Add user agent string to feed requests

[mattbasta]

Hey folks, I work on Pinecast (https://pinecast.com). We recently had a customer write in that their audiobookshelf server stopped updating feeds from us. It turns out that we'd seen a huge uptick in requests from the default Axios user agent string, which caused requests with that user agent to get served a captcha. My suspicion is that companies have started scraping more aggressively.

Podcast apps have (by and large) done a good job of providing descriptive User-Agent headers. Identifying your app allows podcasters to get informed analytics data and helps hosting companies like mine to identify and fix problems more easily.

---

In this PR, I'm adding the user-agent HTTP request header. I've chosen a value that is fairly straightforward: the name of the project, a link to the project, and like iTMS to indicate that the client should be treated similarly to Apple's podcast crawler. I do not have strong feelings about this UA string, but I'll suggest that if you decide to go with something else, you avoid adding a version number. There's a balance between user privacy (less information) and transparency (more descriptive information) and compatibility (more information that looks like another user agent).

This PR does not update the client calls to download media. I am not too familiar with the project and do not want to regress any features. However, I strongly suspect this function will need a similar update:

https://github.com/advplyr/audiobookshelf/blob/master/server/utils/fileUtils.js#L254

I'm glad to make that change as well, if you think it's appropriate.

[advplyr]

Thanks! The fileUtils downloadFile is used for downloading podcast episodes, podcast cover images and audiobook cover images.
I'm not sure about the like iTMS part, is that specific to Pinecast and what would the difference be if that wasn't included?

[mattbasta]

It's not specific to Pinecast, no. iTMS is the UA string that Apple uses when crawling feeds and episodes. Many UAs follow that pattern (you'll see "like Gecko" for things that want to advertise compatibility with Firefox or "like FeedFetcher-Google" for things that pull RSS).

Like I said, I don't have strong feelings either way, but my intuition is that it'll help in cases where folks are doing naive UA detection. I'm glad to update it to your liking, and also update the second function.

[advplyr]

If like iTMS was included in the request for the RSS feed but was not included in the request for the files would that make sense? I'm not sure if it is better to be consistent across all requests.

[mattbasta]

It could be excluded for the asset requests. That's sensible

[advplyr]

I updated the site to https://audiobookshelf.org instead of the repo because the repo is going to move to the audiobookshelf org eventually. Thanks for the help!