Sanitize JSON-LD input before decoding

[rvanlaak]

This PR relies on #38 getting merged, as commit e4e9a18 is the commit that fixes the bug

When the Json ends with a ; the json can not get decoded.

Example

WHO content ends with an ;. As result of that the NewsArticle can not get parsed.

https://www.who.int/news-room/detail/17-10-2019-7-million-people-receive-record-levels-of-lifesaving-tb-treatment-but-3-million-still-miss-out

How to test

Run while parsing above URL

Or test the following JSON snippet on https://jsonlint.com/#:

{"author":{"@type":"Organization","name":"World Health Organization: WHO","logo":{"@type":"ImageObject","url":"http://www.who.int/Images/SchemaOrg/schemaOrgLogo.jpg","width":250,"height":60}},"headline":"7 million people receive record levels of lifesaving TB treatment but 3 million still miss out","description":"<p>More people received life-saving treatment for tuberculosis (TB) in 2018 than ever before, largely due to improved detection and diagnosis. Globally, 7 million people were diagnosed and treated for TB - up from 6.4 million in 2017 &ndash; enabling the world to meet one of the milestones towards the United Nations political declaration targets on TB. </p> [...] <p>Brazil, China, the Russian Federation and Zimbabwe, which all have high TB burdens, achieved treatment coverage levels of more than 80 per cent.</p>","datePublished":"2019-10-17T22:01:00.0000000+00:00","image":"https://www.who.int/images/default-source/wpro/news/feature-stories/cambodia-tb-screening.jpg?sfvrsn=e7bbe3fb_20","publisher":{"@type":"Organization","name":"World Health Organization: WHO","logo":{"@type":"ImageObject","url":"http://www.who.int/Images/SchemaOrg/schemaOrgLogo.jpg","width":250,"height":60}},"dateModified":"2019-10-17T22:01:00.0000000+00:00","mainEntityOfPage":"https://www.who.int/news-room/detail/17-10-2019-7-million-people-receive-record-levels-of-lifesaving-tb-treatment-but-3-million-still-miss-out","@context":"http://schema.org","@type":"NewsArticle"};

How to fix

check if the last character is a ;, and strip that

[coveralls]

Coverage increased (+0.0003%) to 99.555% when pulling e4e9a18 on rvanlaak:patch-1 into eb7d515 on jkphl:master.

[jkphl]

@rvanlaak Thanks!