JAV database

[cyberzen]

There are lots of web hosting companies that allow porn, I used to host hentai websites. Very easy to get a server with adult content. I don't mind paying for it, BUT obviously not out of my own pocket, you either need advertising / charge for app / donations. I'm pretty sure people will buy an application like this.

It's very hard to have a good performing java GUI. I work with eclipse everyday, and I hate it. Have you tried QtCreator, it's awesome, ui designer, quick intellisense, and I don't need 8gb of ram to run it.

 

[cyberzen]

The basic app should be free, but with some ads included, if a user buys the app ($5-$10) then the ads go away and the user gets access to more advanced features. Follow the android app freemium model.

 

[CodeGeek]

There are lots of web hosting companies that allow porn, I used to host hentai websites. Very easy to get a server with adult content. I don't mind paying for it, BUT obviously not out of my own pocket, you either need advertising / charge for app / donations. I'm pretty sure people will buy an application like this.

Okay, I only looked for hosters like that one time for a few hours, but didn't find any. So I thought there isn't one. Maybe I searched the wrong way for it.

It's very hard to have a good performing java GUI. I work with eclipse everyday, and I hate it. Have you tried QtCreator, it's awesome, ui designer, quick intellisense, and I don't need 8gb of ram to run it.

I also work with Eclipse everyday - in the company as well as at home. And I think the GUI is very responsive. I never had a problem. Means: If you don't install any 3rd party plugins. I have to use the Microsoft Team Foundation Server plugin at work. And that really pisses me off. Sometimes the GUI doesn't react at all or takes minutes to perform an action thanks to that plugin.

I also used IBM Lotus Notes 8.0 / 8.5 in the past which also makes use of the Eclipse client / the SWT. And it was also fine to work with it, no problems at all.

8 GB RAM? I never have seen that Eclipse takes so much memory. What kind of plugins do you use? But I have to admit that it takes more memory than other editors do. E.g. Lotus Notes already needed a lot of memory before. But after they switched the client to Eclipse / SWT you needed at least 1 GB RAM for a normal user PC. I as developer needed even more.



But back to the topic: Maybe we can work together in aspect of the content of the database.

I also plan to offer some kind of web service interface. So also 3rd party programs can access the client database. And because it is a local database you can do as many operations on it as you want. E.g. you can use it as a backend for a website. This is not possible if everyone is accessing the same database on the web (like in the case of AniDB.net).

 

[cyberzen]

I do android development, I had to upgrade my laptop from 4gb of ram to 8gb because eclipse would be so freaking slow and kept getting low memory. QtCreator only takes 180mb ram.

Do you mean sharing tags, and actress for the database? We will need to agree on how to identify a video. The simple way would be by video title, for example NHDTA-403. In the future we should have a more robust way of identifying videos, such as:

http://en.wikipedia.org/wiki/Digital_video_fingerprinting
http://www.phash.org/

 

[nanank99]

is it already release for public yet??
please do tell if there is a release, in any kind of forms (apps or web)
as for now i using javlibrary.com too keep track and watching the jav update...

 

[cyberzen]

Basic features that are completed:

tagging videos
actress for videos
rating videos
searching

I would say it is very much alpha stage now, I've not done any testing whatsoever.

 

[CodeGeek]

I do android development, I had to upgrade my laptop from 4gb of ram to 8gb because eclipse would be so freaking slow and kept getting low memory. QtCreator only takes 180mb ram.

Do you mean sharing tags, and actress for the database? We will need to agree on how to identify a video. The simple way would be by video title, for example NHDTA-403. In the future we should have a more robust way of identifying videos, such as:

http://en.wikipedia.org/wiki/Digital_video_fingerprinting
http://www.phash.org/

I have a book here for Android development using Eclipse. But I have to admit that I never used it.

I'm not sure what you mean by identifying videos. Do you mean if someone is searching for it using the application? Or do you mean if the application is reading files on the HDD?
In my case the user can search using all attributes that a movie has. And the application will recognize files on the HDD by hashing them (CRC32, ED2K, MD5, SHA1/SHA160, TTH and Torrent hash). After that it will compare them with the hashes stored in the database. And everyone can submit hashes as CREQs for movies together with the meta data of the file (using the MediaInfo library).
I looked at the links which you posted. Seems you approach works totally different as you want to identify only the movie, but not the file. Means the user in your case will know which movie it is afterwards, but not if there is also a version with e.g. better quality.

is it already release for public yet??
please do tell if there is a release, in any kind of forms (apps or web)
as for now i using javlibrary.com too keep track and watching the jav update...

I don't know the status of the application of cyberzen. But in my case there isn't anything to show yet (unfortunately).

 

[cyberzen]

Hashing a video seems quite brittle to me, a video title could be encoded in different codecs, then the hash will be different. What I would like would be a single UUID for a movie title, meaning it should be normalized as much as possible. If there is a HD version of the movie, then it can be included in the metadata, I don't think it should be part of the UUID.

