NatI icon indicating copy to clipboard operation
NatI copied to clipboard

grooveshark is also a dependency

Open shiruken1 opened this issue 10 years ago • 10 comments

And it was quite a pain to install. Also, I'm getting this:

Resource u'corpora/stopwords' not found. Please use the NLTK Downloader to obtain the resource: >>> nltk.download()

shiruken1 avatar Feb 04 '15 01:02 shiruken1

I have updated the dependencies. Thanks.

You need to download the nltk corpus, since NatI makes use of some of its data, such as the stopword collection from the nltk corpus. The following link has some information of how can you download the nltk corpus. http://www.nltk.org/data.html

rcorcs avatar Feb 04 '15 03:02 rcorcs

Awesome, thanks!

Now onto the next problem: it detects the speech perfectly, but it doesn't execute any commands.

Also, can it take dictations?

Thanks for your work on this! I'm learning Python and will definitely be pitching in, as voice applications are sorely needed in Linux.

shiruken1 avatar Feb 04 '15 04:02 shiruken1

It doesn't support dictations.

Try some simple commands first, such as "What time is it?", and then see what happens. The source code for the commands execution is in the commands.py file and the methods that perform the actual commands starts with a do[Action], such as doWhatTime(). Check if you can make it work, or identify why it's not working.

Thanks for your interest.

rcorcs avatar Feb 04 '15 04:02 rcorcs

Out of curiosity, what could I do to make it support dictations? You've made it sample pretty darn well! :) It even picks up bad words and censors them LOL.

It seems to me that if I adjusted the length of the samples, it should work for dictations: http://i.imgur.com/gbq1aDz.png

Let me know what you think.

On Tue, Feb 3, 2015 at 11:44 PM, Rodrigo Caetano Rocha < [email protected]> wrote:

It doesn't support dictations.

Try some simple commands first, such as "What time is it?", and then see what happens. The source code for the commands execution is in the commands.py file and the methods that perform the actual commands starts with a do[Action], such as doWhatTime(). Check if you can make it work, or identify why it's not working.

Thanks for your interest.

— Reply to this email directly or view it on GitHub https://github.com/rcorcs/NatI/issues/7#issuecomment-72789951.

shiruken1 avatar Feb 04 '15 16:02 shiruken1

I think it is a good idea. There is a sleep mode, using the commands "go to sleep" and "wake up". Perhaps it needs only a dictation mode. When you finish the dictation mode it saves a file with the text. At the moment I'm very busy finishing my master's degree in the area of parallel computing. If you can do that, perhaps I can add it to my version as well.

rcorcs avatar Feb 04 '15 18:02 rcorcs

@shiruken1 Hey there! I'm stuck with the same issue you had (it detects the speech perfectly, but it doesn't execute any commands). Did you achieve any progress for that matter? I didn't want to ask the same question to Rodrigo as he seems to be very busy. Instead I'm asking you...Could you give me any advice on solving it? Thanks!

Coffeer avatar Jul 16 '15 22:07 Coffeer

Hey @Coffeer, the pygrooveshark library recently stopped working, as the grooveshark web site has been shut down for copyright reasons. Until there is no other solution for playing music by streaming implemented, I suggest removing this feature alongside the pygrooveshark dependencies. When possible, I will return my activities with NatI, looking for new solutions for music and also new features. Thanks!

rcorcs avatar Jul 16 '15 22:07 rcorcs

Hi @rcorcs, I'm the same person who sent you an e-mail in Portuguese a couple of days ago. Yep, I already commented all lines referring to pygrooveshark module as you instructed me. That worked! I'll be watching for your news. I'm looking forward to seeing you back on track soon. Thanks a lot for you attention!

Coffeer avatar Jul 16 '15 23:07 Coffeer

That's so funny! Just yesterday I decided to take some time to look at the code, and had downloaded this again to give it another shot. Which lines did you comment out, @Coffeer? Also, to answer your question: I didn't get too far into it, since this year has been a whirlwind for me. Thankfully things are settling down, so I'll definitely keep you guys posted if I come up with anything.

shiruken1 avatar Jul 17 '15 00:07 shiruken1

@shiruken1 I'm glad you're back to this! I have no experience at all with Natural Language Processing. I'm very excited I found this excellent work @rcorcs did! That will definitely be a valuable study resource for me along with some books I got and on-line documentation. At least I have some free time on my hands. Hope you guys will find free time as well. Thanks answering!

Those are the lines I commented out: commented_lines

Coffeer avatar Jul 17 '15 00:07 Coffeer