Fix transform for `f1_retained_erng` xbrl + dbf

[aesharpe]

Updated tasklist

• [ ] Add a method to process_xbrl_metadata() that adds a completely new calculation to a field where one is missing
• [ ] Use this method to add calculations for retained_earnings_ferc1 to the transform

Table Notes

• The DBF table is almost entirely unstructured, with at least 6 freeform subsections, each of which has a subtotal -- the subtotal rows have well defined meanings, but only the row number and overall category is standardized for the rows making up the sections.
• XBRL has again broken out these unstructured tables into their own unstructured, sequenced tables containing a description + amount, with (presumably) the total from those tables matching the totals reported in the coordinating structured table. But we need to confirm that.
• The main table also has a start + end of year balance in the instant table retained_earnings_118_instant.
• There's also an unstructured instant table: retained_earnings_appropriated_118_instant with end of last year, end of this year dates.
• Good table for testing whether unstructured tables sum to reported totals, since everybody reports unstructured data.

Transformation Plan

• Row numbers don't change over the years, but do need to be labeled for DBF + XBRL concatenation.
• Extreme case of unstructured data with only aggregate values showing up in the structured table.

[zaneselvans]

Whoops, forgot there are a bunch of unstructured tables yet to go here.

[e-belfer]

In issue #2016 and #2581 we ran into some errors with this table. Here are some compiled notes:

• Has starting_balance and ending_balance and also amount. Unclear looking at the form what amount corresponds to. Mostly NA start/end balances. Looking at form it seems like it might be unstructured data? Something is wrong here. The primary affected FERC account is getting lost, the freeform item description is also being lost. I think we need to revisit this transform.
• In 2020 and earlier data, they seem to have put some ending_balance values in the amount column.

[e-belfer]

First-pass assessment of explosion-specific problems with this table:

• Fields are described in the metadata as including sums of other fields, but there are no calculations at all defined for this table in the XBRL metadata
• Unclear distinction between starting_balance, ending_balance and amount for some line items
• Item and primary account only found in processed dbf data, not in processed xbrl. This seems to accurately reflect the nature of the instant and duration tables.

[e-belfer]

The plan for the current iteration of this transform:

• [ ] add a calculation which takes the sub-totals and sums them to the totals

Out of scope:

• dealing with the unstructured line items or comparing them to the sub-totals