VrrtepCLI

Started by Tirea Aean, May 22, 2011, 03:40:58 PM

Previous topic - Next topic

0 Members and 1 Guest are viewing this topic.

Swoka Ikran

Quote from: Tirea Aean on October 22, 2011, 05:11:44 PM
But beware(!) words which start with what looks like a prefix but is not. (e.g., tìng, len, nìn, nìltsan etc.)

and words which end with what looks like a suffix but is not. (e.g., Neytiri, riti, pxuntil, muiä, omatikaya, etc.)

and words which contain what looks like an infix but is not (e.g., tswayon, terkup, hamalo, hayalo, toltem, etc.)
Yeah. Those words are what concerned me. I have to think about this some more (and start prototyping), but what I've thought of so far:

Some of these words aren't even complete if the affix is removed (toltem being a good example - ttem isn't valid AFAIK), so that's one method. Things like "tswon" are an issue for this though.

The most inefficient, but often one of the most successful, method that could be used is permutation trials. This method can be very slow, but on the plus side, it can be up to 95% accurate when implemented in combination with other detection methods.


White lists are also an option. It'd need to be implemented alongside the permutation engine.

As an absolute last resort, there's always the "Enter both the affixed word and root word and I'll tell you what the affixes on it mean" option. This one defeats the feature's original goal by rendering it useless for beginners ( :( ), but it would add functionality nonetheless.
2010 was the year of the Na'vi.Vivar 'ivong Na'vi!


 
Avatray | NWOTD Sigbars | Sacred's Sigbar Tool | My collection of Avatar merchandise

Blue Elf

Well, I'll try to add my thoughts (although probably useless):
IMHO all depends on the fact how smart program would be. If you know word type to "dismantle", maybe situation is easier:
You have "tereiok" as input and know it is a verb. Then you can take list of infixes and try to find them inside of word. If you fing any, remove it from the word and continue until all posiible infixes were tried. Now you probably have root verb which should be found in the distionary files. If not, then:
1/ verb on input was already root verb, for example zerok, zamunge (they both contains "infix", which is not real infix)
2/ verb on input was one of verb from 1/ which contains another infix. So probably you should first find all infixes and then try all combinations of removing all infixes and search in dictionary

The same I probably applicable to nouns, adjectives etc.

But if you don't know word type (beginners probably will not take care about word type), things gets very complicated and combinatorial analysis should be used.

BTW ma Tirea - congratulation for 5000 posts :)
Oe lu skxawng skxakep. Slä oe nerume mi.
"Oe tasyätxaw ulte koren za'u oehu" (Limonádový Joe)


Swoka Ikran

Quote from: Blue Elf on October 23, 2011, 11:39:37 AM
Well, I'll try to add my thoughts (although probably useless):
IMHO all depends on the fact how smart program would be. If you know word type to "dismantle", maybe situation is easier:
You have "tereiok" as input and know it is a verb. Then you can take list of infixes and try to find them inside of word. If you fing any, remove it from the word and continue until all posiible infixes were tried. Now you probably have root verb which should be found in the distionary files. If not, then:
1/ verb on input was already root verb, for example zerok, zamunge (they both contains "infix", which is not real infix)
2/ verb on input was one of verb from 1/ which contains another infix. So probably you should first find all infixes and then try all combinations of removing all infixes and search in dictionary

The same I probably applicable to nouns, adjectives etc.

But if you don't know word type (beginners probably will not take care about word type), things gets very complicated and combinatorial analysis should be used.
It's still permutation/combination testing, but optimized to not check words of the wrong type. It'd be useful for increasing speed :) If word type is known, use it. If not, just try everything.

Also, for your #1...that's easy: check the word in the dictionary before doing anything, and go on to testing only if it's not found. Doing it this way eliminates the testing altogether for words like zerok. Of course, testing is still needed if someone sticks an infix in it.
2010 was the year of the Na'vi.Vivar 'ivong Na'vi!


 
Avatray | NWOTD Sigbars | Sacred's Sigbar Tool | My collection of Avatar merchandise

Tirea Aean

Sounds like good progress. :D Also, thank you ma Blue Elf, I did not notice I just made my 5000th post.