I like to name my videos NHDTA-403, some other user might name the video file differently, also he might have a different video codec. Now suppose if 50 users all name their files differently or they all have different encoded video, then there will be 50 different records all referring to the same title NHDTA-403, thats a lot of redundant data.

 

[cyberzen]

Single user mode will not be that useful. It will take me a VERY long time to tag all my videos with the relevant tags, and it will be error prone as I am not familiar with all actresses. If however we can get each user to tag a small portion of their collection and share the same tags to other users, then it will vastly save everyone's time.

 

[CodeGeek]

I guess we're talking at cross-purposes.

• In my case the ID of the movie is a artifical ID starting at 0 increased by 1 for every new entry.
• "NDHAT-403" will be the value of the "DVD code" attribute.
• The movie will have file-sets (because some movies are split in several files) and each file-set contains records, one record for each file. That record contains the different types of hashes of that file as well as the meta data (video codec, audio codec, resolution, bitrate, etc.).
• When the application reads a directory it doesn't care for the name of the files, but the hashes of these files. The files / movies will be identified by that. But I will offer an optional automatism that all the movies will be renamed by a pattern defined by the user. But again: That doesn't have anything to do with the identification of the file itself.

About the fingerprint by doing some kind of meta fingerprint: I guess the Google picture search works the same way. And most of the time it seems to be pretty accurate. But not all the time. And in this case we're talking only about pictures. I'm not sure if it will work well with movies as they have different lengths and sometimes the content is also different (some ripping groups have their own intro or outro, others have watermarks and/or subtitles). The codec or the resolution is surely the smallest problem when creating such fingerprints. But I also have to admit that I haven't tested it. Maybe I will give it a try and add that fingerprint directly to the movie. In this case it is only a guessing, but maybe it is interesting for getting hashes of new files more easily. I mean: It will be more convenient for the users to commit new hashes to the main database.

I'm not sure what you mean by "single user mode". I'm not planning to start by zero. I already have the data of more than 80,000 movies. I think that is a good point to start. But not all actresses are marked in all movies. Sougouwiki, e.g., has some more information about, but unfortunately not so many movies. But currently I don't have any hashes of the movies. It will surely take some time and a lot of people to get them. Also there are many movies out there which are not included in these 80,000 movies. So people also have to commit these movies to the database. But as you already mentioned: If everyone is committing a few movies (e.g. of his favourite actress or genre) it shouldn't be a big deal. The amount of data will be the bigger problem. I also think that it is necessary to review the CREQs. I don't want that the work of true members of the community gets wasted by some idiots which only commit nonsense.

I only can reference to AniDB.net. They have a system like that and it works perfectly for years now.
Of course that doesn't mean that there isn't room for improvement.

 

[cyberzen]

I am not familiar with CREQ, what is it?

Single user mode is when you are operating on your local database, and not connected to the server.

Not sure the audio codec and video codec should be stored in the database? What use case does that provide? What hash are you using? Everyone will have different hashes, then the server will not be able to sync the correct tags and actress to that particular video.

To me the most important data is tag & actress, nothing else matters to me.

About videos split in several parts, I think this should be a special edge case, where the user will need to manually denote it as under the same DVD code.

 

[cyberzen]

BTW, you can even signup for a free 1 yr micro instance at Amazon Web Services, they allow adult content. Obviously the micro instance is not that powerful and you will need to make sure you don't go above the allowed bandwidth (they will charge for overage). You can setup a small MySQL instance and use it for app testing / development.

http://aws.amazon.com/free/terms/

 

[CodeGeek]

cyberzen, I'm sorry that I haven't replied to your posts until now. I thought it would be better if I would explain it by some screenshots. But until now I didn't have the time to take some.

CREQ = change request

This extra information is for people who emphasize on certain codec, bit rate and/or resolution. Some codecs are also very rare so many video player can't play them.

You wrote that you're only interested in the tags and the actress (name). How would a record of yours look like?
Mine would look like this (XML structure as an example):

<movie id="12345>
  <cover id="12305" />
  <titleJapanese>爆乳痴漢祭り 巨大な乳房を揺らしながらイキまくる女達</titleJapanese>
  <titleRomanji>Bakunyū chikan matsuri kyodaina chibusa o yurashinagara iki makuru on'natachi</titleRomanji>
  <titleEnglish>Girls spree alive shaking the breast huge breasts molester festival</titleEnglish>
  <type>DVD</type>
  <releaseDate>2011-09-08</releaseDate>
  <playLength>240</playLength>
  <performers />
  <directors>
    <director id="123" />
  </directors>
  <series id="367" />
  <studio id="12" />
  <label id="782" />
  <categories>
    <category id="4" />
    <category id="12" />
    <category id="16" />
    <category id="44" />
    <category id="45" />
  </categories>
  <code>NHDTA-152</code>
  <descriptionJapanese>服の上からでもわかる爆乳を揉みしだくのは男の夢。しかも夏は薄着のシーズン！無防備に爆乳を見せつける女子校生、女子大生、OLたちのたわわに実った乳に痴漢師の手が伸びる。乳揉み、マ○コいじりで嫌がりながらも興奮する爆乳娘の乳にチ○ポを押しつけ擦りつける！！乳の大きさ、被害者数、痴漢テクetc.全てが盛りだくさんの爆乳痴漢祭り！！！ </descriptionJapanese>
  <descriptionEnglish />
  <resources>
    <resource id="783" />
  </resources>
  <filesets>
    <fileset id="34589">
	  <file id="789765">
	    <video />
		<audio />
		<hashes />
	  </file>
	</fileset>
  </filesets>
</movie>

