intelligent noise filter, metadata trim and blacklist
djryan opened this issue · 13 comments
What version of SLS?
v1.4.7 ?
Latest 1.4.8 beta from here.
The old version from the play store didn’t pay attention to the extra data.
hmm... this is a tough one.
I can't really be sure what to do about this one.
I would need to apply really specific code.
Such as for podcast addicts only and for this particular use case.
Regex or removing everything after the dot would work.
I can't really apply such a thing unless it's a universal certainty.
This falls under the category of intelligent noise filters.
I can't yet decide how I will implement this.
I'll keep it on my todo list however.
Would an interim, but slightly more general solution, be to have rules allowing regex? I’d be happy to put something like (*.) ⚫️*.
converts to /1
Sounds good.
I will try to implement something or see if it's already been done before, as soon as I can.
@djryan what do you think of just isolating parts?
For example you could remove the 'dot' and the date separately and just trim the white space off the ends?
I could add a simple check box page for special characters and another for their contents maybe.
Examples would be:
special characters and square brackets and optionally their contents,
various date formats,
I could add a developer option to enable regex expressions.
There's already a database export developer option.
Oh that’s cute! I’d have been happy taking all the pain on myself (bringing back fond memories of my Perl development past) but I suspect I’d be in the minority!
Just supporting regex (#560) would be good enough - it's probably too much to hardcode intelligent filtering
Hi,
Has anyone figured out how to remove the date from the artist's name when scrobbling from Podcast Addict?
Thanks!
Apart from editing in SLS, no. I did request that the author of Podcast Addict made the feature configurable, but they were adamant it was never going to change.
Yes agreed.
For those interested, in the meantime, I rolled back to version 1.5.7 where the date issue is not present.