Annif Wiki

Welcome to the Annif usage documentation!

🧑‍💻 Introduction & Getting Started

Get started with Annif, understand system requirements, and learn about supported corpus formats and configuration options.

• Getting started: The first steps
• System requirements: OS, hardware, and dependencies
• Subject vocabulary formats: Supported formats for subject vocabularies
• Corpus formats: Supported formats for collections of documents (with or without assigned subjects) used for training, evaluation and testing
• Project configuration: How to set up and manage projects
• Optional features and dependencies: Some Annif features and backends need installing extra dependencies
• Backward compatibility between releases: What parts of Annif can and cannot change
• Architecture: Overview of Annif’s design

🚀 Deployment

Guides for running Annif in production environments.

• Usage with Docker: Containerized deployment
• Running as a WSGI service: Deployment as a traditional web service

🖥️ User Interfaces

Explore the different ways to interact with Annif.

• Command line interface: For Annif projects administration
• Web user interface: For testing and simple end-user usage
• REST API: For integrating other systems

⚙️ Preprocessing & Supporting Features

Features to tweak text inputs and help in workflows.

• Analyzers: Tokenization, stemming, etc.
• Transforms: Truncation and language filtering
• Language detection: Automatic language identification with CLI or API methods
• Subject exclusion and inclusion: Exclude certain subjects from a vocabulary
• 🤗 Hugging Face Hub integration: Share and reuse models via Hugging Face

🎯 Optimization Techniques

Tips and tricks for improving results.

• Achieving good results: Best practices
• Reusing preprocessed training data: Save time and resources
• Generating synthetic training data: Augment your datasets

🧩 Backends

Annif supports a variety of backends (algorithms) for different use cases. See a comparison table and each backend’s page for details and configuration options.

• TF-IDF
• fastText
• Omikuji
• MLLM
• STWFSA
• YAKE
• SVC
• Ensemble
• PAV
• nn_ensemble
• HTTP
• Dummy

🛠️ Development & Contribution

For contributors.

• Development flow, branches and tags: Workflow and versioning
• Release process: How to make a new release
• Creating a new backend: Extend Annif with new algorithms

🆘 Troubleshooting & Support

• Annif-users mailing list/forum: Community support ↗️
• GitHub Issues: Create an issue here to report bugs or request features ↗️
• Security Policy: Guidelines for reporting vulnerabilities and security issues ↗️

---

Getting started →