doccano

Open source annotation tool for machine learning practitioners.

9,385

1,712

9,385

340

View on GitHub

Top Related Projects

label-studio

18,167

Label Studio is a multi-type data labeling and annotation tool with standardized output format

snorkel

5,780

A system for quickly generating training data with weak supervision

Quick Overview

Doccano is an open-source text annotation tool for human-in-the-loop machine learning. It provides a web-based interface for collaborative annotation tasks such as text classification, sequence labeling, and sequence-to-sequence tasks. Doccano is designed to be user-friendly and easily deployable for various natural language processing projects.

Pros

User-friendly interface with a modern, responsive design
Supports multiple annotation types (classification, sequence labeling, seq2seq)
Easy to deploy and scale using Docker
Collaborative features for team-based annotation projects

Cons

Limited built-in machine learning integration
Requires some technical knowledge for setup and customization
May lack advanced features found in some commercial annotation tools
Documentation could be more comprehensive for advanced use cases

Getting Started

To get started with Doccano, follow these steps:

Install Docker and Docker Compose

Clone the repository:

git clone https://github.com/doccano/doccano.git
cd doccano

Run Doccano using Docker Compose:

docker-compose -f docker-compose.prod.yml up

Access the Doccano interface at http://localhost:8000
Create an admin account and start your annotation project

For more detailed instructions and configuration options, refer to the official Doccano documentation.

Competitor Comparisons

label-studio

18,167

Label Studio is a multi-type data labeling and annotation tool with standardized output format

Pros of Label Studio

More versatile with support for a wider range of data types and annotation tasks
Offers a more extensive set of pre-built templates and customization options
Provides integration with popular machine learning frameworks and cloud services

Cons of Label Studio

Steeper learning curve due to its more complex interface and feature set
Requires more system resources, potentially leading to slower performance on large datasets
Some advanced features are only available in the enterprise version

Code Comparison

Label Studio configuration example:

<View>
  <Image name="image" value="$image"/>
  <RectangleLabels name="label" toName="image">
    <Label value="Car" background="red"/>
    <Label value="Pedestrian" background="blue"/>
  </RectangleLabels>
</View>

Doccano configuration example:

{
  "labels": [
    {"text": "Car", "color": "#FF0000"},
    {"text": "Pedestrian", "color": "#0000FF"}
  ]
}

Both projects use different configuration formats, with Label Studio using XML-based syntax for more complex layouts, while Doccano opts for a simpler JSON structure.

snorkel

5,780

A system for quickly generating training data with weak supervision

Pros of Snorkel

Programmatic labeling approach allows for more scalable and flexible data labeling
Supports weak supervision techniques, enabling the use of multiple noisy labeling sources
Integrates machine learning models for label aggregation and denoising

Cons of Snorkel

Steeper learning curve due to its programmatic nature
Requires more setup and configuration compared to Doccano's straightforward interface
May be overkill for simple annotation tasks or small datasets

Code Comparison

Snorkel (labeling function example):

@labeling_function()
def keyword_labeling(text):
    if "positive" in text.lower():
        return POSITIVE
    elif "negative" in text.lower():
        return NEGATIVE
    return ABSTAIN

Doccano (JSON configuration example):

{
  "labels": [
    {"text": "Positive", "prefix_key": "p"},
    {"text": "Negative", "prefix_key": "n"}
  ]
}

Both projects aim to facilitate data labeling, but they take different approaches. Snorkel focuses on programmatic labeling and weak supervision, while Doccano provides a user-friendly interface for manual annotation. Snorkel is more suitable for large-scale labeling tasks and complex scenarios, whereas Doccano excels in simplicity and ease of use for straightforward annotation projects.

Convert designs to code with AI

Introducing Visual Copilot: A new AI model to turn Figma designs to high quality code using your components.

Try Visual Copilot

README

doccano

doccano is an open-source text annotation tool for humans. It provides annotation features for text classification, sequence labeling, and sequence to sequence tasks. You can create labeled data for sentiment analysis, named entity recognition, text summarization, and so on. Just create a project, upload data, and start annotating. You can build a dataset in hours.

Demo

Try the annotation demo.

Demo image

Documentation

Read the documentation at https://doccano.github.io/doccano/.

