adds cache support

[brunocalza]

Adds a --cache flag to the publication create command. It means how many minutes the data will stay in cache.

That information is sent to the provider. See tablelandnetwork/basin-provider#16

Also, it adds a new column EXIPIRES AT to the listing of files

+-------------------------------------------------------------+---------+---------------------------+----------+---------------------+
|                             CID                             |  SIZE   |         TIMESTAMP         | ARCHIVED |     EXPIRES AT      |
+-------------------------------------------------------------+---------+---------------------------+----------+---------------------+
| bafybeifr5njnrw67yyb2h2t7k6ukm3pml4fgphsxeurqcmgmeb7omc2vlq | 1613287 | 2023-11-17T16:29:20-03:00 | Y        | 2023-11-24 05:52:36 |
| bafybeifr5njnrw67yyb2h2t7k6ukm3pml4fgphsxeurqcmgmeb7omc2vlq | 1613287 | 2023-11-17T16:29:20-03:00 | Y        |                     |
| bafybeifr5njnrw67yyb2h2t7k6ukm3pml4fgphsxeurqcmgmeb7omc2vlq | 1613287 | 2023-11-17T16:29:20-03:00 | Y        |                     |
+-------------------------------------------------------------+---------+---------------------------+----------+---------------------+

Note: no changes were made to the retrieve command. Theoretically, it should check if the file is in cache or not and make a decision. But there's no way for the CLI to implement that without hitting the provider to check that and download the file from GCS. That would mean implementing a download method in capnp, which seems too much effort for something we're moving away from. We have two options:

• Make GCS bucket public, so that the CLI can hit the bucket direct and check if it's in cache
• When adding an HTTP API, add a download endpoint

[sanderpick]

What will happen if retrieve command is used before the file is archived?

Nit: In the table you showed, it's not clear what is expiring. It could interpretted as if the archived data is expiring. how about "Cache Expiry"

[brunocalza]

What will happen if retrieve command is used before the file is archived?

Right, now it will work just fine because the file is also in IPFS.

Nit: In the table you showed, it's not clear what is expiring. It could interpretted as if the archived data is expiring. how about "Cache Expiry"

Good suggestion 👍