pentarosarium commited on
Commit
43f24b5
·
1 Parent(s): 00a84c6

mend spacy use

Browse files
Files changed (2) hide show
  1. app.py +4 -1
  2. requirements.txt +2 -1
app.py CHANGED
@@ -17,7 +17,10 @@ spacy.cli.download("ru_core_news_sm")
17
 
18
  class NewsProcessor:
19
  def __init__(self, similarity_threshold=0.75, time_threshold=24):
20
- self.nlp = ru_core_news_sm.load()
 
 
 
21
  self.embeddings = pipeline("feature-extraction",
22
  model="sentence-transformers/paraphrase-multilingual-mpnet-base-v2",
23
  token=st.secrets["hf_token"])
 
17
 
18
  class NewsProcessor:
19
  def __init__(self, similarity_threshold=0.75, time_threshold=24):
20
+ try:
21
+ self.nlp = spacy.load("ru_core_news_sm")
22
+ except:
23
+ self.nlp = spacy.load("en_core_web_sm")
24
  self.embeddings = pipeline("feature-extraction",
25
  model="sentence-transformers/paraphrase-multilingual-mpnet-base-v2",
26
  token=st.secrets["hf_token"])
requirements.txt CHANGED
@@ -5,4 +5,5 @@ transformers
5
  rapidfuzz
6
  openpyxl
7
  tqdm
8
- spacy
 
 
5
  rapidfuzz
6
  openpyxl
7
  tqdm
8
+ spacy
9
+ https://github.com/explosion/spacy-models/releases/download/en_core_web_sm-3.7.0/en_core_web_sm-3.7.0.tar.gz