# PAI python code

**URL:** <https://forum.anaconda.com/t/pai-python-code/85723>\
**Category:** Getting Started\
**Created:** [October 13, 2024, 7:37pm UTC](https://forum.anaconda.com/t/pai-python-code/85723 "2024-10-13T19:37:13Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![rvirtigo.lamson](https://dub1.discourse-cdn.com/anaconda/user_avatar/forum.anaconda.com/rvirtigo.lamson/32/5486_2.png) [@rvirtigo.lamson](https://forum.anaconda.com/u/rvirtigo.lamson)\
**Post date:** [October 13, 2024, 7:37pm UTC](https://forum.anaconda.com/t/pai-python-code/85723/1 "2024-10-13T19:37:13Z")

</div>

import nltk # SAT 10-12-24 with errors under terminal. This is the most recent code from PyCharm  
from nltk.tokenize import word\_tokenize  
from nltk.tree import Tree

# Ensure the necessary NLTK resources are available

def ensure\_nltk\_resources():  
resources = [‘punkt’, ‘averaged\_perceptron\_tagger’]  
for resource in resources:  
try:  
nltk.data.find(resource)  
except LookupError:  
nltk.download(resource)

ensure\_nltk\_resources()

# Define the CFG non-terminals and terminals

NONTERMINALS = “”"  
S → NP VP | S Conj S | VP  
NP → Det N | Det AdjP N | N | N PP | Det NP | Pro | PropN  
VP → V | V NP | V NP PP | V PP | Adv V NP PP | Aux V NP PP | V Conj VP | V AdvP  
PP → P NP  
AdjP → Adj | Adj AdjP  
AdvP → Adv | Adv AdvP  
“”"

TERMINALS = “”"  
Det → ‘a’ | ‘an’ | ‘the’  
N → ‘holmes’ | ‘armchair’ | ‘home’ | ‘thursday’ | ‘country’ | ‘walk’ | ‘mess’  
V → ‘sat’ | ‘chuckled’ | ‘had’ | ‘came’ | ‘walk’ | ‘was’ | ‘were’  
Adj → ‘red’ | ‘paint’ | ‘little’ | ‘dreadful’ | ‘complex’ | ‘country’  
P → ‘in’ | ‘on’ | ‘at’ | ‘with’  
Adv → ‘then’ | ‘later’ | ‘quickly’  
Conj → ‘and’  
Pro → ‘i’  
Aux → ‘did’  
PropN → ‘Thursday’ | ‘home’  
“”"

# Combine nonterminals and terminals to create a CFG

GRAMMAR = NONTERMINALS + TERMINALS  
cfg = nltk.CFG.fromstring(GRAMMAR)

# Define the parser using this grammar

parser = nltk.ChartParser(cfg)

def preprocess(sentence):  
“”"  
Preprocess the sentence to lowercased list of words, excluding non-alphabetic words.  
“”"  
tokens = word\_tokenize(sentence.lower())  
words = [word for word in tokens if any(char.isalpha() for char in word)]  
return words

def np\_chunk(tree):  
“”"  
Return a list of all noun phrase chunks in the sentence tree.  
A noun phrase chunk is defined as a NP subtree that does not itself contain any other NPs.  
“”"  
np\_chunks =   
for subtree in tree.subtrees():  
if subtree.label() == ‘NP’ and not any(child.label() == ‘NP’ for child in subtree if isinstance(child, Tree)):  
np\_chunks.append(subtree)  
return np\_chunks

# Example usage for testing purposes

if **name** == “ **main** ”:  
sentences = [  
“Holmes sat.”,  
“I had a country walk on Thursday and came home in a dreadful mess.”,  
“Holmes sat in the little red armchair.”,  
“He quickly walked on a bright Thursday afternoon.”  
]

```
for sentence in sentences:
    words = preprocess(sentence)
    print(f"Words: {words}")

    trees = list(parser.parse(words))
    if not trees:
        print(f"Could not parse sentence: \"{sentence}\".")
        continue

    for tree in trees:
        tree.pretty_print()
        chunks = np_chunk(tree)
        chunk_text = [" ".join(word for word, pos in chunk.leaves()) for chunk in chunks]
        print("NP Chunks:", chunk_text)

```