[Feature] CBS import data to s3 from new emails

[atalyaalon]

Describe the bug
Relevant issue (partially contains this issue): #808

See the following document: https://github.com/hasadna/anyway/blob/dev/docs/Architecture/CBS.md
See part 6:
Data Loading - Separate to multiple stages - see CBS ETL in process refactoring Make sure data is not loaded multiple times and that no duplicates are created
Current Flow: email -> s3 -> updated tables email -> s3: can be scheduled once a week / even a day s3 -> Data Tables: Needs to be scheduled when both accident type 1 and accident type 3 of that months are in s3 Explanation: Nowadays we pull the last data from last 4 emails and insert data to s3 (after deleting previous data), we need to pull only emails we didn't save to s3 - hence track on the emails we already read and not re-insert them. Optional: We can add CBS data versioning in s3 - right now we delete old data and insert new one.

Expected behavior
Checking email once a day, when a new email arrives that we didn't load to S3 (perhaps create a data versioning table as mentioned above), load its data to s3.