* Add files via upload

Adding nest_asyncio.apply() to resolve RuntimeError: asyncio.run() cannot be called from a running event loop

* Delete experimental/openai_fine_tuning/openai_fine_tuning-2.ipynb

* Update openai_fine_tuning.ipynb with asyncio to fix runtime error

* first commit

update dependancy

update the model configuration

* fix CI issue

fix missing node postprocessor in CondensePlusContextChatEngine async

* wip

* wip return direct

* fix tests

* fully implement with docs

* more verbose logging

* even more verbose

* more docs

update json loader path

* Add in-memory loading for non-default filesystems in PDFReader

* bump version

fixes

* add async task stack

* just use asyncio.current_task

* wip

* delay asyncio Lock assignment in dispatcher

* fix async unit tests

* add async_postprocess_nodes and rankgpt test code

* fix typo and rename async functions

* use pytest.mark.asyncio

---------

Co-authored-by: jeffrey <vkefhdl1@gmail.com>

* Update base.py

I spent more time than what I am willing to admit to figure out why the code was not working.
Since the service account did not have READ access to the folder, and because if you try to read "0" docs, the code will break, I spent a lot of time debugging why it was failing.

* Updated readme.md

* Added multilingual support for the Wikipedia reader.

* fix: Added a check for the language prefix.

Checks if the provided language prefix is supported by the list of languages of Wikipedia.

* add anthropic tool calling

* update docs

Most of the other readers (including OneDrive which is the most similar) use
recursive file reading by default. Do the same thing here.

Also treat empty strings as the value not being present.

nb for vector pe lab talk