Jump to content

Wikipedia:Reference desk/Archives/Computing/2014 February 23

From Wikipedia, the free encyclopedia
Computing desk
< February 22 << Jan | February | Mar >> February 24 >
Welcome to the Wikipedia Computing Reference Desk Archives
The page you are currently viewing is an archive page. While you can leave answers for any questions shown below, please ask new questions on one of the current reference desk pages.


February 23

[edit]

Need a site for publishing details on a DSP algorithm

[edit]

I've recently worked out the details for implementing a specific real-time sample rate conversion algorithm. There is a lot of information online about the basic technique, but next to nothing focused on how to actually get from the theory to functional code. Working it out took a lot of jumping between several different references to tie together all the details. I'd like to write up a nice article describing how the method works and how to actually implement it so that other people can benefit from it, but I don't know of a good place to publish it online. It's the sort of thing I would have published on my blog, but it's been offline for a few years and I don't want to revive it now. Katie R (talk) 02:46, 23 February 2014 (UTC)[reply]

Sounds great. If this is a general sort of algorithm, you might consider the wikibook on DSP, it might go well in multirate systems section. If this is more specialist and/or novel, people sometimes publish on Github, but you'd still want some venue to let people know about it. If this is for an audio application, KVR forums might be a place to discuss the algorithm and musicdsp.org has a lot of great DSP-related source code. --Mark viking (talk) 03:30, 23 February 2014 (UTC)[reply]
If I publish any code it will be psuedocode, because this was done at work and I can't release code developed for the project, so Github is out of the question. The algorithm is pretty common, and is implemented in a few open-source libraries, but I needed to develop it from scratch to avoid licensing issues. Does wikibooks have the same sort of rules as wikipedia for OR? The details are all covered in various sources, but pulling it all together involves some reasoning that isn't covered in any sources I found. Katie R (talk) 15:49, 23 February 2014 (UTC)[reply]
The Wikibooks OR proposal seems compatible with pseudocode examples based on algorithms from published sources. --Mark viking (talk) 23:26, 23 February 2014 (UTC)[reply]

Writing Unicode text to files

[edit]

I'm working on a project where I perform sentiment analysis on tweets, and I'm trying to stream Hindi tweets and write them in a file in a human-readable form. I've tried the following piece of code for writing the Hindi data to a .txt file.

    with open("C://Users/User/Desktop/hindistream.txt","ab") as f:
        f.write(tweet.decode("UTF-8").encode("UTF-8"))
        f.close()

Funnily, the code works when I give it a variable tweet="भारत का इतिहास काफी समृद्ध एवं विस्तृत है।" or something along those lines, but doesn't work when I plug it into the twitter-streaming code as below:

    class  listener(StreamListener):
        def on_data(self, data):
            try:
                tweet = data.split(',"text":"')[1].split('","source":')[0]
                tweet = tweet.decode("utf-8")
                print tweet
                with open("C://Users/User/Desktop/hindistream.txt", "a") as f:
                    f.write(tweet.encode("UTF-8"))
                    f.close()
                return True
            except BaseException, e:
                print 'failed ondata,',str(e)
                time.sleep(5)
        def on_error(self,status):
            print status

The data does get written, but it's encoded (the text becomes something like "\u0939\u0948\u0902", instead of Hindi characters). Can anybody tell me where I'm going wrong? La Alquimista 06:04, 23 February 2014 (UTC)[reply]

JSON strings can contain \uXXXX escapes, and you don't have any code to unescape them. I suspect that's your problem. Use the standard json library instead of str.split to parse the data. -- BenRG (talk) 07:31, 23 February 2014 (UTC)[reply]

I'm not sure how to. >_< I'm a rookie still, and much of the code has been stitched from multiple pre-existing snippets... You mean I should not use the split() function? What equivalent function does the json library have? La Alquimista 08:52, 23 February 2014 (UTC)[reply]

I think that your input data is JSON, and Python's json library will parse it correctly and give you a Python data structure, avoiding the need for ad-hoc parsing. Try replacing the first two lines of the try block with something like
  data = json.loads(data.decode('utf-8'))
  tweet = data['text']
This may not work if the "text" key is nested inside something else, but if the tweets look something like this then it should work. You also need to import json at the top of the file. By the way, you can delete the line f.close(), because exiting the with block closes the file. -- BenRG (talk) 12:03, 23 February 2014 (UTC)[reply]

Thank you! It totally worked! A barnstar to you sir! However, I have a follow up question. Earlier, I was using the split() function to get the main tweet that appeared inside all that code (the tweets always appeared between "text:" and "source:", so I had made that the criteria for my split). But this code appears to work fine just with 'text'. How does that work. How does python know where to stop extracting the tweet from? La Alquimista 15:44, 23 February 2014 (UTC)[reply]

The tweet data looks like ...,"text":"This is a tweet","source":.... The syntax is similar to a Python dict: "text" is the key, and "This is a tweet" is the corresponding value. The ending quotation mark is what ends it. "source" is the name of another key that just happens to be listed next in the tweets you've looked at so far, but probably won't be in general. -- BenRG (talk) 06:57, 25 February 2014 (UTC)[reply]

minecraft server?

[edit]

some friends of mine have decided to set up a small minecraft server, so we can all play together, seems a good idea, but for some reason even the computer experts are flailing around trying to find a way of linking us all to the game. Originally we tried hamatchi, but it kept dropping out or not working properly, so we switched to something called tunngle, seems to be working, but it's throwing adverts at us all the time and generally being annoying. Anyone know of a simpler, less intrusive way we can get something set up? I don't know much about computers, but I get the impression these programs are effectively doing something we could set up ourselves without their help, perhaps with better server hardware/software? or perhaps not, maybe we just need a better program to run the connection for us? any thoughts?

many thanks,

94.8.179.211 (talk) 19:14, 23 February 2014 (UTC)[reply]

There are many free software utilities to create a virtual private network. It can be done! But the reality is, especially if you are using Windows, setting up the VPN server will take a lot of effort, even for a pretty skilled computer technician. Are you running Windows or something else? Nimur (talk) 04:44, 24 February 2014 (UTC)[reply]

at the moment I am running both Windows and Mint, and have the option to install another OS if needed, though it'd need to be something others can access through Windows devices. We have the extra hardware to set up a physical server within my house if that would be easier? (assuming it can't, or at least shouldn't, be hosted off the same commercial server account my websites are on) 94.8.179.211 (talk) 11:31, 24 February 2014 (UTC)[reply]