Anyways, I have had similar ideas myself. Where should we start to implement this?  maybe make a logic/pseudocode type deal I can convert into python.. perhaps we can even intro a new .py file to import which contains this filtering bit. (instead of cluttering vrrtepcli.py with more stuff than it already has)

Swoka Ikran

Quote from: Tirea Aean on October 23, 2011, 12:31:27 PM
Sounds like good progress. :D Also, thank you ma Blue Elf, I did not notice I just made my 5000th post.

Anyways, I have had similar ideas myself. Where should we start to implement this?  maybe make a logic/pseudocode type deal I can convert into python.. perhaps we can even intro a new .py file to import which contains this filtering bit. (instead of cluttering vrrtepcli.py with more stuff than it already has)
I was gonna prototype it in PHP since it has more functionality than AutoIt, is easy to make a crude UI for (HTML form for input, and echo $output;), and I can use the same dictionary (in its much-more-flexible SQL database form) during testing. :)

When it does get converted to .py, I agree it should be in a separate file (e.g. affix.py) to reduce clutter in vcli.py.

Hopefully I'll get to starting on this sometime this week. I am busy with homework and another PHP project though (making a version of my sigbar tool that's suitable for public use).
2010 was the year of the Na'vi.Vivar 'ivong Na'vi!


 
Avatray | NWOTD Sigbars | Sacred's Sigbar Tool | My collection of Avatar merchandise

Key'ìl Nekxetse

#325
A quick bug report:
Installed VrrtepCLI 1.93 on Ubuntu 11.04, 32bit, Python 2.7.1. Mostly working but got this trace when running "vrrtepcli -q"
The bug seems to occur whenever the program is run with -q.

Edit: had a few minutes, so I had a look and it was a typo in quiz.py; line 22 should be "value = maximum + 1" not "value = max + 1".
If you want to do some complex searching of strings, have you though about regexp? I'm not sure exactly how you'd use them for infixes, but I think the ability to match a complex sequence of letters with optional characters and so on might be quite suitable. I don't know how efficient they are in Python, might not be fast enough in a large search.
Key'ìl Nekxetse on "The Revolutionists"
~$ life --help
The program life received signal SIGSEV. Core dumped.

Tirea Aean

#326
Ah yes, I fixed that in the repository source code a while back but I guess I never pushed the update to the download page. Thanks for bringing that up.

EDIT: I would update the quiz module right now, but I'm actually working on it to add quiz/scramble highscore saving. so far it works, but I'm thinking on how best to implement it. I think I will add a print command when you quit the quiz to print your score and your highscore...

Tirea Aean

#327
okay, It seems I do not have ssh or scp access to my webdir on Haxalot's server anymore. that's gonna be a big game-changer... That server is where all the updates have been uploaded and stored. If I change where the updates come from, users will most likely need to download a new version...

I'm gonna try to work this out, I was just now in the middle of updating the dictionary files

EDIT: Got it figured out, update should be out.

Now, in the next version I'm gonna build a way to update the program from within itself instead of going to google code to get the newest version.

Swoka Ikran

Quote from: Tirea Aean on November 01, 2011, 09:47:03 AM
Now, in the next version I'm gonna build a way to update the program from within itself instead of going to google code to get the newest version.
Easiest is to just have it download the app files alongside the dictionary, and maybe add an option to update program, dictionary, or all. I was actually planning this for Windows 1.94, but scrapped it when I decided to do the menu enhancement.

Quote from: Tirea Aean on November 01, 2011, 09:47:03 AM
okay, It seems I do not have ssh or scp access to my webdir on Haxalot's server anymore. that's gonna be a big game-changer... That server is where all the updates have been uploaded and stored. If I change where the updates come from, users will most likely need to download a new version...
They will need a new updater initially.

As for SCP/SSH, why not just use plain FTP? Not sure about Linux, but it works great from a batch script on Windows. I pushed application updates this way for almost 3 years, until early June.
2010 was the year of the Na'vi.Vivar 'ivong Na'vi!


