Confusion about add_endings method

[qiaoyinglin19]

Hi,MashaPo.

The data preprocessing of your paper Automated Word Stress Detection in Russian stated like,

''if the previous word has less than three letters, we remove its word stress and concatenate it with the current word (for example, “te oblaka ́ ” [that-Pl.Nom cloud-Pl.Nom]). If the previ- ous word has 3 or more letters, we use the last three, since Russian endings are typically 2-3 letters long and derivational morphemes are usually located on the right periphery of the word.''

To my understanding, the begining of the sentence should be concatenated with _ as mark. Besizes, every word should be concatenated with context information no matter how many context letters there are.

Thus, I took sentence 'Шла речь о том что где-то в газете напечатан хороший рецепт творожника' as input and call method of add_endings.

However, the result wasn't consistent with the paper statement in two ways.

• Begining of the sentence was not concatenated with _

• The index is clearly 0, but the code doesn't go into the 'elif i == 0'

• The previous word which is less than 3 letters wasn't concatenated with the current word.

Your add_endings stated as follows.
def add_endings(wordlist): pluswords = [] for i, word in enumerate(wordlist): if not bool(re.search(REG, word)): # won't predict, just return (less then two syllables ) pluswords.append(word) elif i == 0 or wordlist[i - 1] == '_': pluswords.append('_' + word) else: context = wordlist[i - 1].replace("'", "") if len(context) < 3: ending = context else: ending = context[-3:] plusword = ending + '_' + word pluswords.append(plusword) return pluswords

It would be appreciated if you could help me with the problems aforementioned.