|
|
--- |
|
|
license: mit |
|
|
language: |
|
|
- ru |
|
|
library_name: transformers |
|
|
--- |
|
|
|
|
|
Annoy vector db of Russian Synodal Translation (RST) of **Bible** |
|
|
|
|
|
Embdding by [intfloat/multilingual-e5-large](https://huggingface.co/intfloat/multilingual-e5-large) |
|
|
|
|
|
Ussage: |
|
|
``` |
|
|
rst_ind = AnnoyIndex(1024, metric='angular') |
|
|
rst_ind.load('rst.indx') |
|
|
rst_df = pandas.read_csv('rst_db.csv') |
|
|
|
|
|
embd = GetE5Emb("Как жить правильно и не совершить греха?") # see the origin e5-large usage |
|
|
|
|
|
indices = rst_ind.get_nns_by_vector(emb, 3) |
|
|
ind = random.choice(indices) |
|
|
str_ = rst_df.loc[ind].Text |
|
|
print(f"{str_} [{rst_df.loc[ind]['Book']}{rst_df.loc[ind]['Chapter']}:{rst_df.loc[ind]['Verse']}]") |
|
|
|
|
|
``` |
