Orange's new web-site

Janez · **Janez** on Wed Jul 22, 2009 14:11

You may have noticed - unless you are here for the first time - that the web site underwent a major reconstruction.

Please excuse us an occasional 404 (or worse) here and there...

**eliasz** on Mon Jul 27, 2009 18:49

I appreciate your hard work with the web site, yet if almost all the docs are 404'd, please at least bring back the PDF leaflets. I'm struggling to make any sensible classification scheme and -- believe me -- any working example would help a lot

Janez · **Janez** on Thu Jul 30, 2009 9:37

This is what you get if you migrate from the sloppy case-insensitive Windows to Linux (and you don't want to cheat). Apart from a few files which were already missing before, all links should work now.

jfjanssen · **jfjanssen** on Thu Jul 30, 2009 14:08

Great!

However: on the front page, if you click "Download Orange 2.0b" (Are you really there already? Great!) and then "Nightly Packed Sources" you get a 403 access denied link...

Janez · **Janez** on Thu Jul 30, 2009 14:19

This is, however, one of the few files which were already missing before. ;)
This is in domain of the guys who maintain the Linux stuff - I just asked them again.

jfjanssen · **jfjanssen** on Thu Jul 30, 2009 14:29

I see

I think my recommendation is to use one of the many free software packages that will try to crawl your website and report broken links. Doing this automatically of course saves a lot of time and allows one to catch -all- issues: I see that some links are broken due to permissions issues in addition to missing (= sometimes improperly named) files, as well.

Janez · **Janez** on Thu Jul 30, 2009 21:44

This is what I used - it would be impossible to check the whole site manually. There are some 15 broken links (excluding the external links and maybe some links in the forum posts - I'll check these later) and they have been assigned to those responsible. I hope they'd have some time soon...

**Guest** on Thu Jul 30, 2009 22:20

http://www.ailab.si/orange/features.html is not working.

Janez · **Janez** on Thu Jul 30, 2009 22:35

Thanks, I know. It's waiting for Blaz. :)

**aonlazio** on Fri Jul 31, 2009 16:50

Could the installer change to support python 2.6? When I install, it automatically goes into python 2.5 but I want to use orange with 2.6.
Thanks

Janez · **Janez** on Fri Jul 31, 2009 17:18

Compiling Orange for Python 2.6 and building the installation files is more or less a matter of adding a string "26" to the make file, except...

Orange requires numpy, PyQt and PyQwt. After a long wait for numpy, we now miss windows binaries for PyQwt for 2.6. If anybody builds them and sends them to us (this might be easy or not, I don't know), then we have, I believe, everything we need.

I hate having to always lag behind, too.

**aonlazio** on Sun Sep 06, 2009 3:13

PyQWT for python 2.6 is out, it's time for orange to change to python 2.6 support.

Janez · **Janez** on Wed Sep 23, 2009 22:09

Thanks. We're almost there. Daily snapshot gets build, we only need to fix a few problems due to changes in PyQwt (mostly just method names).

Edit: I gave it a quick try and it already seems to work without any problems. If you are eager to try it, you can ... if you can guess the link. :) (Some simple notproning won't hurt.)

Janez · **Janez** on Fri Sep 25, 2009 21:38

After fixing the last problems with the new PyQwt, Orange for Python 2.6 is now available for testing. Please report any problems you might still encounter.

**Guest** on Tue Dec 22, 2009 19:52

I have sequences like

ADSFSWWERERGFSDGDFSG,
ETYRGDGFUHGDBVXCNXVNN and so on.

So each character in the sequence is differentiated as important one and unimportant one. For example i label important one has "1" and unimportant as "0" as shown below.

A D S F S W W E R E R G F S D G D F S G
0 0 0 0 1 0 0 1 0 0 0 1 1 0 1 0 1 1 1 1

Now i have the sequences with this information.
I should train these sequences using Naive Bayes classifier and when i give a test sequence, the classifier should give result of the information of each character if it is important or not.

I want to try with different window sizes like 10 or 15 where the target character should be in the middle. I mean i should discover the pattern of the characters either side of the target character so that that can be used in finding the importance of the target character in the test sequence.

So is there any possible of giving the sequences without making the subsets with a window size manually?
how should be the format of the data instances to be used as input data?
Am I going correctly to achieve the objective?if not ,please suggest me.

Thanks in advance.

NET

	Orange Forum Forum Index -> Latest News	All times are GMT + 1 Hour
Page 1 of 1