For Windows, see the quick summary in the Windows screenshot at the end. The macOS steps and a more detailed explanation are given immediately below.
At the Terminal, after installing SWI-Prolog, type
swipl. You should see the Prolog welcome banner and its prompt
?-. Type py_version. and hit Enter. Don't
forget the period (.) at the end of the command (a Prolog quirk). This
one embeds Python 3.9.6 (YMMV, you will probably have a different
version showing).
$ swipl Welcome to SWI-Prolog (threaded, 64 bits, version 9.2.5) SWI-Prolog comes with ABSOLUTELY NO WARRANTY. This is free software. ?- py_version. % Interactive session; added `.` to Python `sys.path` % Janus 1.3.0 embeds Python 3.9.6 (default, Feb 3 2024, 15:58:28) [Clang 15.0.0 (clang-1500.3.9.4)] true.
Next, we can enter the linked Python interpreter directly from
SWI-Prolog. Type py_shell. and Enter after the prompt.
You should see the Python welcome banner and a prompt
>>>. Notice the Python version is 3.9.6, the
important thing is that it's the same one that was reported above.
?- py_shell. Warning: Janus: py_shell/0: Importing janus into the Python shell requires Python 3.10 or later. Warning: Run "from janus import *" in the Python shell to import janus. Python 3.9.6 (default, Feb 3 2024, 15:58:28) [Clang 15.0.0 (clang-1500.3.9.4)] on darwin (InteractiveConsole)
If you type sys.executable and hit Enter (notice no
period is required in Python), you can see the underlying program is
swipl.
>>> sys.executable '/Applications/SWI-Prolog.app/Contents/MacOS/swipl' >>>
Assuming you haven't already installed nltk into this particular
Python, the command import nltk will throw an error (as
it's not installed yet).
>>> import nltk Traceback (most recent call last): File "", line 1, in ModuleNotFoundError: No module named 'nltk'
pip is the command that downloads and install Python
packages such as nltk. The problem is that each install of Python has
its own pip, and if you install nltk to one, it's not visible to other
installs. What's the solution? Use the pip that is inside the Python
we are running now to install nltk. That guarantees it's the right
one. Command is pip.main(['install','nltk']) after first running
import pip.
>>> import pip
>>> pip.main(['install','nltk'])
WARNING: pip is being invoked by an old script wrapper. This will fail in a future version of pip.
Please see https://github.com/pypa/pip/issues/5599 for advice on fixing the underlying issue.
To avoid this problem you can invoke Python with '-m pip' instead of running pip directly.
Defaulting to user installation because normal site-packages is not writeable
Collecting nltk
Downloading nltk-3.8.1-py3-none-any.whl (1.5 MB)
|████████████████████████████████| 1.5 MB 9.2 MB/s
Collecting regex>=2021.8.3
Downloading regex-2024.5.15-cp39-cp39-macosx_10_9_x86_64.whl (281 kB)
|████████████████████████████████| 281 kB 30.4 MB/s
Collecting joblib
Downloading joblib-1.4.2-py3-none-any.whl (301 kB)
|████████████████████████████████| 301 kB 26.8 MB/s
Collecting tqdm
Downloading tqdm-4.66.4-py3-none-any.whl (78 kB)
|████████████████████████████████| 78 kB 10.3 MB/s
Collecting click
Downloading click-8.1.7-py3-none-any.whl (97 kB)
|████████████████████████████████| 97 kB 11.2 MB/s
Installing collected packages: tqdm, regex, joblib, click, nltk
WARNING: The script tqdm is installed in '/Users/sandiway/Library/Python/3.9/bin' which is not on PATH.
Consider adding this directory to PATH or, if you prefer to suppress this warning, use --no-warn-script-location.
WARNING: The script nltk is installed in '/Users/sandiway/Library/Python/3.9/bin' which is not on PATH.
Consider adding this directory to PATH or, if you prefer to suppress this warning, use --no-warn-script-location.
Successfully installed click-8.1.7 joblib-1.4.2 nltk-3.8.1 regex-2024.5.15 tqdm-4.66.4
WARNING: You are using pip version 21.2.4; however, version 24.1.2 is available.
You should consider upgrading via the '/Applications/SWI-Prolog.app/Contents/MacOS/swipl -m pip install --upgrade pip' command.
0
The above largely ignorable junk tells you version 3.8.1 of nltk is being
installed. Where does all that go? Well, it appear it goes into
/Users/sandiway/Library/Python/3.9/, which makes sense as
we're running embedded Python 3.9.6. What it doesn't tell you is that
nltk actually went into
/Users/sandiway/Library/Python/3.9/lib/python/site-packages
for me. Shell command (not Python, not Prolog) ls on this
directory will
show you something like this:
$ ls click nltk tqdm click-8.1.7.dist-info nltk-3.8.1.dist-info tqdm-4.66.4.dist-info joblib regex joblib-1.4.2.dist-info regex-2024.7.24.dist-info
How does Python 3.9.6 know where on your computer to look for the
nltk package? Well, there's a system variable called
sys.path. Typing it will show something like:
>>> sys.path ['', '/Applications/SWI-Prolog.app/Contents/swipl/library/ext/swipy/python', '/Applications/Xcode.app/Contents/Developer/Library/Frameworks/Python3.framework/Versions/3.9/lib/python39.zip', '/Applications/Xcode.app/Contents/Developer/Library/Frameworks/Python3.framework/Versions/3.9/lib/python3.9', '/Applications/Xcode.app/Contents/Developer/Library/Frameworks/Python3.framework/Versions/3.9/lib/python3.9/lib-dynload', '/Applications/Xcode.app/Contents/Developer/Library/Frameworks/Python3.framework/Versions/3.9/lib/python3.9/site-packages']
If you are not confident, it can find nltk (via Python command
import nltk), you can add it to sys.path as
follows:
>>>
sys.path.append('/Users/sandiway/Library/Python/3.9/lib/python/site-packages')
At this point, we've installed nltk for Python 3.9. It resides
somewhere in your user directory structure on this computer. And we've
added it to sys.path. So, now import nltk
followed by nltk.download() (for corpora including
WordNet) should work.
>>> import nltk >>> nltk.download() DEPRECATION WARNING: The system version of Tk is deprecated and may be removed in a future release. Please don't rely on it. Set TK_SILENCE_DEPRECATION=1 to suppress this warning. showing info https://raw.githubusercontent.com/nltk/nltk_data/gh-pages/index.xml
The above warning message is because the Tk graphics library on this laptop is out of date. But it still worked. And popped up a menu, and you can select the corpora you want. I recommend selecting all as you don't really want to do this again, do you?
After, it's done, WordNet should be somewhere on your computer. It
doesn't tell you where. It's in the directory
nltk_data/corpora under your home directory (of
course). The following is the result of running the Shell command
ls on ~/nltk_data (tilde: ~ is shorthand
for my home directory on the computer).
$ ls chunkers grammars misc sentiment taggers corpora help models stemmers tokenizers
We're done. Inside Python, you should be able load and run Wordnet as follows:
>>> from nltk.corpus import wordnet as wn
>>> wn.morphy('did','v')
'do'
The SMT parser uses wn.morphy to do stemming (regular
and irregular). Above, 'did' as a verb 'v'
is stemmed down to 'do'. Below is an example with a
plural noun.
>>> wn.morphy('corridors','n')
'corridor'
You can also call morphy from SWI-Prolog on the
command line with a slightly different syntax once you've installed
the parser.
An example of running Python via the py command in
PowerShell, and installing nltk (using pip from inside
Python) and nltk_data (via the popup menu) following the
instructions above.
