Lingtrain Aligner — ML powered library for the accurate texts alignment.

Overview

Lingtrain Aligner

ML powered library for the accurate texts alignment in different languages.

Cover

Purpose

Main purpose of this alignment tool is to build parallel corpora using two or more raw texts in different languages. Texts should contain the same information (i.e., one text should be a translated analog oh the other text). E.g., it can be the Drei Kameraden by Remarque in German and the Three Comrades — it's translation into English.

Process

There are plenty of obstacles during the alignment process:

  • The translator could translate several sentences as one.
  • The translator could translate one sentence as many.
  • There are some service marks in the text
    • Page numbers
    • Chapters and other section headings
    • Author and title information
    • Notes

While service marks can be handled manually (the tool helps to detect them), the translation conflicts should be handled more carefully.

Lingtrain Aligner tool will do almost all alignment work for you. It matches the sentence pairs automatically using the multilingual machine learning models. Then it searches for the alignment conflicts and resolves them. As output you will have the parallel corpora either as two distinct plain text files or as the merged corpora in widely used TMX format.

Supported languages and models

Automated alignment process relies on the sentence embeddings models. Embeddings are multidimensional vectors of a special kind which are used to calculate a distance between the sentences. Supported languages list depend on the selected backend model.

  • distiluse-base-multilingual-cased-v2
    • more reliable and fast
    • moderate weights size — 500MB
    • supports 50+ languages
    • full list of supported languages can be found in this paper
  • LaBSE (Language-agnostic BERT Sentence Embedding)
    • can be used for rare languages
    • pretty heavy weights — 1.8GB
    • supports 100+ languages
    • full list of supported languages can be found here

Profit

  • Parallel corpora by itself can used as the resource for machine translation models or for linguistic researches.
  • My personal goal of this project is to help people building parallel translated books for the foreign language learning.
You might also like...
Sentence boundary disambiguation tool for Japanese texts (日本語文境界判定器)

Bunkai Bunkai is a sentence boundary (SB) disambiguation tool for Japanese texts. Quick Start $ pip install bunkai $ echo -e '宿を予約しました♪!まだ2ヶ月も先だけど。早すぎ

Code for EMNLP'21 paper "Types of Out-of-Distribution Texts and How to Detect Them"

Code for EMNLP'21 paper "Types of Out-of-Distribution Texts and How to Detect Them"

Neural text generators like the GPT models promise a general-purpose means of manipulating texts.

Boolean Prompting for Neural Text Generators Neural text generators like the GPT models promise a general-purpose means of manipulating texts. These m

Biterm Topic Model (BTM): modeling topics in short texts
Biterm Topic Model (BTM): modeling topics in short texts

Biterm Topic Model Bitermplus implements Biterm topic model for short texts introduced by Xiaohui Yan, Jiafeng Guo, Yanyan Lan, and Xueqi Cheng. Actua

This repository contains Python scripts for extracting linguistic features from Filipino texts.

Filipino Text Linguistic Feature Extractors This repository contains scripts for extracting linguistic features from Filipino texts. The scripts were

Text Classification in Turkish Texts with Bert
Text Classification in Turkish Texts with Bert

You can watch the details of the project on my youtube channel Project Interface Project Second Interface Goal= Correctly guessing the classification

Code for our paper
Code for our paper "Mask-Align: Self-Supervised Neural Word Alignment" in ACL 2021

Mask-Align: Self-Supervised Neural Word Alignment This is the implementation of our work Mask-Align: Self-Supervised Neural Word Alignment. @inproceed

A pytorch implementation of the ACL2019 paper
A pytorch implementation of the ACL2019 paper "Simple and Effective Text Matching with Richer Alignment Features".

