/og\/fb_landscape_share.png",
- "Title": "Minecraft 1.8 - The Bountiful Update"
-} \ No newline at end of file
+ "Title": "Minecraft 1.8 - The Bountiful Update",
+ "SiteName": "Minecraft Update News"
+}
diff --git a/test/test-pages/wapo-1/expected-metadata.json b/test/test-pages/wapo-1/expected-metadata.json
index c7c946e..d3aad65 100644
--- a/test/test-pages/wapo-1/expected-metadata.json
+++ b/test/test-pages/wapo-1/expected-metadata.json
@@ -3,5 +3,6 @@
"Direction": null,
"Excerpt": "The assault on Tunisia’s most renowned museum, in which gunmen killed at least 19 people, could heighten tensions in a nation that has become deeply divided between pro- and anti-Islamist factions.",
"Image": "http:\/\/img.washingtonpost.com\/rw\/2010-2019\/WashingtonPost\/2015\/03\/18\/Foreign\/Images\/Nic6429927.jpg",
- "Title": "Attack stokes instability fears in North Africa"
-} \ No newline at end of file
+ "Title": "Attack stokes instability fears in North Africa",
+ "SiteName": "Washington Post"
+}
diff --git a/test/test-pages/wapo-2/expected-metadata.json b/test/test-pages/wapo-2/expected-metadata.json
index 87459c2..57f81c1 100644
--- a/test/test-pages/wapo-2/expected-metadata.json
+++ b/test/test-pages/wapo-2/expected-metadata.json
@@ -3,5 +3,6 @@
"Direction": null,
"Excerpt": "Few foreign leaders have so brazenly stood up to President Obama and the relationship could face its next test this month.",
"Image": "http:\/\/img.washingtonpost.com\/rw\/2010-2019\/WashingtonPost\/2015\/03\/18\/National-Economy\/Images\/Nic6429750-1138.jpg",
- "Title": "Where do strained U.S.-Israeli relations go after Netanyahu’s victory?"
-} \ No newline at end of file
+ "Title": "Where do strained U.S.-Israeli relations go after Netanyahu’s victory?",
+ "SiteName": "Washington Post"
+}
diff --git a/test/test-pages/webmd-1/expected-metadata.json b/test/test-pages/webmd-1/expected-metadata.json
index 146268d..2a7bced 100644
--- a/test/test-pages/webmd-1/expected-metadata.json
+++ b/test/test-pages/webmd-1/expected-metadata.json
@@ -3,5 +3,6 @@
"Direction": null,
"Excerpt": "Life-threatening peanut allergies have mysteriously been on the rise in the past decade, with little hope for a cure. But a groundbreaking new study may offer a way to stem that rise, while another may offer some hope for those who are already allergic.",
"Image": "http:\/\/img.webmd.com\/dtmcms\/live\/webmd\/consumer_assets\/site_images\/miscellaneous\/webmd_FB_logo_bluebkgrd.jpg?v=1",
- "Title": "Babies Who Eat Peanuts Early May Avoid Allergy"
-} \ No newline at end of file
+ "Title": "Babies Who Eat Peanuts Early May Avoid Allergy",
+ "SiteName": null
+}
diff --git a/test/test-pages/webmd-2/expected-metadata.json b/test/test-pages/webmd-2/expected-metadata.json
index 7d13fdc..9fee7bc 100644
--- a/test/test-pages/webmd-2/expected-metadata.json
+++ b/test/test-pages/webmd-2/expected-metadata.json
@@ -3,5 +3,6 @@
"Direction": null,
"Excerpt": "Drug-resistant bacteria, dubbed",
"Image": "http:\/\/img.webmd.com\/dtmcms\/live\/webmd\/consumer_assets\/site_images\/miscellaneous\/webmd_FB_logo_bluebkgrd.jpg?v=1",
- "Title": "Superbugs: What They Are and How You Get Them"
-} \ No newline at end of file
+ "Title": "Superbugs: What They Are and How You Get Them",
+ "SiteName": null
+}
diff --git a/test/test-pages/wikia/expected-metadata.json b/test/test-pages/wikia/expected-metadata.json
