Defaults to INFO, both on read (to make sure that old setups without the value in config.yaml don’t fail) and generation of new config file.
Sets the peewee logger to WARNING

Allows for a regex to be applied to each article url when first encountered in a feed.
Born from noticing visir.is editing errors in headlines by changing the url.

Example:

http://www.visir.is/g/2017170629354/glaepagengi-i-arhusum-ogna-oryggi-almennings

Changed to:

http://www.visir.is/g/2017170629354/glaepagengi-i-arosum-ogna-oryggi-almennings

With a regex that only captures ‘http://www.visir.is/g/2017170629354’ (which is enough for the article to be displayed) then such a change is picked up.

Example config:

urlregex: (.*\/g\/\d*)

Allows for a regex to be applied to each article url when first encountered in a feed.
Born from noticing visir.is editing errors in headlines by changing the url.

Example:

http://www.visir.is/g/2017170629354/glaepagengi-i-arhusum-ogna-oryggi-almennings

Changed to:

http://www.visir.is/g/2017170629354/glaepagengi-i-arosum-ogna-oryggi-almennings

With a regex that only captures ‘http://www.visir.is/g/2017170629354’ (which is enough for the article to be displayed) then such a change is picked up.

Example config:

urlregex: (.*\/g\/\d*)

Squashing stray print statement removal commit

Instead of one

Prompted by visir.is returning bogus text when True:

https://twitter.com/visir_diff/status/879553652842745857

Prompted by visir.is inserting them in headlines:

https://twitter.com/visir_diff/status/879742393389547520

See: https://github.com/buriy/python-readability/blob/master/readability/readability.py#L88

Supply a list of classes that are removed with lxml prior to readability doing it’s parsing.
It’s messy but will suffice for now.
Prompted because readability does not properly parse pages like this:

http://bleikt.pressan.is/lesa/dottir-kallar-mommu-sina-feita-svar-modur-hennar-hefur-vakid-mikla-athygli/

(It Is picking up excerpts at the bottom - with the config: ‘kill_classes: ["articles-with-excerpt”]’ then we are fine)

Similar to kill_classes, but uses lxml’s ‘drop_tag’ that drops the tag, but keeps its children and text.

Currently used to collapse div in ruv_diff