feat(document-search): parallelize document ingestion

[kdziedzic68]

Feature description

• Implement parallelization for document ingestion to speed up processing.
• Support configurable parallelization strategies (e.g., local multiprocessing, ray, remote endpoints on distributed machines etc.).

.

Motivation

We need fast and scalable, production ready process of document processing

Additional context

Out of the box python parallelization can be somehow problematic - each of the processing threads would either need to send requests to external services (eg. unstructured) which may cause throttling or will need to use multiple instances of processing models increasing gpu requirements.

.