Author Topic: The PORN man, the porn.  (Read 4605 times)

Offline nrk666

  • Junior Indexer
  • **
  • Posts: 41
  • Helpful: +0/-0
The PORN man, the porn.
« on: 2017-06-18, 10:49:55 am »
a.b.erotica posts are using some new form of obfuscation. Is this something we can expect nZEDb to adapt to or are we left to our own devices on this? I realize the age old issue of "if everyone knows how to decrypt it, whats the point" but I'd think this was a fairly central function for an indexer. I've already got par2/nfo/decrypt all turned on, but I'm not even collecting the new releases in a.b.e except for a few oddballs, much less post processing proper names using the latest release in master.

Any suggestions? (Collection regex, decrpyting, etc?)


Offline nrk666

  • Junior Indexer
  • **
  • Posts: 41
  • Helpful: +0/-0
Re: The PORN man, the porn.
« Reply #1 on: 2017-07-01, 10:49:30 am »
nvm, figured it out. No one coulda just said "renametopre.php" ?


Offline slayer99

  • Junior Indexer
  • **
  • Posts: 9
  • Helpful: +0/-0
Re: The PORN man, the porn.
« Reply #2 on: 2017-07-01, 11:34:45 am »
The odder issue to me is that it's not renaming based on the filename (par2 has the info).  The rename script is running...hell, I've run it with options to rename by filename...yet the name in the db remains the same.  I don't have this issue in other groups.

Offline nrk666

  • Junior Indexer
  • **
  • Posts: 41
  • Helpful: +0/-0
Re: The PORN man, the porn.
« Reply #3 on: 2017-07-28, 03:40:55 pm »
WARNING : I am not an expert on nZEDb or MySQL so all of this might be a totally fucked way of doing anything and I should probably kill myself from the shame of doing this wrong.

So, from observation poking around the PreDB table inside the database I noticed that most porn titles pre'd (like 95% of them) include the filenames in the predb.filename, but WITH the vowels. So, I wrote a perl script (I don't know PHP, don't really want to learn PHP, so there is probably a much better way to do this, but I used what I know) to search the predb table for releases >= X hours (command line arg) for entries where predb.category LIKE %XXX% or NULL, checked to see if predb.filename exists, then I strip out the vowels and compare it to entries in the releases.searchname (around the same time frame or later) and if I find a match, I manually update the releases.searchname and predb_id entries for that release using the predb.title and predb.id. I also set it to categorized = 0 and finish with a run of "php recategorize.php" to make sure it follows my categorization regexs. I had to make a few exceptions for a couple different group releases (mangling the filenames a bit) and for the most part, now have a index of porn with accurate scene predb searchnames again. I run the script every 15 minutes or so looking back like 4-8 hrs, do a big one in the middle of the night for 48hrs just to catch any stragglers, and its nearly perfect. I'm not super excited about posting my script (its a total kludge) but if you want to ask any questions, I can tell you how I handled certain things.

The only thing that is still totally fucked is the PACK collections. I seem to collect them all individually instead of in a Part1 or Part2 collection release after monkeying with the collection regex for packs a bit. So, each single "title" gets its search name set to the PACK requestid title, so I have like 40-50 unique release.id's for "Pornstart.Part1.PACK.XXX-U4A". If anyone knows how to start hacking on that, or how pack collection works (I've studied the collection regex and the new naming scheme make identifying the collections harder because things are out of order, I don't know how to hack up a regex to put it in the old order.) and can point me in a direction, I'd probably take a stab at that too.

I realize that an out of band script doing direct edits on the DB is probably a colossally stupid thing to do, but this is a personal indexer that only me and a few friends use, so just keep calm I guess. I'm just hacking on it for fun.