Some CNN personalities are not being filtered

[ericaryan]

There are a few CNN personalities who are especially likely to sneak through into the Alerts based on my review. They include:

JIM SCIUTTO, CNN ANCHOR (example transcript: http://transcripts.cnn.com/TRANSCRIPTS/1909/09/cnr.01.html) DAVE BRIGGS, CO-ANCHOR, EARLY START (example transcript: http://transcripts.cnn.com/TRANSCRIPTS/1909/09/es.03.html) CHRISTIANE AMANPOUR, CHIEF INTERNATIONAL CORRESPONDENT (example transcript: http://transcripts.cnn.com/TRANSCRIPTS/1909/09/ampr.01.html)

The above first references are consistent across the transcripts I've checked for these three people when this issue has cropped up.

A couple of others that have made it through: POPPY HARLOW, CNN ANCHOR (http://www.cnn.com/TRANSCRIPTS/1908/15/cnr.09.html) ZAIN ASHER, CNN INTERNATIONAL HOST (http://transcripts.cnn.com/TRANSCRIPTS/1908/30/qmb.01.html)

[ericaryan]

Tech & Check Alerts: CNN, Twitter 09/11/19

RICHARD QUEST on QUEST MEANS BUSINESS (CNN): If you look again, it's actually not done that badly since the beginning of the year, from 28.54 up to 36.18, it's a United States dollar premium, 30 percent on the original.

http://transcripts.cnn.com/TRANSCRIPTS/1909/09/qmb.01.html

First ref: RICHARD QUEST, CNN INTERNATIONAL HOST, QUEST MEANS BUSINESS

[reefdog]

Here's the list of CNN folks the old system was specifically filtering out.

I wonder if this should be a function of the scraper or the newsletter composer? In a misty future where subscribers are subscribing to individual speakers, will we want to give them the option of subscribing to these speakers? (Not that the answer to that is determinative. We can filter on speakers now and then lift that filter when we add individual newsletter composition later.)

[ericaryan]

Can I ask how you are currently filtering CNN personalities?

[slifty]

If CNN is in the person's affiliation we remove them -- we don't have a list of specific names (I would love to avoid that but understand if we can't avoid it)

[ericaryan]

Thanks! My intent with opening this issue was to address instances where people are slipping through the existing filtering. So if there's a way to improve it to catch the ones above (a few of whom do already have CNN in their affiliations), then that's preferable to me too. "Anchor" and "correspondent" seem like good additional terms to me?

[slifty]

Those terms sound quite reasonable to add in! And this is a good issue to have open since those that slip through the cracks are indeed a bug. I may modify the title to reflect that the feature does exist already and this is a requested bug fix / enhancement.