Skip to content

amitjaju82/tika_parser

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

1 Commit
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Tika Document Parser Plugin

Author: Amit Jaju (amitjaju@gmail.com)
License: MIT
GitHub: amitjaju82/tika_parser


Universal document parsing pipeline for Agent Zero, powered by Apache Tika.

What It Does

Turns any input file into extracted text + metadata by routing through a single Apache Tika service:

  • PDF (digital + scanned/OCR)
  • Office (DOCX, XLSX, PPTX, ODT, ODS, ODP)
  • Images (PNG, JPEG, TIFF, etc. -- OCR'd automatically)
  • Email (MSG, EML -- with attachments extracted)
  • Archives (ZIP, TAR, GZIP -- recursive extraction)
  • Web (HTML, XML)
  • And 1000+ other formats

Requirements

  • An Apache Tika server (4.0.0+) running and accessible from the Agent Zero host.
  • The default URL is http://192.168.0.167:9998 -- change it in the plugin settings.

Installation

Install from the Agent Zero Plugin Hub, or manually place this folder in:

/a0/usr/plugins/tika_parser/

Configuration

All settings are configurable via the plugin UI (Settings > Agent > Tika Document Parser) or environment variables:

Setting Env Var Default Purpose
Tika Server URL TIKA_URL http://192.168.0.167:9998 Tika base URL
Read Timeout TIKA_READ_TIMEOUT_S 600 HTTP read timeout (OCR can be slow)
Max Concurrency TIKA_MAX_CONCURRENCY 6 Max simultaneous requests
Max Input Size TIKA_MAX_INPUT_MB 100 Max input file size in MB
Retries TIKA_RETRIES 3 Retries on transient errors
Dedup Cache -- true SHA-256 dedup cache

Architecture

File -> [Size Guard] -> [Dedup Check] -> PUT /rmeta -> [Retry/Backoff] -> Assemble Result
                                       ^                              |
                              Concurrency Cap              Text + Metadata + Provenance

All Tika I/O goes through one client module (helpers/tika_client.py).

Tool: tika_parse

The plugin exposes a tika_parse tool with four actions:

Action Args Description
parse file_path Parse any local file
parse_url url Fetch + parse from URL
batch file_paths (array) Parse multiple files concurrently
health -- Check Tika service connectivity

Files

File Purpose
plugin.yaml Plugin manifest
helpers/tika_client.py Core async Tika client (single I/O point)
tools/tika_parse.py Agent-callable tool
api/tika_health.py Health check API handler
api/tika_parse_api.py Direct parse API handler
prompts/agent.system.tool.tika_parse.md Agent tool prompt
hooks.py Install/uninstall lifecycle hooks
webui/config.html Settings UI
default_config.yaml Default configuration

About

Agent Zero plugin: Universal document parsing via Apache Tika - PDF, Office, images (OCR), email, archives, and 1000+ formats

Resources

License

Stars

0 stars

Watchers

0 watching

Forks

Releases

No releases published

Packages

 
 
 

Contributors