Add baseline validation tool to resstockpostproc by rajeee · Pull Request #1529 · NatLabRockies/resstock

rajeee · 2025-11-05T17:54:55Z

Pull Request Description

This tool will compare ResStock baseline run with various truth data sources such as EIA 861, EIA 176, RECS, LRD and others and generate comparison graphics and dashboard.

See more details here in README.

Live dashboard is here.

Validation test on CI

The baseline-validation CI job runs python resstockpostproc/baseline_validation/main.py --test --no-parallel end-to-end on every PR. This exercises the full pipeline — BSQ/Athena queries, RECS/EIA/LRD truth-data loads, RECS uncertainty computation, and plot rendering for the test subset — and uploads the generated dashboard tree as a CI artifact (baseline-validation-plots.tar.gz) for reviewer inspection. It only contains a small fraction of the plots compared to when doing the full run without --test flag.

The test run hits real AWS sources (Athena in the rescore workgroup, s3://resstock-core/ for truth data, s3://oedi-data-lake/ for ResStock baseline parquets), but the data exchanged is small (~100 MB from the paid resstock-core bucket; the rest is free OEDI egress), so the AWS cost per CI run is negligible (roughly $0.01 or less).

Checklist

Required:

Add to the changelog_dev.rst file
No unexpected regression test changes on CI (comparison artifacts have been checked)
Update documentation (one is required)
- Technical reference guide
- Technical development guide

Optional (not all items may apply):

Tests (and test files) have been updated
Supporting resource files have been updated (related to resstock-estimation)
openstudio tasks.rb update_measures has been run

Delete unused plotters (eia_plotter, base_plotter, timeseries_plotter, monthly_plotter), empty stubs (eiaid_mapping), and obsolete tests (test_eia_viz). Fix test_data column assertions to match current DataCol-normalized output names. Uncomment utility_vertical rename in lrd_plotter for day_of_year resolution.

- README: match upgrade_comparison's declarative tone; cross-reference it - Add thorough field-by-field Configuration section; remove duplicate Configuration Details block that repeated the Input YAML - Replace 'Plot Types' bullet lists with a link to plot_showcase.md - main.py: --help description updated to match - New plot_showcase.md gallery covers 9 plot families with 36 images (dashboard + EIA + RECS + cross-filter dims + box plots + histograms + difference views + LRD) - All images constrained via <img width> so single-entity plots don't dominate the page (400 px), with 700 px for portrait histograms and 900 px for multi-panel tilemaps - Exclude .svg and .png from typos pre-commit hook; Plotly-generated SVGs have random short substrings (e.g. 'daa') in internal IDs that aren't real typos

GitHub auto-renders a README.md when the user navigates to a directory, so clicking into example_plots/ in the GitHub web UI now shows the gallery directly without needing a sibling file. Strip the 'example_plots/' prefix from all in-gallery image paths since they're now relative to the gallery itself, and point the top-level README's link at the new location.

- Remove unused 'workflow: WorkflowConfig' parameter from all 5 plotters (bar, box, heatmap, histogram, choropleth). The param was added but never read in any plotter body; two of its three callers (all_plots_generator, highlights_generator) were never updated to pass it, which silently broke those paths. - Update the dashboard caller to drop the 'workflow=run_workflow' kwarg accordingly so all 3 call sites use a uniform 2-arg signature. - Revert the default port change (8052 -> 8051) in dynamic_dashboard.py — a personal-preference drive-by unrelated to the shared refactor. All other upgrade_comparison changes on this branch (theme/colors/ end_use_dicts moves to shared_utils, local plotter-helper dedup, diagonal -> diagonal_relaxed in input_manager) are legitimate and retained.

Copilot

Pull request overview

Adds a new baseline validation tool under resstockpostproc to compare ResStock baseline runs against external datasets (EIA, RECS, LRD, etc.), generate plots, and assemble a browsable dashboard. The PR also introduces shared utilities (caching, S3 helpers, generic plotters) and refactors parts of upgrade_comparison to reuse them.

