Vietnamese NLP Toolkit
Project description
Open-source Vietnamese Natural Language Process Toolkit
Underthesea is:
🌊 A Vietnamese NLP toolkit. Underthesea is a suite of open source Python modules data sets and tutorials supporting research and development in Vietnamese Natural Language Processing. We provides extremely easy API to quickly apply pretrained NLP models to your Vietnamese text, such as word segmentation, part-of-speech tagging (PoS), named entity recognition (NER), text classification and dependency parsing.
🎁 Support Us! Every bit of support helps us achieve our goals. Thank you so much. 💝💝💝
🎉 New in v9.1.5! Conversational AI Agent is here! Use agent("Xin chào") to chat with an AI assistant specialized in Vietnamese NLP. Supports OpenAI and Azure OpenAI. 🚀✨
Installation
To install underthesea, simply:
$ pip install underthesea
✨🍰✨
Satisfaction, guaranteed.
Install with extras (note: use quotes in zsh):
$ pip install "underthesea[deep]" # Deep learning support
$ pip install "underthesea[voice]" # Text-to-Speech support
$ pip install "underthesea[agent]" # Conversational AI agent
Tutorials
Sentence Segmentation - Breaking text into individual sentences
📜
-
📜 Usage
>>> from underthesea import sent_tokenize >>> text = 'Taylor cho biết lúc đầu cô cảm thấy ngại với cô bạn thân Amanda nhưng rồi mọi thứ trôi qua nhanh chóng. Amanda cũng thoải mái với mối quan hệ này.' >>> sent_tokenize(text) [ "Taylor cho biết lúc đầu cô cảm thấy ngại với cô bạn thân Amanda nhưng rồi mọi thứ trôi qua nhanh chóng.", "Amanda cũng thoải mái với mối quan hệ này." ]
Text Normalization - Standardizing textual data representation
📜
-
📜 Usage
>>> from underthesea import text_normalize >>> text_normalize("Ðảm baỏ chất lựơng phòng thí nghịêm hoá học") "Đảm bảo chất lượng phòng thí nghiệm hóa học"
Word Segmentation - Dividing text into individual words
📜
-
📜 Usage
>>> from underthesea import word_tokenize >>> text = "Chàng trai 9X Quảng Trị khởi nghiệp từ nấm sò" >>> word_tokenize(text) ["Chàng trai", "9X", "Quảng Trị", "khởi nghiệp", "từ", "nấm", "sò"] >>> word_tokenize(sentence, format="text") "Chàng_trai 9X Quảng_Trị khởi_nghiệp từ nấm sò" >>> text = "Viện Nghiên Cứu chiến lược quốc gia về học máy" >>> fixed_words = ["Viện Nghiên Cứu", "học máy"] >>> word_tokenize(text, fixed_words=fixed_words) "Viện_Nghiên_Cứu chiến_lược quốc_gia về học_máy"
POS Tagging - Labeling words with their part-of-speech
📜
-
📜 Usage
>>> from underthesea import pos_tag >>> pos_tag('Chợ thịt chó nổi tiếng ở Sài Gòn bị truy quét') [('Chợ', 'N'), ('thịt', 'N'), ('chó', 'N'), ('nổi tiếng', 'A'), ('ở', 'E'), ('Sài Gòn', 'Np'), ('bị', 'V'), ('truy quét', 'V')]
Chunking - Grouping words into meaningful phrases or units
📜
-
📜 Usage
>>> from underthesea import chunk >>> text = 'Bác sĩ bây giờ có thể thản nhiên báo tin bệnh nhân bị ung thư?' >>> chunk(text) [('Bác sĩ', 'N', 'B-NP'), ('bây giờ', 'P', 'B-NP'), ('có thể', 'R', 'O'), ('thản nhiên', 'A', 'B-AP'), ('báo', 'V', 'B-VP'), ('tin', 'N', 'B-NP'), ('bệnh nhân', 'N', 'B-NP'), ('bị', 'V', 'B-VP'), ('ung thư', 'N', 'B-NP'), ('?', 'CH', 'O')]
Dependency Parsing - Analyzing grammatical structure between words
⚛️
-
⚛️ Deep Learning Model
$ pip install underthesea[deep]
>>> from underthesea import dependency_parse >>> text = 'Tối 29/11, Việt Nam thêm 2 ca mắc Covid-19' >>> dependency_parse(text) [('Tối', 5, 'obl:tmod'), ('29/11', 1, 'flat:date'), (',', 1, 'punct'), ('Việt Nam', 5, 'nsubj'), ('thêm', 0, 'root'), ('2', 7, 'nummod'), ('ca', 5, 'obj'), ('mắc', 7, 'nmod'), ('Covid-19', 8, 'nummod')]
