How to count words in a Libre Office file using a script?

Question

I am trying to write a script that takes a folder with X .odt files and compute the number of words. It has to write it in a csv file, with the date.

I tried to do it using odfpy.

import odf
import glob
import pandas as pd
import os
from odf.opendocument import load as load_odt

filenames = []
word_counts = []
for f in glob.glob('*.odt'):
    print(f)
    doc = load_odt(f)
    if doc.text.hasChildNodes():
        n = 0
        for e in doc.text.childNodes:
            if ":text:" in e.qname[0]:
                words = [w for w in str(e).split(" ") if len(w) > 0]
                n += len(words)
            else:
                print(e.qname[0])
        filenames.append(f)
        word_counts.append(n)

df = pd.DataFrame({'date':[pd.Timestamp.now() for i in range(len(filenames))], 'filename':filenames, 'word_count':word_counts})
print(df)
csv_filename = 'word_count.csv'

it somehow works but there are some missing files from the CSV. Any ideas ?

Answer 1

It looks like this works :

import odf
import glob
import pandas as pd
import os
from odf.opendocument import load as load_odt

filenames = []
word_counts = []
for f in glob.glob('*.odt'):
    print(f)
    doc = load_odt(f)
    n = 0
    for e in doc.body.childNodes:
        if type(e) == odf.element.Text or type(e) == odf.element.Element:
            words = [w for w in str(e).split(" ") if len(w) > 0]
            n += len(words)
        else:
            print(type(e))
    
    filenames.append(f)
    word_counts.append(n)

df = pd.DataFrame({'date':[pd.Timestamp.now() for i in range(len(filenames))], 'filename':filenames, 'word_count':word_counts})
print(df)
csv_filename = 'word_count.csv'

df.to_csv(csv_filename, index = False, mode='a', header=not os.path.exists(csv_filename))
print(df.sum(axis = 0))

It's not exactly the same word count as LibreOffice but it will be enough.