index f7557f2..ec9aa47 100644
--- a/test/test-pages/wikia/expected-metadata.json
+++ b/test/test-pages/wikia/expected-metadata.json
@@ -3,5 +3,6 @@
"Direction": null,
"Excerpt": "As a 40th birthday present to the Star Wars Saga and its fans, Lucasfilm could re-release the original versions of the original trilogy films.",
"Image": "https:\/\/vignette.wikia.nocookie.net\/1fb5ee36-d9ae-4125-96d9-f52eb403f1c9\/thumbnail-down\/width\/1280\/height\/720",
- "Title": "'Star Wars' Original Cuts Might Get Released for 40th Anniversary"
-} \ No newline at end of file
+ "Title": "'Star Wars' Original Cuts Might Get Released for 40th Anniversary",
+ "SiteName": "Fandom powered by Wikia"
+}
diff --git a/test/test-pages/wikipedia/expected-metadata.json b/test/test-pages/wikipedia/expected-metadata.json
index 4630de7..b6e3d76 100644
--- a/test/test-pages/wikipedia/expected-metadata.json
+++ b/test/test-pages/wikipedia/expected-metadata.json
@@ -3,5 +3,6 @@
"Direction": "ltr",
"Excerpt": "Mozilla is a free-software community, created in 1998 by members of Netscape. The Mozilla community uses, develops, spreads and supports Mozilla products, thereby promoting exclusively free software and open standards, with only minor exceptions.[1] The community is supported institutionally by the Mozilla Foundation and its tax-paying subsidiary, the Mozilla Corporation.[2]",
"Image": null,
- "Title": "Mozilla - Wikipedia"
-} \ No newline at end of file
+ "Title": "Mozilla - Wikipedia",
+ "SiteName": null
+}
diff --git a/test/test-pages/wordpress/expected-metadata.json b/test/test-pages/wordpress/expected-metadata.json
index f77c78a..9154ef0 100644
--- a/test/test-pages/wordpress/expected-metadata.json
+++ b/test/test-pages/wordpress/expected-metadata.json
@@ -3,5 +3,6 @@
"Direction": null,
"Excerpt": "Stack Overflow published its analysis of 2017 hiring trends based on the targeting options employers selected when posting to Stack Overflow Jobs. The report, which compares data from 200 companies…",
"Image": "https:\/\/i0.wp.com\/wptavern.com\/wp-content\/uploads\/2016\/07\/stack-overflow.png?fit=1200%2C470&ssl=1",
- "Title": "Stack Overflow Jobs Data Shows ReactJS Skills in High Demand, WordPress Market Oversaturated with Developers"
-} \ No newline at end of file
+ "Title": "Stack Overflow Jobs Data Shows ReactJS Skills in High Demand, WordPress Market Oversaturated with Developers",
+ "SiteName": "WordPress Tavern"
+}
diff --git a/test/test-pages/yahoo-1/expected-metadata.json b/test/test-pages/yahoo-1/expected-metadata.json
index abab3bc..cfd3a69 100644
--- a/test/test-pages/yahoo-1/expected-metadata.json
+++ b/test/test-pages/yahoo-1/expected-metadata.json
@@ -3,5 +3,6 @@
"Direction": null,
"Excerpt": "To help you decide what’s what, I’ve put together this list of the 8 PSVR games worth considering. Beloved cult hit “Rez” gets the VR treatment to help launch the PSVR, and the results are terrific. Chaos, for sure, and also “Thumper.” Called a “violent rhythm game” by its creators, “Thumper” is, well",
"Image": "http:\/\/l3.yimg.com\/uu\/api\/res\/1.2\/4eRCPf9lJt_3q29.outekQ--\/aD02Njk7dz03NDQ7c209MTthcHBpZD15dGFjaHlvbg--\/http:\/\/media.zenfs.com\/en\/homerun\/feed_manager_auto_publish_494\/4406ef57dcb40376c513903b03bef048",
- "Title": "These are the 8 coolest PlayStation VR games"
-} \ No newline at end of file
