jMule

[Guest]

I'm writing a program to search on the emule network, store search results and compare them to what the user already has. I'd like to use jmule for searching.

Now, some doubts:
- how stable searching is?
- do you think it's easier to start the full jmule or is it "easy" to separate the searching portion from the rest?

By the way, congratulations for the great job!

[emarant]

Use jmule, yes. because:
1. File sharing means to share files when ever possible. So if a user connects to internet to search something in a filesharing network he/she should also give credit back.

2. Have a plugin to jmule that maintains a database of all files the user have allready (but not in share because it is not good to share every file the user have in edonkey2000/emule network) is good idea also to include *protection* for new download from ed2k links. It could also maintain lists of files user had deleted and don't want to download again, or things he/she don't want to download now.

ps. Such kind of tools should always be opensource to prove that there is no backdoor for any spying propose, as plugin written in java that shouldn't be a problem. Any result of your work at least a link to would be usefull for jmule users, if you don't want add plugin directly to jmule project on sourceforge.net.

pps. Global search is limited to not search on every known server, also results for *local* search are limited in number and search for some kind of metatags (like mp3 info tag) isn't supported now. But every thing of this could be fix (if someone has time to take care of this features).

[Guest]

The code is available at http://sourceforge.net/projects/cdbrowser Do you have any "good place" to put this link?

I already started to "extract" the search engine, and it's almost done. Maybe this is the way to go for now...

On 1, I don't think it's much usefull to share stuff while the search is running. Probably the guy will also have a client running. It doesn't make much sense to search for stuff you don't plan to download...

On 2, there's a plugin "system" for jmule? This is a UI-based application (it's really usefull for browsing files), I'm not sure how it would fix into jmule. If you have time, take a look at it and let me know what you think. Maybe it would make sense once jmule has a UI environment, then it would fit as a plugin.

On the features side, it's really planned as a way to collect ed2k links, compare to what you already have and decide what to download. Say, you can have a folder with stuff you already downloaded and another one with the search results. Then you can, for instance, create a folder on the search results and move there the search results you're not interested in. When you update the ed2k search, it won't place the same result again, as it's there already (just on another folder). Obviously I'm still working it out, and suggestions are welcome.

I'm currently using local search only on one server, then connects to another and search in it, and loops for all servers. By the way, to get more results I added
public final static int OP_SEARCHREQUESTMORE = 0x21;

[RastaMooGle]

Why you want to take our code to make a new prog ? Why you don't join jmule team ?
Now we have an unstable gui in improvement (in cvs dev), but we don't make yet a realease. I hope it'll arrive soon.

Therefore someHelp could be really interesting.

But I have a question : what is the main goal that to have a separate search engine from a client.
Coz after with a running client we could improve searchEngine that to get from source the shared file. I know some client don't give them but we could get them by another way...

And don't forget jmule is a multi P2PProtocol gateway..
[/url]

[Guest]

I'll donwload the CVS and have a look. Maybe this software would be a great interface for searching (and even for managin files as a whole, with some changes).

The original goal was to write a program to catalog CDs, but very flexible in terms of editing the position on the tree and the details of the files. It also seemed a great place to store ed2k links (I was tired of searching, getting a lot of results and don't know what to do with them). Note that it SAVES search results, so you can go back on them later and do things like check if you already downloaded that file and deleted it, or move search results to another folder.

There's not much logic in it, but the UI is really nice (well, at least for me...). And it can handle A LOT of files (I have catalogs her with over 100K files).