 
Avatray | NWOTD Sigbars | Sacred's Sigbar Tool | My collection of Avatar merchandise

Tirea Aean

well, that's what I was gonna do: make it download and copy the source code to ~/.vrrtepcli/ including the dictionaries and all .py and .sh scripts. and include switch arg for vrrtepcli -u --all and default is just update dicts vrrtepcli -u

disregard the scp/ssh thing, Hax changed the port due to loads of unnecessary port 22 traffic, is all. The update should have been out as of yesterday.

Eywa'eveng-tìranyu

If I use the IPA option, it doesn't show it.
@ubuntu:~/vrr$ ./vrrtepcli.sh -ipa txep
n. fire


I'm using Ubuntu 11.0, vrrtepCLI v1.93 and Python v. 2.7.2+

Swoka Ikran

Quote from: Eywa'eveng-tìranyu on November 04, 2011, 10:18:41 AM
If I use the IPA option, it doesn't show it.
@ubuntu:~/vrr$ ./vrrtepcli.sh -ipa txep
n. fire


I'm using Ubuntu 11.0, vrrtepCLI v1.93 and Python v. 2.7.2+
It does work if you use vrrtepcli -l -ipa fire. It's probably a bug...I don't maintain Linux ver though, so for all I know, this could be intended behavior.

I don't know too much about the IPA function since the Windows version doesn't support it.
2010 was the year of the Na'vi.Vivar 'ivong Na'vi!


 
Avatray | NWOTD Sigbars | Sacred's Sigbar Tool | My collection of Avatar merchandise

Tirea Aean

#332
Quote from: Eywa'eveng-tìranyu on November 04, 2011, 10:18:41 AM
If I use the IPA option, it doesn't show it.
@ubuntu:~/vrr$ ./vrrtepcli.sh -ipa txep
n. fire


I'm using Ubuntu 11.0, vrrtepCLI v1.93 and Python v. 2.7.2+

As I said earlier in the thread, -ipa currently ONLY WORKS in the LOCAL --> NA'VI direction. That is, you need to know the local translation of the Na'vi word you want the IPA for, and call the IPA of the word from the Local word, e.g


@ubuntu:~/vrr$ ./vrrtepcli.sh -l -ipa fire


I am trying to figure out how to make it so you can go Na'vi -> Na'vi to call the IPA of a Na'vi word from Na'vi input. e.g.


@ubuntu:~/vrr$ ./vrrtepcli.sh -ipa txep


it does not exist yet.

Eywa'eveng-tìranyu

maybe just search the metaWords.txt for the word and display the IPA that's
in that line?

Tirea Aean

That's what I'm workin for, but I'm actually busy with a Java project for school, and other homework and random timekilling things in my schedule.

I'll post here when I have updated the Google Code repo. And as I said, from the next version, after initial installation, the user can just update the program from within running it.

Tirea Aean

#335
NEW Linux/Mac version! New features including:

  • get IPA of a Na'vi word from Na'vi input
  • get infix positions of a Na'vi verb from Na'vi input
  • upgrade the whole program from within itself, much quicker than constantly returning to Google Code

    Vrrtep for Linux/Mac on http://tirea.skxawng.lu/source

Tirea Aean

#336
What is up with this??? I am not sure how I feel about this, honestly. It's an old version, and all the google code downloads are up to date. I don't want more than one place to develop this. Google code svn repository has been great. Anyone have any idea what this madness is about? What do you guys think about this Softpedia thing? I'm kinda happy,flattered, and annoyed at the same time...

