---
license: mit
---
The model uses only sign **ӏ** for explosive consonants (small cyrillic palochka letter)!


The model was teached by folloving David Dale's instructions for erzya language and using code from his repository. Commentaries in Russian were left untouched.

```python
import torch
from transformers import BertTokenizer, AutoModel
import numpy as np
import pandas as pd
import razdel
import matplotlib.pyplot as plt
from tqdm.auto import tqdm, trange
```
Download the model from Huggingface repository:
```python
model_name = 'NM-development/labse-en-ru-ce-prototype'
tokenizer = BertTokenizer.from_pretrained(model_name)
model = AutoModel.from_pretrained(model_name)
```
Assign files with the texts you want to split into parallel sentences:
```python
file_ru = None
file_nm = None

if file_ru is None or file_nm is None:
    nm_text = 'Ламро. Сахьт. Къена. Адам. Зуда. Вокха. Тӏулг.'
    ru_text = 'Горец. Час. Старый. Человек. Жена. Высокий. Камень.'
else:
    with open(file_nm, 'r') as f1, open(file_ru, 'r') as f2:
        nm_text = f1.read()
        ru_text = f2.read()
```
In the following section define auxillary functions for parallel sentence comparison:
```python
def embed(text):
    encoded_input = tokenizer(text, padding=True, truncation=True, max_length=128, return_tensors='pt')
    with torch.inference_mode():
        model_output = model(**encoded_input.to(model.device))
    embeddings = model_output.pooler_output
    embeddings = torch.nn.functional.normalize(embeddings)
    return embeddings[0].cpu().numpy()

def center_norm(v):
    v = v - v.mean(0)
    return v /  (v**2).sum(1, keepdims=True) ** 0.5


def center_dot(x, y):
    m = (x.sum(0) + y.sum(0)) / (x.shape[0] + y.shape[0])
    x = x - m
    y = y - m
    x =  x /  (x**2).sum(1, keepdims=True) ** 0.5
    y =  y /  (y**2).sum(1, keepdims=True) ** 0.5
    return np.dot(x, y.T)

def get_top_mean_by_row(x, k=5):
    m, n = x.shape
    k = min(k, n)
    topk_indices = np.argpartition(x, -k, axis=1)[:, -k:]
    rows, _ = np.indices((m, k))
    return x[rows, topk_indices].mean(1)

def align3(sims):
    
    #sims = np.dot(center_norm(orig_vecs), center_norm(sum_vecs).T) ** 3
    #sims = center_dot(orig_embeds, sum_embeds) #** 3

    rewards = np.zeros_like(sims)
    choices = np.zeros_like(sims).astype(int)  # 1: choose this pair, 2: decrease i, 3: decrease j

    # алгоритм, разрешающий пропускать сколько угодно пар, лишь бы была монотонность
    for i in range(sims.shape[0]):
        for j in range(0, sims.shape[1]):
            # вариант первый: выровнять i-тое предложение с j-тым
            score_add = sims[i, j]
            if i > 0 and j > 0:  # вот как тогда выровняются предыдущие 
                score_add += rewards[i-1, j-1]
                choices[i, j] = 1
            best = score_add
            if i > 0 and rewards[i-1, j] > best:
                best = rewards[i-1, j]
                choices[i, j] = 2
            if j > 0 and rewards[i, j-1] > best:
                best = rewards[i, j-1]
                choices[i, j] = 3
            rewards[i, j] = best
    alignment = []
    i = sims.shape[0] - 1
    j = sims.shape[1] - 1
    while i > 0 and j > 0:
        if choices[i, j] == 1:
            alignment.append([i, j])
            i -= 1
            j -= 1
        elif choices[i, j] == 2:
            i -= 1
        else:
            j -= 1
    return alignment[::-1]

def make_sents(text):
    sents = [s.text.replace('\n', ' ').strip() for p in text.split('\n\n') for s in razdel.sentenize(p)]
    sents = [s for s in sents if s]
    return sents
```
Firstly split your texts into sentences:
```python
sents_nm = make_sents(nm_text)
sents_ru = make_sents(ru_text)
```
Then embed all the chunks:
```python
emb_ru = np.stack([embed(s) for s in tqdm(sents_ru)])
emb_nm = np.stack([embed(s) for s in tqdm(sents_nm)])
```
Now compare sentenses' semanics vectors and build correlation heatmap:
```python
pen = np.array([[min(len(x), len(y)) / max(len(x), len(y)) for x in sents_nm] for y in sents_ru])
sims = np.maximum(0, np.dot(emb_ru, emb_nm.T)) ** 1 * pen

alpha = 0.2
penalty = 0.2
sims_rel = (sims.T - get_top_mean_by_row(sims) * alpha).T - get_top_mean_by_row(sims.T) * alpha - penalty

alignment = align3(sims_rel)

print(sum(sims[i, j] for i, j in alignment) / min(sims.shape))
plt.figure(figsize=(12, 6))
plt.subplot(1, 2, 1)
plt.imshow(sims_rel)
plt.subplot(1, 2, 2)
plt.scatter(*list(zip(*alignment)), s=5);
```
Finally, save the parallel corpus into a json file:
```python
nm_ru_parallel_corpus = pd.DataFrame({'nm_text' : [sents_nm[x[1]] for x in alignment], 'ru_text' : [sents_ru[x[0]] for x in alignment]})
corpus_filename = 'nm_ru_corpus.json'
with open(corpus_filename, 'w') as f:
    nm_ru_parallel_corpus.to_json(f, force_ascii=False, indent=4)
```