unstructured/requirements/ingest/embed-openai.in

-c ../constraints.in
-c ../base.txt
langchain
tiktoken
openai
Roman/optimize ingest ci (#1799) ### Description Currently the CI caches the CI dependencies but uses the hash of all files in `requirements/`. This isn't completely accurate since the ingest dependencies are installed in a later step and don't affect the cached environment. As part of this PR: * ingest dependencies were isolated into their own folder in `requirements/ingest/` * A new cache setup was introduced in the CI to restore the base cache -> install ingest dependencies -> cache it with a new id * new make target created to install all ingest dependencies via `pip install -r ...` * updates to Dockerfile to use `find ...` to install all dependencies, avoiding the need to update this when new deps are added. * update to pip-compile script to run over all `*.in` files in `requirements/` 2023-10-24 10:54:00 -04:00			`-c ../constraints.in`
			`-c ../base.txt`
			`langchain`
			`tiktoken`
			`openai`