About IbSearch

First version was launched back in November 2009. For a year it was operating on Pentium II 367,50 MHz with 256 DIMM RAM handling about 2 million images. It was largely abandoned after reaching 5 million posts in May 2011 (those were duplicate accross indexed boorus, not separate images with list of sources as now). During this period it ran on the single server in my bathroom (mind you, it was upgraded to Pentium-D 3 GHz with 1 GiB RAM in mid-2010).

In March 2015 I suddenly decided to revamp the entire thing – something I have been planning for years and have almost started a few times. This attempt turned to be a success and now you see what has become of it. IbSearch has moved to its own domain – or rather two (.ch and .xxx).

Unlike previous two versions, IbSearch 3 runs on whole lot of servers including several dedicated robots that automatically crawl various boorus adding, updating, deleting and linking together images that are both old and just uploaded. Having such an unattended crawler was my dream from day one. Dreams come true, you see.

Oh, and since this swarm of servers can no more fit in my bathroom (although I could use their heat in the winter) they live in a neightboor data center.

Because IbSearch is my hobby project I would very much like to hear what you have to say – cheers, feedback, suggestions, bug reports. Say Hi!, the form is right there

How it was

19.04.2015 Version 3 released
14.03.2015 Version 3 development
7.05.2011 Fifth million
19.08.2010 Third million
28.06.2010 Pentium-D 3 GHz
7.05.2010 First million indexed
30.04.2010 Brand old Pentium II
17.04.2010 Version 2 released
30.09.2009 First version launch


with priority (+/-/0):
8 fix mod's -NOT for spx
8 fix trunc*
5 does ign_chars make sense - sphinx tags are normalized? search for "foo!" becomes fooX00 bypassing ign_c
5 put search history, hits, online, tag stats, etc. to mmc; push to db in bg thread
5 use sphinx tags index stats instead of db tags table?
5 show kw aliases/total results from META
5 fix w/h of ani GIFs (0x0, 1x1 or 1 px less than original w/h)
4 cli min - get latest from db, may not be present in sphinx index yet
4 cli min - track last run (no parallels) and adjust delay
4 no db calls for: /, /images/?q=, /images/123
4 mmc for all bare-tc queries
4 like:ibid[,ibid,...][=num_tags_(8)][/quorum]
4 show unblock.ibsear.ch headsup for proxies
4 sq: reindex mismatching tags on kona
4 fix infoURL not defined on SPA
3 implied mods (.ch/.xxx rating, default tseg sort) not shown in stdTitle
3 slideshot & hotkeys (see msgs)
2 hili/categorize tags as on dan/gelbooru
2 HR rate tags
2 HR review synonyms
2 HR logo
2 HR review rating of site:r34,x
1 PR
1 mobile version (bye bye Ani)
1 sq: in-memory db for queue
1 api/images: sources=1, relations=X - fetch joins after id list (?)
1 add note about max pages (max_matches); new max:N mod for spx
JSONP API format?
automatize refetch broken images
query modifier for image props (ani GIF, alpha PNG)
fix auto scripts on static after changing file struct (+ ignore if file md5 is as expected)
resizable thumbs
comments + sync from boorus
fav image streams
swf, webm
rating:p (popularity)
is max_results=1500 enough? collect stats
sort:tsegs; default sort:tsegs,weight
send admin pm form
no resize url for ani gifs
rating is not filtered in similar images: https://ibsear.ch/images/22894
fix alpha on thumbs (https://ibsear.ch/images/5885649)
-1 search's RSS
-1 index booru.touhouradio.com, rule34.paheal.net
-1 directory of boorus: http://pastebin.com/HNp6NuSA
-3 enqueue images from IbS v2 when have all sqsrs
-3 check archived to-do's

While you're on it…

  1. oh and uh coldtoon sent ya to animebox.es
  2. ok so you need mobile? then it may be wise to invest with the animebox.es team they have done high level work and still are. please i would love to see you two team up
  3. What about indexing chan.sankakucomplex? Some pics are there but not on the other sites.