xlsx working, room for improvments

added docx support

improved chunking: merged_item mode, chunk-debug path, subprocess worker

Improve docling chunking: hierarchical cleanup, table heuristics, config flags

added batching on upsert

fixed markdown has attribute bug

added some logging and disabled docx

removed doc id from project id

added ping endpoint and namespace creation on /vectorize call

updated logging to BE

??

added batching on embedding for big files

added .xlsx and .docx support

solved pinecone api bug

added fallback to 1 markdown chunk

improved logging

disabled picture classification

fixed some dependency bug of missed repo

set env var of OMP_NUM_THREADS and solved dependency issue

testing with more classification models

again :(

fixed docker bug

updated dockerfile to install CUDA drivers

fixed GPU inference auto option

pushed some broke code fix tomorrow

Ignore PDFs in app/engine

Stop tracking test PDF; add to gitignore

init push

initial commit