# TanaT

> TanaT (Temporal ANalysis of Trajectories) is a Python library for temporal sequence analysis, focused on patient care pathways. It supports multi-sequence trajectories combining events, intervals, and states.
> [HTML documentation](https://tanat.gitlabpages.inria.fr/core/tanat)
> Version: 0.0.post1

## Documentation Files

Documentation is available as plain text files optimized for LLMs:

- [Getting Started](https://tanat.gitlabpages.inria.fr/core/tanat/llms/getting-started.txt) (~2,967 tokens)
- [API Reference](https://tanat.gitlabpages.inria.fr/core/tanat/llms/reference.txt) (~11,539 tokens)
- [Examples](https://tanat.gitlabpages.inria.fr/core/tanat/llms/examples.txt) (~27,700 tokens)
- [Tutorials](https://tanat.gitlabpages.inria.fr/core/tanat/llms/tutorials.txt) (~14,105 tokens)
- [Community](https://tanat.gitlabpages.inria.fr/core/tanat/llms/community.txt) (~979 tokens)

- [All content combined](https://tanat.gitlabpages.inria.fr/core/tanat/llms-full.txt) (~57,290 tokens)

## Table of Contents

### Getting Started

- Installation
- Core Concepts
- First Steps

### Community

- Changelog
- Contributing
- Citing TanaT

### Examples

- Clustering: CLARA (Clustering LARge Applications)
- Clustering: Hierarchical Clustering
- Clustering: Partitioning Around Medoids (PAM)
- The Three Types of Sequences
- Trajectories
- EntityCriterion
- LengthCriterion
- PatternCriterion
- RankCriterion
- StaticCriterion
- TimeCriterion
- Custom Entity Metric
- Entity Metric: Hamming Distance
- Sequence Metrics: Chi²
- Custom Sequence Metric
- Sequence Metrics: DTW
- Sequence Metrics: Edit Distance
- Sequence Metrics: LCP
- Sequence Metrics: LCS
- Sequence Metrics: LinearPairwise
- Sequence Metrics: SoftDTW
- Trajectory Metrics: Aggregation Trajectory Metric
- Custom Trajectory Metric
- Barplot
- Distribution
- Spanplot
- Timeline
- Sequence Level Zeroing
- Trajectory-Level Zeroing

### Tutorials

- Discretizing time series into sequences
- Building pools from multiple sources
- Analysing and clustering cohort sequences
- Exploring a patient cohort
- Filtering and preparing a cohort
- Learning clinical temporal patterns with SWoTTeD
- Survival analysis from cohort clusters
- Clustering learning sessions by action patterns
- Exploring learner activity sequences

### Reference

- Builder
- Clustering
- Criterion
- Manipulation
- Metadata
- Metrics
- Zeroing