search

wellcometrust / reach

Wellcome tool to parse references scraped from policy documents using machine learning
MIT License
25 stars 4 forks source link

Output file not linking correctly in deployed code #20

Closed lizgzil closed 5 years ago

lizgzil commented 5 years ago

In parser_references>reference the id_document does not seem to be correct as it links to a predicted title (in parser_references>document.id) which is nothing like the actually title. Otherwise the output match info (the file hash and pubmed id) seem to be correctly matched.

This situation does not happen when the code is run locally.

LOCALLY RUN OUPUT: file hash: e54e005e8f9f736189e52fbd03a3d9f2 Predicted title: (2006) Does breastfeeding influence risk of type 2 diabetes in later life?, A quantitative analysis of published evidence Actual title: Does breastfeeding influence risk of type 2 diabetes in later life? A quantitative analysis of published evidence. Pubmed id: pub.1077320364

RDS OUTPUT: file hash: e54e005e8f9f736189e52fbd03a3d9f2 Predicted title: and cardiovascular disease: metaanalysis of prospective studies Actual title: Does breastfeeding influence risk of type 2 diabetes in later life? A quantitative analysis of published evidence. Pubmed id: pub.1077320364

lizgzil commented 5 years ago

I've checked and this is still an issue (although this particular example is no longer in the matched data).

lizgzil commented 5 years ago

screenshot 2019-01-07 at 16 18 31

SamDepardieu commented 5 years ago

We're not using RDS anymore and the ouputs have completely changed. Closing this.