Update Canvas sandbox dependencies

[sydneynotthecity]

The sandbox_update DAG broke because we changed the name of the upstream DAG that the workflow was dependent on. This PR addresses the name change and refactors the code to be less brittle.

• The DAG originally updated data based on current_date(). If the pipeline broke, backfilling missing data required manual intervention. The query is now date parameterized by batch_start_date and will backfill missing days because catchup=True.
• copy statements pointed to the internal dataset, hubble-261722.crypto_stellar_internal_2. Updated the statements to use crypto-stellar.crypto_stellar.
• Added the state table DAG as an additional dependency. We missed this

Note: We will probably need to run the sandbox_create DAG because the column order on the public tables differ from internal. It'll be easier to just recreate the tables so that the schemas match.

Future considerations:

• add data quality checks
• use a merge statement to ensure duplicate data is not inserted into the table (alternative is to delete/reinsert)
• utilize table clone feature to copy data instead of manually inserting new data via SQL