All entries with an ID attribute are references to other records. And I didn't fill the <file> part as I'm to lazy for this now. But I hope it shows how my database will work.
By the way: That's the movie: http://www.dmm.co.jp/mono/dvd/-/detail/=/cid=1nhdta152/

 

[cyberzen]

Currently this is what my local database schema looks like, http://sqlfiddle.com/#!7/f4423/3

I have started work on the server, my goal is to be able to sync tags, acts for videos by the end of the month. For the server I am using MongoDB as the database. Each video will hold the following schema:

{
"_id" : ObjectId("52f440cb1c99dbb06bd7ddcb"),
"title" : "RBD-555",
"engDesc" : "some description about the video",
"japDesc": "服の上からでもわかる爆乳を揉み",
"rating" : 7,
"tags" : [ "r***", "incest", "cuckold" ],
"actress" : [ "Yui Hatano" ],
"hashes": [ "sdfg897sdfg98s7d0fg987sd", "fgsd87f6g89s7dfgy87sddg", "sdfg9s8df7g09s8df7g098" ],
"recommendations" : [ "RBD-455", "JUX-155" ],
"comments": [ "I love this movie", "I hate this movie" ]
}

 

[cyberzen]

Currently this is what my local database schema looks like, http://sqlfiddle.com/#!7/f4423/3

I have started work on the server, my goal is to be able to sync tags, acts for videos by the end of the month. For the server I am using MongoDB as the database. Each video will hold the following schema:

{ 
  "_id" : ObjectId("52f440cb1c99dbb06bd7ddcb"), 
  "title" : "RBD-555", 
  "engDesc" : "some description about the video", 
  "japDesc": "服の上からでもわかる爆乳を揉み",
  "rating" : 7, 
  "tags" : [  "r***",  "incest",  "cuckold" ], 
  "actress" : [  "Yui Hatano" ],
  "hashes": [ "sdfg897sdfg98s7d0fg987sd", "fgsd87f6g89s7dfgy87sddg", "sdfg9s8df7g09s8df7g098" ],
  "recommendations" : [ "RBD-455", "JUX-155" ],
  "comments": [ "I love this movie", "I hate this movie" ]
}

 

[CodeGeek]

It seems that it would be possible to use the same master data.

• Okay, the IDs have to mapped.
• And "title" in your record is "code" in mine. But I don't see any problems there.
• I don't have any rating yet and also not comments. But I will add recensions later on (like on Amazon.com). Your tags seems to be the categories of DMM (My record is strongly based on the DMM record.). But I will also add user defined tags.
• Your actress field is also no problem. I guess the "hashes" in your record are that media fingerprint, right? Okay, I don't have that.
• Recommendations are a good idea. I thought about a system which automatically creates the recommendations by certain criteria (same categories, same performers, same studio, same label, etc., but maybe also on the movies the user already has).

 

[cyberzen]

I am currently now trying to think of a solution for tags normalization, there are 2 options:

1) think up of several tags that will encompass 99% of videos and have users use those, in this option users will not be able to add their own user defined tags.
2) follow the way stackoverflow handles tags. There will be a primary tag used, and a list of tag synonyms that map back to the official tag. This option users will be able to add their own user defined tags.

Obviously option 1 would be easier to implement at the moment, but I think in the future option 2 would be most desirable.

Ya for recommendations, we can use recommend automatically based of your suggested heuristic, and allow other users to recommend (like myanimelist)

 

[cyberzen]

I've uploaded to github a very alpha build, this is more for developers atm. It's not user friendly at all.

https://github.com/zenwong/tagu

tagu means tag in japanese, well according to google translate

Lemme know if you got any problems compiling on linux. I'm not too sure about the windows / mac OSX side.

 

[cyberzen]

Added a thumbnail grid view

 

[iori11]

I've uploaded to github a very alpha build, this is more for developers atm. It's not user friendly at all.

https://github.com/zenwong/tagu

tagu means tag in japanese, well according to google translate

Lemme know if you got any problems compiling on linux. I'm not too sure about the windows / mac OSX side.

i will give it a try great work:byebye: