Advanced Topics

Table of Contents

This section covers advanced usage and internals of Apache Tika.

Topics

Language Detection - Built-in bigram language detector, training pipeline, and comparison with OpenNLP
Building the Language Detector - Full training log, decisions, and benchmark results vs OpenNLP
Robustness - Process isolation and fault tolerance when parsing untrusted content
Setting Limits - Limiting metadata size and resource consumption when processing untrusted content
TikaInputStream and Spooling - Understanding how TikaInputStream handles buffering, caching, and spooling to disk
Embedded Document Metadata - Understanding how Tika tracks embedded documents and their paths
ZIP Detection and Salvaging - How Tika detects and recovers truncated ZIP-based files
Running a Local VLM Server - Run an open-source VLM locally as an OpenAI-compatible endpoint for air-gapped OCR