Parenting & md5 questions

There is several mentions on here of "dont upload full manga/doujin unless you have a really good reason", but little on the correct parent-> child sequence.

1. should (short) doujin uploads be parented in sequence (1,2,3,4) or all from page 1? Should pools be used for such short uploads? The specific case i'm querying of is post #52905, which I have the full 4 pages of.

2. Would it be possible to get a copy of the filename (or md5, whichever way round it works) table, so that I can perform automated duplicate checking on my systems? I'd rather not waste danbooru bandwidth and my hdd space on the rape script to get them all manually.

love

r.

1. Doujins should no longer be parented, instead make a pool for them. See forum #8674. Parenting should be reserved for variations of a pic, and because it no longer hides the child posts in the main index it is pointless to use it for whole comics. Its main use is to allow people who found one pic to quickly see the other variations of it.

2. I have a copy of the MD5 list for use in IQDB, so I could make that if nobody objects.

thank you on both counts.

Doujins should only be uploaded if each page is of comparable quality to an acceptable upload.

I'm very interested in the md5 list as well.

The md5 is included in the result from http://danbooru.donmai.us/post/index.xml

zatchii said:
The md5 is included in the result from http://danbooru.donmai.us/post/index.xml

Yes, but do you have any idea just how many queries it would take to get the MD5 of ALL posts on Danbooru?

In any case I've exported post ID and MD5 now. It's at http://iqdb.org/danbooru_md5.txt.bz2 - though some of these will probably have been deleted now, I haven't run the deletion check script recently since deleted posts still contain useful information now.

In a duplicate check, deleted MD5's should probably be included anyway!

The simple deletes will be checked of course, but if a post has been removed from the database entirely, then not even the MD5 will be remembered here. In that case even the deletion record itself is deleted so my script won't ever find out about this.

I don't know how common that is nowadays though.

http://iqdb.org/danbooru_md5.txt.bz2

I've often wished for this. Thanks! Will it be automatically updated? Would it be worth hosting an uncompressed version so people can just grab the tail with 'wget --continue'?

legga said:
Will it be automatically updated?

No. But bzip2 archives are appendable so wget --continue would still work when I do update it. I suppose I could make this a weekly thing if there's demand for it.

I don't think there'd be much need for more than a bimonthly update if you feel so inclined; the bulk of backdated duplicates are dealt with after all.

But thanks again, this saves a lot of time.