Fix delete functionality

[jenhagg]

Purpose

Fix error when deleting data, and handle clearing local data correctly.

What the code is doing

The self.path_config attribute needs to be set in the base class since we don't actually call the Delete constructor. Also, data is now cached in the same tree structure as the server, so changed the glob to match that.

I removed the implementation in DataAccess since that will fail when glob matches are directories instead of files, and didn't feel a strong need for it but I can put it back if people are using it.

Testing

Shell session testing the glob change

In [1]: from powersimdata.utility import server_setup

In [2]: import glob, os

In [3]: glob.glob(os.path.join(server_setup.LOCAL_DIR, "data", "**", "1714123_*"))
Out[3]: []

In [4]: glob.glob(os.path.join(server_setup.LOCAL_DIR, "data", "**", "1714_*"))
Out[4]:
['/Users/jonhagg/ScenarioData/data/input/1714_ct.pkl',
 '/Users/jonhagg/ScenarioData/data/input/1714_grid.mat',
 '/Users/jonhagg/ScenarioData/data/output/1714_AVERAGED_CONG.pkl',
 '/Users/jonhagg/ScenarioData/data/output/1714_PF.pkl']

In [5]: glob.glob(os.path.join(server_setup.LOCAL_DIR, "foo", "**", "1714_*"))
Out[5]: []

Separate session testing the delete state:

In [1]: from powersimdata.scenario.scenario import Scenario

In [2]: scenario = Scenario('1714')
Failed to download ScenarioList.csv from server
Falling back to local cache...
Failed to download ExecuteList.csv from server
Falling back to local cache...
SCENARIO: test | jon1

--> State
analyze
--> Loading grid
Loading bus
Loading plant
Loading heat_rate_curve
Loading gencost_before
Loading gencost_after
Loading branch
Loading sub
Loading bus2sub
--> Loading ct

In [3]: from powersimdata.scenario.delete import Delete

In [4]: scenario.change(Delete)
State switching: analyze --> delete

In [5]: scenario.state.path_config
Out[5]: <powersimdata.utility.server_setup.PathConfig at 0x1041aab80>

Time estimate

5 min

[rouille]

Since the class is not instantiated, I guess we could just remove the __init__, no?

[rouille]

This has been successfully tested. Thanks