Quote from: my email, minutes ago
On Saturday, December 10, 2011, Softpedia Editorial Team <[email protected]> wrote:
> Congratulations,
>
> Vrrtep CLI, one of your products, has been added to Softpedia's database of
> software programs for the Windows operating system. It is featured with a
> description text, screenshots, download links and technical details on this
> page:
> http://www.softpedia.com/get/Others/Miscellaneous/Vrrtep-CLI.shtml
>
> The description text was created by our editors, using sources such as text
> from your product's homepage, information from its help system, the PAD
> file (if available) and the editor's own opinions on the program itself.
>
>
> "Vrrtep CLI" has been tested in the Softpedia labs using several
> industry-leading security solutions and found to be completely clean of
> adware/spyware components. We are impressed with the quality of your
> product and encourage you to keep these high standards in the future.
>
> To assure our visitors that Vrrtep CLI is clean, we have granted it with
> the "100% FREE" Softpedia award. To let your users know about this
> certification, you may display this award on your website, on software
> boxes or inside your product.
>
> More information about your product's certification and the award is
> available on this page:
> http://www.softpedia.com/progClean/Vrrtep-CLI-Clean-202095.html
>
> Feel free to link to us using the URLs above. If you choose to link to the
> clean award page for your product, you may use the award graphic or a text
> link: "100% FREE award granted by Softpedia".
>
> We're also providing download mirrors for your product on our dedicated
> servers. If you want to link to our download page please do so by using the
> URL below:
> http://www.softpedia.com/progDownload/Vrrtep-CLI-Download-202095.html
> We have some buttons (GIF format) if you want to provide an image-based
> link to that page instead of a text-based one. You can download a small ZIP
> archive (16KB) that contains various size formats of these buttons at the
> address below:
> http://www.softpedia.com/files/softpedia_download_buttons.zip
>
>
> Your developer page on Softpedia can be reached at the URL below. It
> contains the list of software products and a link to your website.
> http://www.softpedia.com/developer/Tirea-Aean-82905.html
>
> If you feel that having your product listed on Softpedia is not a benefit
> for you or simply need something changed or updated, please contact us via
> email at [email protected] and we will work with you to fix any
> problem you may have found with the product's listing.
>
> --
> Sincerely,
> The Softpedia Team
>
> -----------------------------------------------------------------------
> Softpedia is a library of over 400,000 free and free-to-try software
> programs for Windows, Mac OS and Linux, games and gaming tools, Windows
> device drivers, mobile devices and IT-related articles.
> -----------------------------------------------------------------------
> Softpedia - the encyclopedia of free software downloads
> http://www.softpedia.com/
>
>

Swoka Ikran

#337
One of the few things I don't like about softpedia is that they do exactly this (add stuff without asking). The programs do see more downloads from being listed there though.

As for current vCLI listing, it's mostly accurate. It's not an outdated version, they listed the Windows version of the program. 1.93 is officially current for Windows since I never released a 1.94 zip file.

Also, the version numbers for Win and Linux do not correspond. My 1.94 was a menu for GUI users, yours was score saving and program updating. There is no Windows version with Linux 1.94's feature set. The updater will come with 1.95, the score stuff is TBD.

Also, some things to note about the listing that could be improved:

  • The description could be better. A more descriptive explanation of what it does could be useful to downloaders.
  • The caption describes it as cross platform, but there's no mention of what other platforms or where to get a non-Windows version of it.
  • Author should be "Multiple" since there's more than one
  • You have to click TA's name, then click "homepage" to find the SVN currently. Would be better if this was noted on the main listing
2010 was the year of the Na'vi.Vivar 'ivong Na'vi!


 
Avatray | NWOTD Sigbars | Sacred's Sigbar Tool | My collection of Avatar merchandise

Tirea Aean

I didn't write any of that stuff and idk how to edit it. I'll look into it more when I get back to a computer.

Blue Elf

I think you can be calm, ma Tirea.
Vrrtep is listed under Windows category, and version 1.93 is last Windows version of vrrtep (as Swoka ikran said already).
What is interesting, they didn't add it also under Linux category. Your developer page at softpedia links back to google code, it is not mirro or something like this, if I see well.
The only bad thing is that they did ask you for agreement. I see softpedia first time, so i have no idea how it works - maybe it provides only free software?

BTW, while playing with vrrtep, i found this:
C:\Windows\System32>vrrtepcli -l
Vrrtep CLI v1.93 by Tirea Aean
Windows version by Swoka Ikran
Standalone version

vrrtep:> language
Query matches:
n. Traceback (most recent call last):
 File "vrrtepcli.py", line 349, in <module>
 File "vrrtepcli.py", line 317, in main
 File "encodings\cp437.pyc", line 12, in encode
UnicodeEncodeError: 'charmap' codec can't encode character u'\xcc' in position 1
: character maps to <undefined>

I expect it is because Czech locale (using Win 7 english, Czech locale installed, but default locale for nonunicode progs is english )
Oe lu skxawng skxakep. Slä oe nerume mi.
"Oe tasyätxaw ulte koren za'u oehu" (Limonádový Joe)