Welcome!
This is the community forum for my apps Pythonista and Editorial.
For individual support questions, you can also send an email. If you have a very short question or just want to say hello — I'm @olemoritz on Twitter.
Installing NLTK
-
Yes I do hav the book
-
I didn't run the setup. I just tried the following
import nltk nltk.download()And that allowed me to download the corpora.
-
How did u simply import nltk
When i do that it gives me an error that nltk doesnt exist
Can u walk me through the steps of dowloading this through pipista all the way theough dowloading the corpora?
Thank you -
I used the version of shellista that includes the pdown command (not pipista). Then after unzipping and untaring, I moved everything to an nltk directory in Documents (within the Pythonista app). I was then able to import nltk and run samples from the book. I found it easier to move things around on the iPad using Diskaid from my Mac but you should be able to create the directories and move things from within Pythonista.
-
So u moved the entire nltk-2.0.4 folder to another folder named nltk and just opened up the console and run your commands?
-
@karthikmaiya, are you using the latest (1.5) version of pythonista? Do you have site-packages > nltk-2.0.4 > nltk or do you have site-packages > nltk or something else? I had to move things around a bit (using shellista). Just having the simple nltk right at the top level was the easiest way to get it working for me. I'm away from my iPad until tonight, so I can't look at all the details.
-
Yes i am on 1.5. My site-packages is separate from the folder nltk-2.0.4. So my nltk-2.0.4 folder is on the top level.
-
I have a folder nltk at the top level (same level as pipista.py). Within the nltk folder are all the files and folders that were inside the nltk-2.0.4 folder.
-
So how did u then import nltk. It still does not work for me. Did you have to run any setup scripts such as setuptools? And if so can you give me any specifics. I really want to start going through the book.
-
I haven't installed nltk, but for other toolboxes I've found that setup.py does not work.
Part of this is because the folder structure is different in pythonista than other environments. You don't have write access to ~ for example! you want to keep everything under ~/Documents.Putting the folder under Site-packages is your best bet. Not sure, but you may need to restart pythonista to get it to reload that folder.
You can also os.chdir to your nltk folder (the one containing the actual nltk.py), and import from there.
You can also append to sys.path with the pointer to the folder containing nltk.py -
Ok so i finally got it to allow me to import and open the downloader. Do I need to do any config like set the download path or anything??
-
I have not changed any other settings.
-
It will not let be dowload brown. How did u guys do that and how did u guys run brown.words()
-
For example how can i set the NLTKDATA variable to the location of nltk_data
-
I think I used pythonista to open the data.py file and set it by hand. I think I added it as a path in the list on line 66. Don't know if that is the proper way to do it, but I figured I might have to get a bit hacky to make it work on an iPad.
-
Thats what I don't know though. Where am I setting it to? How do I find the path of nltk_data
-
My nltk_data directory is at the same level as Documents and NLTK works without setting the NLTKDATA variable.
-
Using the console I used
import os print os.path.expanduser('~/')To get the path to the top level of my pythonista install. You can then add on whatever directories you have.
-
Playing with it some more I've found that it also wants Tkinter. Depending on which packages you load you might get a warning:
Documents/site-packages/nltk/draw/__init__.py:14: UserWarning: nltk.draw package not loaded (please install Tkinter library). warnings.warn("nltk.draw package not loaded "So text#.dispersion_plot() won't work, but oddly FreqDist(text#).plot(#) does. Evidently plot uses Matplotlib, but dispersion_plot() uses Tkinter.
-
Spoke too soon. I changed line 456 in nltk.text.py to
from nltk.draw.dispersion import dispersion_plotAnd now the dispersion_plot works. Not sure what in the draw package requires Tkinter, but that wasn't it.
