RiverTex is an open-source library for modeling and training different incremental word vector architectures proposed by the state-of-the-art.
It seeks to standardize many existing incremental word vector algorithms into a unified framework to provide a standardized interface and facilitate the development of new methods.
RiverTex provides two training paradigms:
-
learn_one, which trains one instance at a time; -
and
learn_many, which trains a mini-batch of instances at a time.
This allows for more efficient training of text representation models with text data streams.
RiverText also provides an interface similar to the river package, making it easy for developers to use the library to quickly
and easily train text representation models.
The official documentation can be found at this link.
RiverText is meant to work with Python 3.10 and above. Installation can be done via pip:
pip install rivertext
These package will be installed along with the package, in case these have not already been installed:
- nltk
- numpy
- river
- scikit_learn
- scipy
- torch
- tqdm
- word-embeddings-benchmarks
All unit tests are in the rivertext/tests folder. It uses pytest as a framework to run them.
To run the test, execute:
pytest tests
To check the coverage, run:
pytest tests --cov-report xml:cov.xml --cov rivertext
And then:
coverage report -m
The documentation is created using mkdocs and mkdocs-material. It can be found in the docs folder at the root of the project. First, you need to install:
pip install mkdocs
pip install "mkdocstrings[python]"
pip install mkdocs-material
Then, to compile the documentation, run:
mkdocs build
mkdocs serve
Please cite the following paper if you use this package in an academic publication:
G. Iturra-Bocaz and F. Bravo-Marquez RiverText: A Python Library for Training and Evaluating Incremental Word Embeddings from Text Data Stream. In Proceedings of the 46th International ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR 2023), Taipei, Taiwan.
@inproceedings{10.1145/3539618.3591908,
author = {Iturra-Bocaz, Gabriel and Bravo-Marquez, Felipe},
title = {RiverText: A Python Library for Training and Evaluating Incremental Word Embeddings from Text Data Streams},
year = {2023},
isbn = {9781450394086},
publisher = {Association for Computing Machinery},
address = {New York, NY, USA},
url = {https://doi.org/10.1145/3539618.3591908},
doi = {10.1145/3539618.3591908},
abstract = {Word embeddings have become essential components in various information retrieval and natural language processing tasks, such as ranking, document classification, and question answering. However, despite their widespread use, traditional word embedding models present a limitation in their static nature, which hampers their ability to adapt to the constantly evolving language patterns that emerge in sources such as social media and the web (e.g., new hashtags or brand names). To overcome this problem, incremental word embedding algorithms are introduced, capable of dynamically updating word representations in response to new language patterns and processing continuous data streams.This paper presents RiverText, a Python library for training and evaluating incremental word embeddings from text data streams. Our tool is a resource for the information retrieval and natural language processing communities that work with word embeddings in streaming scenarios, such as analyzing social media. The library implements different incremental word embedding techniques, such as Skip-gram, Continuous Bag of Words, and Word Context Matrix, in a standardized framework. In addition, it uses PyTorch as its backend for neural network training.We have implemented a module that adapts existing intrinsic static word embedding evaluation tasks for word similarity and word categorization to a streaming setting. Finally, we compare the implemented methods with different hyperparameter settings and discuss the results.Our open-source library is available at https://github.com/dccuchile/rivertext.},
booktitle = {Proceedings of the 46th International ACM SIGIR Conference on Research and Development in Information Retrieval},
pages = {3027–3036},
numpages = {10},
keywords = {data streams, word embeddings, incremental learning},
location = {Taipei, Taiwan},
series = {SIGIR '23}
}Please write to gabrieliturrab at ug.chile.cl for inquiries about the software. You are also welcome to do a pull request or publish an issue in the RiverText repository on Github.