Sentence structure analysis

问题

I am trying to look at the structure similarity of sentences, specifically to the position of verbs, adj, nouns. For instance, I have three (or more) sentences which look likes as follows:

I ate an apple pie, yesterday. 
I ate an orange, yesterday.
I eat a lemon, today.

All of them starts with a pronoun (I) followed by a verb (ate/eat) and a noun (apple pie, orange, lemon) and, finally, an adverb (yesterday/tomorrow).

I would like to know if there is a way to identify the structure, i.e. PRONOUN VERB NOUN ADVERB in the sentence.

If I think of it as a pandas dataframe:

SENTENCE
    I ate an apple pie, yesterday. 
    I ate an orange, yesterday.
    I eat a lemon, today.

I would need to have something like as follows:

SENTENCE                                    STRUCTURE
    I ate an apple pie, yesterday.        PRONOUN VERB NOUN ADJECTIVE
    I ate an orange, yesterday.           PRONOUN VERB NOUN ADJECTIVE         
    I eat a lemon, today.                 PRONOUN VERB NOUN ADJECTIVE

Do you know how I can get this (or similar) results?

回答1:

Here is a simple example using spacy:

import spacy
import pandas as pd

# load english language model
nlp = spacy.load('en_core_web_sm',disable=['ner','textcat'])

text = "I ate an apple pie, yesterday."

# create spacy 
doc = nlp(text)
pos = ""
for token in doc:
    pos += token.pos_ + " "
    
# create dataframe
df = pd.DataFrame([[text, pos]], columns=['Sentence', 'Structure'])
print(df)

Output is:

                      Sentence                                  Structure
0  I ate an apple pie, yesterday.  PRON VERB DET NOUN NOUN PUNCT NOUN PUNCT

来源：https://stackoverflow.com/questions/63763542/sentence-structure-analysis

标签

python

nlp

nltk