Release Universal Scraper for Music Artists
#82
(2012-07-20, 22:45)olympia Wrote:
(2012-07-20, 22:23)deh2k7 Wrote: The Red Hot Chili Peppers (searched if I omitted the 'The')
The Pusscat Dolls (searched if I omitted the 'The')
Motley Crue (couldn't get it search, but it has the umlauts above the 'o' and the 'u' in my library. Returned no results, but also did not give me a chance to correct it)
DJ Keoki (scraped if I just searched for Keoki)
Sorry to say, but did you actually look up these on MB before you waste my time with them?
How the scraper would find these if they are named differently there (with the only exception of Motley Crue)?

(2012-07-20, 22:23)deh2k7 Wrote: I'm guessing maybe the scraper is overloading the request source and we're getting timed out or something?
Yes, this is a known issue with MusicBrainz. XBMC is scraping faster than MB would allow.

Olympia - I appreciate the help and all of the work that you have done to date, but response seems a bit harsh, as I have been seeing some legitimate issues and it's hard to tell which issues are known and which are not, or which are non-issues and simply how the scraper works. In each case above, with or without the 'The' Musicbrainz produces a list of search results with each artist appearing with a 100 match score AND the first entry. I would think that the scraper would select these automatically as it seems to do in most cases.

The next closest score was below 50 in ALL cases (21,33,43). It's not like it was a 98 vs 100. With the Chili Peppers, searching with and without the leading 'The' still produces a list of results and has the Chili Peppers as #1, and with a 100% score. I'm not sure why or how it's wasting your time to raise a relatively common use case where the scraper fails to match an artist without manual intervention, when it could easily accept the highest scoring result (or the first one in the list in the case of a near match). This also would have addressed the Motley Crue scraping issue as it matches with or without the umlauts in the first position with a 100 match score.

I was also under the assumption that the scraper would be a bit less sensitive to non-specific leading articles, like 'the' or 'a' (thinking A Perfect Circle as an example), just in case.

I would respectfully ask that you consider at least the Musicbrainz match score or result list when matching artists to facilitate a more streamlined automated scraping process.
Reply


Messages In This Thread
RE: [Release] Universal Scraper for Music Artists - by deh2k7 - 2012-07-21, 04:09
Doesn't work at all - by LastCoder - 2013-03-10, 22:05
Logout Mark Read Team Forum Stats Members Help
Universal Scraper for Music Artists9