+ "Title": "These are the 8 coolest PlayStation VR games",
+ "SiteName": null
+}
diff --git a/test/test-pages/yahoo-2/expected-metadata.json b/test/test-pages/yahoo-2/expected-metadata.json
index 4fe290c..8ca33da 100644
--- a/test/test-pages/yahoo-2/expected-metadata.json
+++ b/test/test-pages/yahoo-2/expected-metadata.json
@@ -3,5 +3,6 @@
"Direction": null,
"Excerpt": "The latest news and headlines from Yahoo! News. Get breaking news stories and in-depth coverage with videos and photos.",
"Image": "https:\/\/s.yimg.com\/os\/mit\/media\/m\/social\/images\/social_default_logo-1481777.png",
- "Title": "Yahoo News - Latest News & Headlines"
-} \ No newline at end of file
+ "Title": "Yahoo News - Latest News & Headlines",
+ "SiteName": null
+}
diff --git a/test/test-pages/yahoo-3/expected-metadata.json b/test/test-pages/yahoo-3/expected-metadata.json
index 8b8a5c7..8af68c0 100644
--- a/test/test-pages/yahoo-3/expected-metadata.json
+++ b/test/test-pages/yahoo-3/expected-metadata.json
@@ -3,5 +3,6 @@
"Direction": null,
"Excerpt": "A photographer and Navy veteran is fighting back after a photo she posted to Facebook started an online backlash. Vanessa Hicks said she had no idea her photo would be considered controversial. The photo, from a military family’s newborn photo shoot, showed a newborn infant wrapped in an American flag held by his father, who was in his military uniform. Hicks, a Navy veteran herself and the wife of an active-duty Navy member, said her intention was to honor the flag as well as her clients, who wanted to incorporate their military service in the photo shoot.",
"Image": "https:\/\/s.yimg.com\/bt\/api\/res\/1.2\/qZaM9MLUOrxLg4IfXt_Niw--\/YXBwaWQ9eW5ld3NfbGVnbztxPTc1O3c9NjAw\/http:\/\/media.zenfs.com\/en-US\/video\/video.abcnewsplus.com\/559ecdbafdb839129816b5c79a996975.cf.png",
- "Title": "Veteran Wraps Baby in American Flag, Photo Sparks Controversy"
-} \ No newline at end of file
+ "Title": "Veteran Wraps Baby in American Flag, Photo Sparks Controversy",
+ "SiteName": "Yahoo"
+}
diff --git a/test/test-pages/yahoo-4/expected-metadata.json b/test/test-pages/yahoo-4/expected-metadata.json
index 707b15b..c8459ad 100644
--- a/test/test-pages/yahoo-4/expected-metadata.json
+++ b/test/test-pages/yahoo-4/expected-metadata.json
@@ -3,5 +3,6 @@
"Direction": null,
"Excerpt": "トレンドマイクロは3月9日、Wi-Fi利用時の通信を暗号化し保護するスマホ・タブレット - Yahoo!ニュース(CNET Japan)",
"Image": "http:\/\/i.yimg.jp\/images\/jpnews\/cre\/common\/all\/images\/fbico_ogp_600x600.png",
- "Title": "トレンドマイクロ、公衆無線LANを安全に使うためのアプリ「フリーWi-Fiプロテクション」(CNET Japan) - Yahoo!ニュース"
-} \ No newline at end of file
+ "Title": "トレンドマイクロ、公衆無線LANを安全に使うためのアプリ「フリーWi-Fiプロテクション」(CNET Japan) - Yahoo!ニュース",
+ "SiteName": "Yahoo!ニュース"
+}
diff --git a/test/test-pages/youth/expected-metadata.json b/test/test-pages/youth/expected-metadata.json
index 17c34d6..f84c6ec 100644
--- a/test/test-pages/youth/expected-metadata.json
+++ b/test/test-pages/youth/expected-metadata.json
@@ -3,5 +3,6 @@
"Direction": null,
"Excerpt": "图为马素湘在澳大利亚悉尼游玩时的近影。出国前后关注点大不同出国前:政治科目会出啥考题?出国后:国家未来将如何发展?在采访中,我们了解到不少学子在出国前就每年守在电脑前观看两会直播。但是,随着年龄和阅历的增长,学子对两会的关注点在出国前后发生了很大的变化。在法国里昂国立应用科学院留学的卢宇表示,他还是个中学生时,就开始关注两会了。“我高中毕业后就出国留学了。",
"Image": null,
- "Title": "海外留学生看两会:出国前后关注点大不同_教育频道_中国青年网"
-} \ No newline at end of file
+ "Title": "海外留学生看两会:出国前后关注点大不同_教育频道_中国青年网",
+ "SiteName": null
+}