Changes:

Add baseline_validation package: schema, data loading/processing, plotters, dashboard generation, and extensive pytest coverage.
Add shared utilities (shared_utils) for caching, S3 downloads, sorting/mapping/colors, and reusable plotter helpers.
Refactor upgrade_comparison plotters/schema to use shared plot theme and shared end-use dictionaries.

Reviewed changes

Copilot reviewed 43 out of 157 changed files in this pull request and generated 9 comments.

Show a summary per file

File	Description
postprocessing/tests/test_caching.py	Adds unit tests for new disk-cache root and `@cached` decorator behavior.
postprocessing/ruff.toml	Updates Ruff ignores and adds E501 per-file ignores for HTML/JS-heavy dashboard modules.
postprocessing/resstockpostproc/upgrade_comparison/schema/workflow_schema.py	Switches end-use dict import to shared_utils location.
postprocessing/resstockpostproc/upgrade_comparison/plotters/histogram_plotter.py	Uses shared theme and new palette builder.
postprocessing/resstockpostproc/upgrade_comparison/plotters/heatmap_plotter.py	Uses shared theme module.
postprocessing/resstockpostproc/upgrade_comparison/plotters/choropleth_plotter.py	Uses shared theme module.
postprocessing/resstockpostproc/upgrade_comparison/plotters/box_plotter.py	Replaces local box plot implementation with shared generic box plotter.
postprocessing/resstockpostproc/upgrade_comparison/plotters/bar_plotter.py	Replaces local bar plot implementation with shared generic bar plotter.
postprocessing/resstockpostproc/upgrade_comparison/io_managers/input_manager.py	Changes Polars concat mode to `diagonal_relaxed`.
postprocessing/resstockpostproc/upgrade_comparison/data_processing/data_processor.py	Switches end-use dict import to shared_utils location.
postprocessing/resstockpostproc/upgrade_comparison/dashboard/callbacks/plotting.py	Calls plotting function with keyword arg for `plot_spec`.
postprocessing/resstockpostproc/upgrade_comparison/.gitignore	Ignores `sdr_plots/` output folder for this submodule.
postprocessing/resstockpostproc/shared_utils/sorting.py	Adds shared “human sort” helper with custom sort orders for dashboard labels.
postprocessing/resstockpostproc/shared_utils/s3_manager.py	Adds shared S3 download + local caching utilities.
postprocessing/resstockpostproc/shared_utils/mapping.py	Adds shared month/state/utility mappings used by baseline validation.
postprocessing/resstockpostproc/shared_utils/generic_plotters/theme.py	Refactors theme to use shared color constants and adds category palette logic.
postprocessing/resstockpostproc/shared_utils/generic_plotters/range_utils.py	Adds shared axis range computation helper.
postprocessing/resstockpostproc/shared_utils/generic_plotters/hover_formatting.py	Adds shared hover formatting utilities (compact/precise/CI/count).
postprocessing/resstockpostproc/shared_utils/generic_plotters/box_plotter.py	Adds shared Plotly box/violin plot renderer used by baseline validation and upgrade comparison.
postprocessing/resstockpostproc/shared_utils/colors.py	Adds shared qualitative color series and fuel color mapping.
postprocessing/resstockpostproc/shared_utils/caching.py	Adds disk-backed caching decorator and repo-local cache root.
postprocessing/resstockpostproc/shared_utils/init.py	Introduces `shared_utils` package marker/docstring.
postprocessing/resstockpostproc/baseline_validation/workflow.yaml	Adds baseline validation workflow config example (data sources, labels, outputs).
postprocessing/resstockpostproc/baseline_validation/tests/test_workflow_schema_histogram.py	Tests workflow schema path inference and download behavior for histogram fast path.
postprocessing/resstockpostproc/baseline_validation/tests/test_utils.py	Tests utility helpers (`add_us_total`, `add_missing_states`, `apply_aggregation`).