RE2 This is a pytorch implementation of the ACL 2019 paper "Simple and Effective Text Matching with Richer Alignment Features". The original Tensorflo

Tensorflow Implementation of A Generative Flow for Text-to-Speech via Monotonic Alignment Search

Tensorflow Implementation of A Generative Flow for Text-to-Speech via Monotonic Alignment Search

Comments
  • File Already Exists

    File Already Exists

    Делаю docker pull lingtrain/aligner:v4 Загружаю текстовый файл и...

    image

    После вот такого предупреждения ничего не происходит Причём оно вылазит на любой текстовый файл

    opened by puffofsmoke 1
  • Fix XML creation:

    Fix XML creation:

    • prevent parent tag duplication for (langs, author, title)
    • add tags for tmx export
    • use 'direction' for splitting paragraphs
    • do not use bs4 (generates incorrect xml), change to lxml
    opened by BorisNA 0
  • A error when I use “splitter.split_by_sentences_wrapper”,please help check the error

    A error when I use “splitter.split_by_sentences_wrapper”,please help check the error

    when I use “splitted_from = splitter.split_by_sentences_wrapper(text1_prepared, lang_from)” return list,

    But I see that there will be a conflict when insert sqlite ,specific error:

    File "ling_test.py", line 36, in aligner.fill_db(db_path, splitted_from, splitted_to) File "lingtrain_aligner/aligner.py", line 498, in fill_db db.executemany("insert into languages(key, val) values(?,?)", [("from", lang_from), ("to", lang_to)]) sqlite3.InterfaceError: Error binding parameter 1 - probably unsupported type.

    opened by Amen-bang 5
  • Add text splitting into small parts

    Add text splitting into small parts

    The current version ignores the H1-H5 headers that were added by user. But when book was translate text from chapter 1 will be translate as a chapter 1 text into another language. You can use this fact and split a big text to small parts.

    Next idea - try split a big text to small blocks automatically: Select a few sentences from original text(for example 10 sentences) and using loop try to find translate block in the thanslated text.

    You can use the next psedocode:

    left_array = original_sentences[100:110]
    sum=[]
    for i=50;i<150 do:
       right_array_candidate=translated_sentences[i:i+10]
       sum[i]=sum(cosunuse_distance(left_array,right_array_candidate))
    
    rigth_array=get_index_with_max_value(sum)
    
    left_text_split_index=left_array[0]
    rigth_text_split_index=rigth_array[0]
    
    opened by AigizK 0
Releases(0.1.0)
Owner
Sergei Averkiev
Software Engineer. Eager to learn languages and machine learning approaches. Live in Moscow.
Sergei Averkiev
Augmenty is an augmentation library based on spaCy for augmenting texts.

Augmenty: The cherry on top of your NLP pipeline Augmenty is an augmentation library based on spaCy for augmenting texts. Besides a wide array of high

Kenneth Enevoldsen 124 Dec 29, 2022
This library is testing the ethics of language models by using natural adversarial texts.

prompt2slip This library is testing the ethics of language models by using natural adversarial texts. This tool allows for short and simple code and v

null 9 Dec 28, 2021
:id: A python library for accurate and scalable fuzzy matching, record deduplication and entity-resolution.

Dedupe Python Library dedupe is a python library that uses machine learning to perform fuzzy matching, deduplication and entity resolution quickly on

Dedupe.io 3.6k Jan 2, 2023
:id: A python library for accurate and scalable fuzzy matching, record deduplication and entity-resolution.

Dedupe Python Library dedupe is a python library that uses machine learning to perform fuzzy matching, deduplication and entity resolution quickly on

Dedupe.io 2.9k Feb 11, 2021
:id: A python library for accurate and scalable fuzzy matching, record deduplication and entity-resolution.

Dedupe Python Library dedupe is a python library that uses machine learning to perform fuzzy matching, deduplication and entity resolution quickly on

Dedupe.io 2.9k Feb 17, 2021
C.J. Hutto 3.8k Dec 30, 2022
Python package to easily retrain OpenAI's GPT-2 text-generating model on new texts

gpt-2-simple A simple Python package that wraps existing model fine-tuning and generation scripts for OpenAI's GPT-2 text generation model (specifical

Max Woolf 3.1k Jan 7, 2023
C.J. Hutto 2.8k Feb 18, 2021
Python package to easily retrain OpenAI's GPT-2 text-generating model on new texts

gpt-2-simple A simple Python package that wraps existing model fine-tuning and generation scripts for OpenAI's GPT-2 text generation model (specifical

Max Woolf 2.5k Feb 17, 2021