Named Entity Recognition - Identifying named entities (e.g., names, locations)
📜 ⚛️
-
📜 Usage
>>> from underthesea import ner >>> text = 'Chưa tiết lộ lịch trình tới Việt Nam của Tổng thống Mỹ Donald Trump' >>> ner(text) [('Chưa', 'R', 'O', 'O'), ('tiết lộ', 'V', 'B-VP', 'O'), ('lịch trình', 'V', 'B-VP', 'O'), ('tới', 'E', 'B-PP', 'O'), ('Việt Nam', 'Np', 'B-NP', 'B-LOC'), ('của', 'E', 'B-PP', 'O'), ('Tổng thống', 'N', 'B-NP', 'O'), ('Mỹ', 'Np', 'B-NP', 'B-LOC'), ('Donald', 'Np', 'B-NP', 'B-PER'), ('Trump', 'Np', 'B-NP', 'I-PER')]
-
⚛️ Deep Learning Model
$ pip install underthesea[deep]
>>> from underthesea import ner >>> text = "Bộ Công Thương xóa một tổng cục, giảm nhiều đầu mối" >>> ner(text, deep=True) [ {'entity': 'B-ORG', 'word': 'Bộ'}, {'entity': 'I-ORG', 'word': 'Công'}, {'entity': 'I-ORG', 'word': 'Thương'} ]
Text Classification - Categorizing text into predefined groups
📜 ⚡
-
📜 Usage
>>> from underthesea import classify >>> classify('HLV đầu tiên ở Premier League bị sa thải sau 4 vòng đấu') ['The thao'] >>> classify('Hội đồng tư vấn kinh doanh Asean vinh danh giải thưởng quốc tế') ['Kinh doanh'] >> classify('Lãi suất từ BIDV rất ưu đãi', domain='bank') ['INTEREST_RATE']
-
⚡ Prompt-based Model
$ pip install underthesea[prompt] $ export OPENAI_API_KEY=YOUR_KEY
>>> from underthesea import classify >>> text = "HLV ngoại đòi gần tỷ mỗi tháng dẫn dắt tuyển Việt Nam" >>> classify(text, model='prompt') Thể thao
Sentiment Analysis - Determining text's emotional tone or sentiment
📜
-
📜 Usage
>>> from underthesea import sentiment >>> sentiment('hàng kém chất lg,chăn đắp lên dính lông lá khắp người. thất vọng') 'negative' >>> sentiment('Sản phẩm hơi nhỏ so với tưởng tượng nhưng chất lượng tốt, đóng gói cẩn thận.') 'positive' >>> sentiment('Đky qua đường link ở bài viết này từ thứ 6 mà giờ chưa thấy ai lhe hết', domain='bank') ['CUSTOMER_SUPPORT#negative'] >>> sentiment('Xem lại vẫn thấy xúc động và tự hào về BIDV của mình', domain='bank') ['TRADEMARK#positive']
Translation - Translating Vietnamese text to English
⚛️
-
⚛️ Deep Learning Model
$ pip install underthesea[deep]
>>> from underthesea import translate >>> translate("Hà Nội là thủ đô của Việt Nam") 'Hanoi is the capital of Vietnam' >>> translate("Ẩm thực Việt Nam nổi tiếng trên thế giới") 'Vietnamese cuisine is famous around the world' >>> translate("I love Vietnamese food", source_lang='en', target_lang='vi') 'Tôi yêu ẩm thực Việt Nam'
Lang Detect - Identifying the Language of Text
⚛️
Lang Detect API. Powered by FastText language identification model, using pure Rust inference via underthesea_core.
Usage examples in script
```python
>>> from underthesea import lang_detect
>>> lang_detect("Cựu binh Mỹ trả nhật ký nhẹ lòng khi thấy cuộc sống hòa bình tại Việt Nam")
vi
```
Text-to-Speech 🗣️ - Converting written text into spoken audio
⚛️
Text to Speech API. Thanks to awesome work from NTT123/vietTTS
Install extend dependencies and models
```bash
$ pip install "underthesea[voice]"
$ underthesea download-model VIET_TTS_V0_4_1
```
Usage examples in script
```python
>>> from underthesea.pipeline.tts import tts
>>> tts("Cựu binh Mỹ trả nhật ký nhẹ lòng khi thấy cuộc sống hòa bình tại Việt Nam")
A new audio file named `sound.wav` will be generated.