postprocessing/resstockpostproc/baseline_validation/tests/test_template_expansion.py	Tests template expansion and work-item generation behavior.
postprocessing/resstockpostproc/baseline_validation/tests/test_rse_column_resolution.py	Tests bound-column resolution logic for CI/RSE columns.
postprocessing/resstockpostproc/baseline_validation/tests/test_resstock_annual_fast_path.py	Tests raw-parquet “fast path” vs Athena fallback for ResStock annual data.
postprocessing/resstockpostproc/baseline_validation/tests/test_recs_rse.py	Tests RECS replicate-weight CI implementation (log-normal bounds) and batching.
postprocessing/resstockpostproc/baseline_validation/tests/test_range_utils.py	Tests shared axis range helper for raw/bounded data.
postprocessing/resstockpostproc/baseline_validation/tests/test_plot_spec.py	Tests PlotSpec validators and label/display formatting rules.
postprocessing/resstockpostproc/baseline_validation/tests/test_main.py	Tests CLI entry behavior and rejection of removed flags.
postprocessing/resstockpostproc/baseline_validation/tests/test_html_utils.py	Tests Plotly HTML post-processing including CDN/local fallback behavior.
postprocessing/resstockpostproc/baseline_validation/tests/test_get_recs_data_bounds.py	Tests RECS loader helpers for producing uncertainty bounds.
postprocessing/resstockpostproc/baseline_validation/tests/test_gather_data.py	Tests data routing for histogram pipeline and bounds column filtering.
postprocessing/resstockpostproc/baseline_validation/tests/test_footnotes.py	Tests generated plot/table footnotes and coverage notes.
postprocessing/resstockpostproc/baseline_validation/tests/test_error_bars.py	Tests error bar computation for bar plotter bounds.
postprocessing/resstockpostproc/baseline_validation/tests/test_data.py	Tests basic loading for EIA/LRD reference datasets.
postprocessing/resstockpostproc/baseline_validation/tests/test_create_html_metric_order.py	Tests metric ordering used in HTML viewer.
postprocessing/resstockpostproc/baseline_validation/tests/test_create_html_filter_pair_tabs.py	Tests HTML viewer Filter 1/2 tab and swap behavior.
postprocessing/resstockpostproc/baseline_validation/tests/test_create_html.py	Tests sharded HTML index generation and viz cell parsing.
postprocessing/resstockpostproc/baseline_validation/tests/_helpers.py	Adds shared PlotSpec builders for tests.
postprocessing/resstockpostproc/baseline_validation/schema/workflow_schema.py	Adds Pydantic workflow schema and ResStock parquet download helpers.
postprocessing/resstockpostproc/baseline_validation/schema/recs_chars_mapping.py	Adds RECS↔ResStock characteristic mapping for grouping/filtering.
postprocessing/resstockpostproc/baseline_validation/plotters/stacked_plotter.py	Adds stacked plot orchestrator (bar/box/histogram) with range logic and splitting.
postprocessing/resstockpostproc/baseline_validation/plotters/main_plotter.py	Adds main plot dispatch across tilemap/stacked/single-entity renderers.
postprocessing/resstockpostproc/baseline_validation/plotters/lrd_plotter.py	Adds LRD-specific plotting behaviors across multiple resolutions/views.
postprocessing/resstockpostproc/baseline_validation/plotters/box_plot_data.py	Adds quartile extraction + box plot dataframe preparation.
postprocessing/resstockpostproc/baseline_validation/plot_helpers/utils.py	Adds baseline validation helper utilities (month/season handling, BSQ creation).
postprocessing/resstockpostproc/baseline_validation/plot_helpers/theme.py	Adds baseline-validation-specific Plotly theming and template registration.
postprocessing/resstockpostproc/baseline_validation/plot_helpers/resstock_raw.py	Adds raw-parquet column resolution for schema variants and quantity mappings.
