grmwld / scholar_bot Goto Github PK

Basically, check that the number of comments is 0

If several links are provided and one of them fails to point to a pdf, the share is destroyed.

• authenticate
• search for "full text source"
• follow link
• search for pdf

In dry mode, link are not posted to reddit.

All rate limits are almost eliminated, providing an easy way to test the behavior

example:

http://www.sciencemag.org/content/341/6145/526.abstract

fetches the pdf at the bottom, which are suggested articles.
The true PDF lies in the sidebar

A generic PDF retriever is almost impossible to write, and sometimes results in the wrong PDF being fetched if several links are available on the same page (see issue #7) .

As much as possible, specific PDF retriever should be written for common websites/journals, and fallback to the generic one otherwise.

Here is a list of specific websites that come to my mind:

• Nature
• Science
• Cell (PDFs are usually available on sciencedirect)
• ScienceDirect