```
Usage examples in command line
```sh
$ underthesea tts "Cựu binh Mỹ trả nhật ký nhẹ lòng khi thấy cuộc sống hòa bình tại Việt Nam"
```
Conversational AI Agent 🤖 - Chat with AI for Vietnamese NLP tasks
⚡
Conversational AI Agent with OpenAI and Azure OpenAI support.
Install extend dependencies
```bash
$ pip install "underthesea[agent]"
$ export OPENAI_API_KEY=your_api_key
# Or for Azure OpenAI:
# export AZURE_OPENAI_API_KEY=your_key
# export AZURE_OPENAI_ENDPOINT=https://xxx.openai.azure.com
```
Usage examples in script
```python
>>> from underthesea import agent
>>> agent("Xin chào!")
'Xin chào! Tôi có thể giúp gì cho bạn?'
>>> agent("NLP là gì?")
'NLP (Natural Language Processing) là xử lý ngôn ngữ tự nhiên...'
>>> agent("Cho ví dụ về word tokenization tiếng Việt")
'Word tokenization trong tiếng Việt là quá trình...'
# Reset conversation
>>> agent.reset()
```
Supports Azure OpenAI
```python
>>> agent("Hello", provider="azure", model="my-gpt4-deployment")
```
Agent with Custom Tools (Function Calling)
```python
>>> from underthesea.agent import Agent, Tool
# Define tools as functions
>>> def get_weather(location: str) -> dict:
... """Get current weather for a location."""
... return {"location": location, "temp": 25, "condition": "sunny"}
>>> def search_news(query: str) -> str:
... """Search Vietnamese news."""
... return f"Results for: {query}"
# Create agent with tools
>>> my_agent = Agent(
... name="assistant",
... tools=[
... Tool(get_weather, description="Get weather for a city"),
... Tool(search_news, description="Search Vietnamese news"),
... ],
... instruction="You are a helpful Vietnamese assistant."
... )
# Agent automatically calls tools when needed
>>> my_agent("Thời tiết ở Hà Nội thế nào?")
'Thời tiết ở Hà Nội hiện tại là 25°C và nắng.'
>>> my_agent.reset() # Clear conversation history
```
Using Default Tools (like LangChain/OpenAI tools)
```python
>>> from underthesea.agent import Agent, default_tools
# Create agent with built-in tools:
# calculator, datetime, web_search, wikipedia, shell, python, file ops...
>>> my_agent = Agent(
... name="assistant",
... tools=default_tools,
... )
>>> my_agent("What time is it?") # Uses datetime tool
>>> my_agent("Calculate sqrt(144) + 10") # Uses calculator tool
>>> my_agent("Search for Python tutorials") # Uses web_search tool
```
Vietnamese NLP Resources
List resources
$ underthesea list-data
| Name | Type | License | Year | Directory |
|---------------------------+-------------+---------+------+------------------------------------|
| CP_Vietnamese_VLC_v2_2022 | Plaintext | Open | 2023 | datasets/CP_Vietnamese_VLC_v2_2022 |
| UIT_ABSA_RESTAURANT | Sentiment | Open | 2021 | datasets/UIT_ABSA_RESTAURANT |
| UIT_ABSA_HOTEL | Sentiment | Open | 2021 | datasets/UIT_ABSA_HOTEL |
| SE_Vietnamese-UBS | Sentiment | Open | 2020 | datasets/SE_Vietnamese-UBS |
| CP_Vietnamese-UNC | Plaintext | Open | 2020 | datasets/CP_Vietnamese-UNC |
| DI_Vietnamese-UVD | Dictionary | Open | 2020 | datasets/DI_Vietnamese-UVD |
| UTS2017-BANK | Categorized | Open | 2017 | datasets/UTS2017-BANK |
| VNTQ_SMALL | Plaintext | Open | 2012 | datasets/LTA |
| VNTQ_BIG | Plaintext | Open | 2012 | datasets/LTA |
| VNESES | Plaintext | Open | 2012 | datasets/LTA |
| VNTC | Categorized | Open | 2007 | datasets/VNTC |
$ underthesea list-data --all
Download resources
$ underthesea download-data CP_Vietnamese_VLC_v2_2022
Resource CP_Vietnamese_VLC_v2_2022 is downloaded in ~/.underthesea/datasets/CP_Vietnamese_VLC_v2_2022 folder
Up Coming Features
- Automatic Speech Recognition
Contributing
Do you want to contribute with underthesea development? Great! Please read more details at CONTRIBUTING.rst
💝 Support Us
If you found this project helpful and would like to support our work, you can just buy us a coffee ☕.
