Skip to content

assignment-sets/docAgent-MCPServer

Repository files navigation

sample urls to test

https://storage.googleapis.com/doc-agent-buck-1/temp/hehe.txt
https://storage.googleapis.com/doc-agent-buck-1/temp/haha.md
https://storage.googleapis.com/doc-agent-buck-1/temp/Algorithms_Data-Structures.ppt
https://storage.googleapis.com/doc-agent-buck-1/temp/Algorithms_Data-Structures.pptx
https://storage.googleapis.com/doc-agent-buck-1/temp/Algorithms_Data%20Structures.docx
https://storage.googleapis.com/doc-agent-buck-1/temp/sampleImg.png
https://storage.googleapis.com/doc-agent-buck-1/temp/sample.pdf
https://storage.googleapis.com/doc-agent-buck-1/temp/Iris.csv
https://storage.googleapis.com/doc-agent-buck-1/temp/Iris.xlsx
https://storage.googleapis.com/doc-agent-buck-1/temp/converted.pdf
https://storage.googleapis.com/doc-agent-buck-1/temp/haha.pdf

handled types

txt, md, png, jpeg, pdf, csv, xlsx, pptx, docx, pkl ...

Docker container py runtime build script with container name(mandatory)

docker build -t py-runtime .

meta data

  • pdf compression: Ghostscript
  • image cocompression: pillow

conversions:

  • docx -> pdf : spire-doc

  • csv -> xlsx -> pdf : spire-xls

  • pptx -> pdf : Spire.Presentation

  • pdf -> docx: spire.pdf

  • pdf -> pptx: spire.pdf

  • pdf -> xlsx: spire.pdf

TODO:

  • use pydantic schema for tool desc in main file for clear non repeated code

  • Add more tools as required

  • fix bugs and file ext extracttion logic for presigned urls

  • Sync your server clock with NTP:

sudo timedatectl set-ntp true
On Docker containers, ensure the container inherits the host time.

About

mcp server with document utility tools implementation for document utility AI agent client

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors