gcp-storage-emulator VS astro-sdk

What do you mean with running App Engine locally? There used to be a local development server, but that did not survive the 2to3 transition. GCP does ship a datastore and a pubsub emulator, but no Cloud Storage for example (someone wrote this one [1], which sort-of works but it does not behave 1:1 like GCS, so ironically my tests have to work around bugs in the emulator...).

[1] https://github.com/oittaa/gcp-storage-emulator

Have you tried the Astro SDK? https://github.com/astronomer/astro-sdk

One interesting point about putting the data into s3, is that if the data is in an S3 file then OP can use the Astro SDK to pretty easily upload that data into a table or a dataframe (there's even an s3 dynamic task function in the SDK that might fit the use-case well here).

I think you should take a look at the Astro SDK It’s an open source python package that removes the complexity of writing DAGs , particularly in the context of Extract, Load, Transform (ELT) use cases. Look at the doc here, especially aql.transform, aql.run_raw_sql, etc. That will definitely help you

- Astro SDK*: Made for folks who are doing their ETL in airflow and want to simplify movement between DBs and Pandas

More of a general principle but when you don't have design patterns, you get varying levels of results right? I think what Astro is doing to introduce "strong defaults" through projects like the astro-sdk or the cloud ide are interesting experiments to remove some of the busy work of common dags (load from s3, do something, push to database) will HELP reduce the cognitive load of really common, simple actions and give them a better single pattern to optimize on. I don't think those efforts reduce the optionality of true power users at all who want to custom code their s3 log sink to have some unique implementation while at the same time maybe solving some of the fragmentation to very frequently performed operations. 🤞

Have you looked into the astro python SDK? My team and I built this out over the last year to do exactly this :). You can you use the `@dataframe` decorator to pull the API data into a dataframe, store it in GCS and the access it in future steps. Lemme know if you have any questions!

I have an example here using COVID data. basically you just write a python function that reads the API and returns a dataframe (or any number of dataframes) and downstream tasks can then read the output as either a dataframe or a SQL table.