Features

Collaborative annotation
Multi-language support
Mobile support
Emoji :smile: support
Dark theme
RESTful API

Usage

There are three options to run doccano:

pip (Python 3.8+)
Docker
Docker Compose

pip

To install doccano, run:

pip install doccano

By default, SQLite 3 is used for the default database. If you want to use PostgreSQL, install the additional dependencies:

pip install 'doccano[postgresql]'

and set the DATABASE_URL environment variable according to your PostgreSQL credentials:

DATABASE_URL="postgres://${POSTGRES_USER}:${POSTGRES_PASSWORD}@${POSTGRES_HOST}:${POSTGRES_PORT}/${POSTGRES_DB}?sslmode=disable"

After installation, run the following commands:

# Initialize database.
doccano init
# Create a super user.
doccano createuser --username admin --password pass
# Start a web server.
doccano webserver --port 8000

In another terminal, run the command:

# Start the task queue to handle file upload/download.
doccano task

Go to http://127.0.0.1:8000/.

Docker

As a one-time setup, create a Docker container as follows:

docker pull doccano/doccano
docker container create --name doccano \
  -e "ADMIN_USERNAME=admin" \
  -e "ADMIN_EMAIL=admin@example.com" \
  -e "ADMIN_PASSWORD=password" \
  -v doccano-db:/data \
  -p 8000:8000 doccano/doccano

Next, start doccano by running the container:

docker container start doccano

Go to http://127.0.0.1:8000/.

To stop the container, run docker container stop doccano -t 5. All data created in the container will persist across restarts.

If you want to use the latest features, specify the nightly tag:

docker pull doccano/doccano:nightly

Docker Compose

You need to install Git and clone the repository:

git clone https://github.com/doccano/doccano.git
cd doccano

Note for Windows developers: Be sure to configure git to correctly handle line endings or you may encounter status code 127 errors while running the services in future steps. Running with the git config options below will ensure your git directory correctly handles line endings.

git clone https://github.com/doccano/doccano.git --config core.autocrlf=input

Then, create an .env file with variables in the following format (see ./docker/.env.example):

# platform settings
ADMIN_USERNAME=admin
ADMIN_PASSWORD=password
ADMIN_EMAIL=admin@example.com

# rabbit mq settings
RABBITMQ_DEFAULT_USER=doccano
RABBITMQ_DEFAULT_PASS=doccano

# database settings
POSTGRES_USER=doccano
POSTGRES_PASSWORD=doccano
POSTGRES_DB=doccano

After running the following command, access http://127.0.0.1/.

docker-compose -f docker/docker-compose.prod.yml --env-file .env up

One-click Deployment

Service	Button
AWS¹
Heroku

FAQ

See the documentation for details.

Contribution

As with any software, doccano is under continuous development. If you have requests for features, please file an issue describing your request. Also, if you want to see work towards a specific feature, feel free to contribute by working towards it. The standard procedure is to fork the repository, add a feature, fix a bug, then file a pull request that your changes are to be merged into the main repository and included in the next release.

Here are some tips might be helpful. How to Contribute to Doccano Project

Citation

@misc{doccano,
  title={{doccano}: Text Annotation Tool for Human},
  url={https://github.com/doccano/doccano},
  note={Software available from https://github.com/doccano/doccano},
  author={
    Hiroki Nakayama and
    Takahiro Kubo and
    Junya Kamura and
    Yasufumi Taniguchi and
    Xu Liang},
  year={2018},
}

Contact

For help and feedback, feel free to contact the author.

(1) EC2 KeyPair cannot be created automatically, so make sure you have an existing EC2 KeyPair in one region. Or create one yourself. (2) If you want to access doccano via HTTPS in AWS, here is an instruction. ↩

Top Related Projects

label-studio

18,167

Label Studio is a multi-type data labeling and annotation tool with standardized output format

snorkel

5,780

A system for quickly generating training data with weak supervision

Convert designs to code with AI

Introducing Visual Copilot: A new AI model to turn Figma designs to high quality code using your components.

Try Visual Copilot

doccano

Top Related Projects

label-studio

snorkel

Quick Overview

Pros

Cons

Getting Started

Competitor Comparisons

label-studio

Pros of Label Studio

Cons of Label Studio

Code Comparison

snorkel

Pros of Snorkel

Cons of Snorkel

Code Comparison

Convert designs to code with AI

README

doccano

Demo

Documentation

Features

Usage

pip

Docker

Docker Compose

One-click Deployment

FAQ

Contribution

Citation

Contact

Footnotes

Top Related Projects

label-studio

snorkel

Convert designs to code with AI