Define ingestion workflow

[amindadgar]

We want our ingestion to be real time, so the data should be passed to a temporal workflow to be called and ingest the data whenever it was called.
Here's the data format

{
  communityId: str,
  platformId: str,
  text: str,
  metadata: dict,
  docId: str = str(uuid())
  excludedEmbedMetadataKeys: list[str] = [],
  excludedLlmMetadataKeys: list[str] = [],
  collectionName: str | None = None,
}

Note: In cases that the collectionName, was not passed in I would set the name using the previous collection name pattern, else I would persist the data into the collectionName. (passing the collectionName will override the default pattern of communityId_platformId)