Implement print to excel for PR3

[jvendries]

cc:
Closes:

What changed? Why?

Adds a new analysis module for exporting three-stage electricity pipeline matrices (V, extended U, and Y) at each checkpoint — after waste disaggregation, after electricity reallocation, and after electricity disaggregation — to an Excel workbook for offline PR3 inspection.

The disaggregation_matrices.py module introduces a PipelineStage dataclass to hold per-checkpoint tables, a _build_stage_tables() function that replicates the waste → reallocation → disaggregation pipeline read-only, and write_electricity_disaggregation_intermediate_outputs() which writes a 16-sheet workbook alongside GO weight and disaggregation weight CSVs. An electricity_balance sheet is included to verify preservation of key metrics (gross output, commodity totals, value added, final demand) across the reallocation → disaggregation boundary.

A CLI entry point (write_221100_disaggregation_matrices.py) is provided to invoke the export with a config YAML that has all three pipeline flags enabled. Generated output files are excluded from version control via .gitignore.

Testing

Unit tests cover:

• assert_disaggregation_export_config raises when any required flag is disabled
• write_electricity_disaggregation_intermediate_outputs writes the expected Excel workbook with all 16 sheets and both CSV files, and confirms the aggregate electricity sector (221100) is absent from the post-disaggregation V matrix
• _build_electricity_balance_summary correctly computes a zero delta and passing check for gross output row
• _derive_y_after_waste_disagg does not call the electricity row split function and preserves the aggregate electricity sector in the output index

[jvendries]

Warning

This pull request is not mergeable via GitHub because a downstack PR is open. Once all requirements are satisfied, merge this PR as a stack on Graphite.
Learn more

• Implement unit conversion #492
• Implement changes from resolutions from discussion #85 #490
• implement analysis for decisions 3, 5, and 7 in dicussion #85 #481
• Implement print to excel for PR3 #463 👈 (View in Graphite)
• Implement 221100 split #462
• main

This stack of pull requests is managed by Graphite. Learn more about stacking.

[bl-young]

My main concern with this approach is unintended opportunities for drift or deviation from the core pipeline given that this script to some extent hard codes some of the derivation. How different would it be to use the existing publish tools to get out what you need across 2 or 3 model configs?

[bl-young]

This feels like something that doesn't need additional tests

[bl-young]

is this necessary to have this again? wouldn't this be covered already?

[jvendries]

My main concern with this approach is unintended opportunities for drift or deviation from the core pipeline given that this script to some extent hard codes some of the derivation. How different would it be to use the existing publish tools to get out what you need across 2 or 3 model configs?

The main driver for this decision was to keep all the printing and analysis pathing independent from the core pipeline and contained within bedrock/analysis/electricity/ to avoid the confusion between intermediate outputs required for the electricity disaggregation, specifically the after waste tables.

A secondary consideration was ease of review for this PR vs #463 by avoiding any confusion on which printing was related to which PR.

What I'd propose is once #463 is merged, then we can re-work this PR to use existing publish tools. Would this be an acceptable course of action?

[WesIngwersen]

Successfully generated electricity_pipeline_stages_V_U_Y.xlsx