How to extract rentprice from text? - nlp

Question

For real estate objects I'm trying to extract rent prices from the description text.
E.g.:
"Für maximalen Komfort sorgt ein Concierge Service, der den Bewohnern im Haus zur Verfügung steht. <br/>Die Wohnung ist seit dem 15.10.2002 zu einer monatlichen Nettokaltmiete in Höhe von 546,25 EUR vermietet. Es bestehen weder Sperrfrist noch Vorkaufsrecht."
The important part is this here
Nettokaltmiete in Höhe von 546,25 EUR
What is a simple way to extract that information?
Caution: There is often other information in EUR - so we can not just search for EUR.
One of my ideas would be:
Split sentences. Check if a sentence includes EUR or €. Check if the sentence also says something about Miete/Jahresmiete and get then the next closest number in €.
I would like to know if there is a more elegant way. What are your thoughts on this? :)

Answer 1

Try using dependency parsing.
I wrote this example which should get you started:
import spacy
nlp = spacy.load("de_core_news_sm")
s = "Die Wohnung ist seit dem 15.10.2002 zu einer monatlichen Nettokaltmiete in Höhe von 546,25 EUR vermietet."
doc = nlp(s)
try:
l = [t for t in doc if 'miete' in t.text.lower() and t.pos_ == "NOUN"]
if l:
rent = l[0]
subtree = [t for t in rent.subtree]
euro = [t for t in subtree if t.text == "EUR" or t.text == "€"]
if euro:
price = euro[0].nbor(-1)
if price.like_num == True:
print(rent, ':', price)
except:
pass
output:
Nettokaltmiete : 546,25