Your support is our biggest encouragement 🎁!
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
Filter files by name, interpreter, ABI, and platform.
If you're not sure about the file name format, learn more about wheel file names.
Copy a direct link to the current filters
File details
Details for the file underthesea-9.2.11.tar.gz.
File metadata
- Download URL: underthesea-9.2.11.tar.gz
- Upload date:
- Size: 7.0 MB
- Tags: Source
- Uploaded using Trusted Publishing? Yes
- Uploaded via: twine/6.1.0 CPython/3.13.7
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
524ed429d1697573259aa0bfaef9d9e1b22f63ec4cfe2f8fa0ad5368086830cd
|
|
| MD5 |
bff414d40549ecf6fa33dc5b80a802e4
|
|
| BLAKE2b-256 |
d7f04d0b1942c65ab41a0bd804df1f7a73bb74e2448f7590900401e7b117c4f6
|
Provenance
The following attestation bundles were made for underthesea-9.2.11.tar.gz:
Publisher:
release-pypi.yml on undertheseanlp/underthesea
-
Statement:
-
Statement type:
https://in-toto.io/Statement/v1 -
Predicate type:
https://docs.pypi.org/attestations/publish/v1 -
Subject name:
underthesea-9.2.11.tar.gz -
Subject digest:
524ed429d1697573259aa0bfaef9d9e1b22f63ec4cfe2f8fa0ad5368086830cd - Sigstore transparency entry: 925761248
- Sigstore integration time:
-
Permalink:
undertheseanlp/underthesea@3df03aea93e450e9b861393980e0f19f0a803bdb -
Branch / Tag:
refs/tags/underthesea-v9.2.11 - Owner: https://github.com/undertheseanlp
-
Access:
public
-
Token Issuer:
https://token.actions.githubusercontent.com -
Runner Environment:
github-hosted -
Publication workflow:
release-pypi.yml@3df03aea93e450e9b861393980e0f19f0a803bdb -
Trigger Event:
push
-
Statement type:
File details
Details for the file underthesea-9.2.11-py3-none-any.whl.
File metadata
- Download URL: underthesea-9.2.11-py3-none-any.whl
- Upload date:
- Size: 7.0 MB
- Tags: Python 3
- Uploaded using Trusted Publishing? Yes
- Uploaded via: twine/6.1.0 CPython/3.13.7
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
fa2d2f77fadcb6fb459e0846a4bfec23ca4977129fea29f8f1384da4f3a4aa61
|
|
| MD5 |
e2aa3a4f5cb8ab11eb89af1f60716c9e
|
|
| BLAKE2b-256 |
259dd74fcb393d045433b9b16917a7ab057a87b9fbd0a54d462753da6627059c
|
Provenance
The following attestation bundles were made for underthesea-9.2.11-py3-none-any.whl:
Publisher:
release-pypi.yml on undertheseanlp/underthesea
-
Statement:
-
Statement type:
https://in-toto.io/Statement/v1 -
Predicate type:
https://docs.pypi.org/attestations/publish/v1 -
Subject name:
underthesea-9.2.11-py3-none-any.whl -
Subject digest:
fa2d2f77fadcb6fb459e0846a4bfec23ca4977129fea29f8f1384da4f3a4aa61 - Sigstore transparency entry: 925761259
- Sigstore integration time:
-
Permalink:
undertheseanlp/underthesea@3df03aea93e450e9b861393980e0f19f0a803bdb -
Branch / Tag:
refs/tags/underthesea-v9.2.11 - Owner: https://github.com/undertheseanlp
-
Access:
public
-
Token Issuer:
https://token.actions.githubusercontent.com -
Runner Environment:
github-hosted -
Publication workflow:
release-pypi.yml@3df03aea93e450e9b861393980e0f19f0a803bdb -
Trigger Event:
push
-
Statement type: