HavocXphere
Honorary Master
I buy pretty much everything except perishable food off Amazon, so I've got price alerts set up on camelcamelcamel.
Problem is lately I'm getting alerts where by the time I get home the price has been pushed up again. So sellers are obviously using a bot and possibly even intentionally triggering alerts to get me to look at their page. Look at these fkin shenanigans:

I'd say 15 is about fair market price in the above example.
A watchlist is no longer cutting it & thinking of setting up a buying bot. And if you check the price differential on the above example we're talking real money savings if I can catch the price on the bottom end.
So far so good.
Part where I need advice from the programming crew:
Scrape Amazon directly or rely on camel's emails as a trigger? Camel is probably programatically easier, but bit of research shows it introduces a 2 hour delay into the equation (their scraping cycle). Neither has an API.
If I scrape directly what are the chances of Amazon nix'ing that as malicious web traffic?
Given that this thing will be armed with a credit card how would you go about doing this somewhat safely?
(Aside from the obvious if toothpaste bought don't buy more type code).
Current thinking is Selenium via Python running on a raspberry pi that is on 24/7. And probably hourly scraping over say 40 products.
Thoughts?
Problem is lately I'm getting alerts where by the time I get home the price has been pushed up again. So sellers are obviously using a bot and possibly even intentionally triggering alerts to get me to look at their page. Look at these fkin shenanigans:

I'd say 15 is about fair market price in the above example.
A watchlist is no longer cutting it & thinking of setting up a buying bot. And if you check the price differential on the above example we're talking real money savings if I can catch the price on the bottom end.
So far so good.
Part where I need advice from the programming crew:
Scrape Amazon directly or rely on camel's emails as a trigger? Camel is probably programatically easier, but bit of research shows it introduces a 2 hour delay into the equation (their scraping cycle). Neither has an API.
If I scrape directly what are the chances of Amazon nix'ing that as malicious web traffic?
Given that this thing will be armed with a credit card how would you go about doing this somewhat safely?
Current thinking is Selenium via Python running on a raspberry pi that is on 24/7. And probably hourly scraping over say 40 products.
Thoughts?