postprocessing/resstockpostproc/baseline_validation/plot_helpers/plot_semantics.py	Adds shared semantic helpers (units, timeseries col, source labels, quartile semantics).
postprocessing/resstockpostproc/baseline_validation/plot_helpers/footnotes.py	Adds note generation for plots/tables (RECS CI, RSE, penetration notes, histogram note).
postprocessing/resstockpostproc/baseline_validation/main.py	Adds baseline validation CLI entry point.
postprocessing/resstockpostproc/baseline_validation/io_managers/utils.py	Adds aggregation helpers and state/US-total normalization for loaded data.
postprocessing/resstockpostproc/baseline_validation/io_managers/stats.py	Adds weighted quantile helper used in annual aggregations.
postprocessing/resstockpostproc/baseline_validation/io_managers/output_manager.py	Adds figure/data saving logic incl. Plotly bundle vendoring and HTML postprocessing.
postprocessing/resstockpostproc/baseline_validation/io_managers/get_lrd_data.py	Adds LRD loader, caching, aggregation by resolution, and mapping to eiaid.
postprocessing/resstockpostproc/baseline_validation/io_managers/data_table_columns.py	Adds table column config builder and humanized labels for dashboard tables.
postprocessing/resstockpostproc/baseline_validation/io_managers/comparison_data_paths.py	Adds S3 paths to EIA/LRD/RECS truth datasets.
postprocessing/resstockpostproc/baseline_validation/generation/index_rows.py	Adds index-row assembly helpers (incl. LRD sidebar semantics normalization).
postprocessing/resstockpostproc/baseline_validation/generation/init.py	Adds generation package marker file.
postprocessing/resstockpostproc/baseline_validation/example_plots/recs_electricity_ustotal_difference.svg	Adds example SVG output artifact.
postprocessing/resstockpostproc/baseline_validation/example_plots/eia_electricity_perunit_difference_ustotal.svg	Adds example SVG output artifact.
postprocessing/resstockpostproc/baseline_validation/data_scraping/scrap_eia861M.py	Adds EIA 861M scraping script for monthly sales.
postprocessing/resstockpostproc/baseline_validation/data_scraping/scrap_eia861.py	Adds EIA 861 annual scraping script for sales/territories.
postprocessing/resstockpostproc/baseline_validation/data_scraping/scrap_eia176.py	Adds EIA 176 scraping/conversion script for gas consumption/heat content.
postprocessing/resstockpostproc/baseline_validation/data_processing/recs_rse.py	Adds RECS replicate-weight CI calculation helpers (single + batch).
postprocessing/resstockpostproc/baseline_validation/data_processing/recs_mapping.py	Adds mapping helpers for RECS columns/enduses/chars into internal schema.
postprocessing/resstockpostproc/baseline_validation/data_processing/metrics.py	Adds metric computations (e.g., per-source MAPE) over assembled plot data.
postprocessing/resstockpostproc/baseline_validation/data_processing/histogram_data.py	Adds exact histogram pipeline using raw ResStock parquet + RECS microdata.
postprocessing/resstockpostproc/baseline_validation/data_processing/dataset_adapters.py	Adds dataset adapters to pair reference + ResStock data and define join keys.
postprocessing/resstockpostproc/baseline_validation/dashboard/dashboard_paths.py	Adds path helpers for dashboard layout and dataset output directories.
postprocessing/pyproject.toml	Adds baseline-validation dependencies (excel, playwright, pyathena, etc.).
postprocessing/.typos.toml	Adds domain acronyms to typos allowlist and adjusts ignore patterns.
postprocessing/.pre-commit-config.yaml	Extends pre-commit scope to baseline_validation and updates exclusions.
postprocessing/.gitignore	Ignores baseline_validation scripts directory under postprocessing.
.gitignore	Adds `.history/`, `.cache/`, and `test_output/` ignores.

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

Copilot · 2026-04-22T01:26:10Z

+    @classmethod
+    def from_yaml(cls, yaml_path: str | Path) -> WorkflowConfig:
+        with open(yaml_path) as f:
+            config_dict = yaml.safe_load(f)
+        return cls.model_validate(config_dict)
+
+    def get_names_of_data_sources(self) -> list[str]:
+        return [ds.name for ds in self.data_sources]
+
+
+workflow_config_path = Path(__file__).parent.parent / "workflow.yaml"
+workflow = WorkflowConfig.from_yaml(workflow_config_path)


This module performs file I/O at import time by loading workflow.yaml and instantiating a global workflow. Import-time side effects make it hard to override config (e.g., via CLI --config), complicate tests, and can raise exceptions on import if the YAML is missing/invalid. Prefer moving the load into the CLI entry point (or a load_workflow(path) function) and keeping this module limited to schema definitions.

Copilot · 2026-04-22T01:26:10Z

+@lru_cache(maxsize=None)
+def get_df_from_s3(full_s3_path, cache_dir: Path | None = None) -> pl.DataFrame:
+    """Download (if needed) and read an S3 file as a Polars DataFrame.
+
+    Results are cached in-memory so repeated calls with the same arguments
+    skip the S3 HEAD check, local MD5 computation, and disk read entirely.
+    """
+    s3bucket, s3path = full_s3_path.replace("s3://", "").split("/", 1)
+    local_path = cache_dir / s3path
+    if not _is_file_same(s3bucket, s3path, local_path):


get_df_from_s3(..., cache_dir: Path | None = None) will raise a TypeError when cache_dir is left as None because local_path = cache_dir / s3path is unconditional. Either make cache_dir required, or provide a sensible default (and/or raise an actionable ValueError when it is not provided). Also consider validating that full_s3_path starts with s3:// (as download_s3_file does) to avoid confusing split errors.

Copilot · 2026-04-22T01:26:11Z

+from typing import Dict
+
+
+NUM2MONTH = {
+    1: "JAN",


from typing import Dict is unused and will trigger unused-import linting. Remove the import (or use it for the exported constants’ type annotations if needed).

Copilot · 2026-04-22T01:26:11Z

+def test_cached_decorator_creates_cache_under_postprocessing():
+    cache_name = "unit_test_cache_location"
+
+    @caching.cached(cache_name)
+    def cached_value(x: int) -> int:
+        return x + 1
+
+    assert cached_value(1) == 2
+    assert (caching.CACHE_ROOT / cache_name).is_dir()


This test writes cache artifacts into the repo-level .cache directory and doesn’t clean them up. That makes the test suite stateful across runs and can interfere with parallel testing. Prefer using tmp_path and monkeypatching caching.CACHE_ROOT (or providing a way to override it) so the test is isolated and leaves no persistent files behind.

Copilot · 2026-04-22T01:26:11Z

+def _is_file_same(bucket, s3_key, local_path):
+    if not local_path.exists():
+        return False
+    client = _get_s3_client()
+    local_md5 = _calculate_md5(local_path)
+    s3_metadata = client.head_object(Bucket=bucket, Key=s3_key)
+    s3_etag = s3_metadata["ETag"].strip('"')
+    return local_md5 == s3_etag


For multipart-uploaded S3 objects, the ETag is not a plain MD5 of the object content, so comparing a locally computed MD5 to head_object()['ETag'] can incorrectly report mismatches and trigger unnecessary re-downloads. Consider using ContentLength + LastModified, storing a known checksum in object metadata, or handling multipart ETags (the -<parts> suffix) explicitly.

Copilot · 2026-04-22T01:26:11Z

+df_heat_content_monthly = pd.read_excel(raw_path, sheet_name="Data 1", skiprows=2, skipfooter=1, na_values=["."])
+
+annul_heatcontent_file = "NG_CONS_HEAT_A_EPG0_VGTH_BTUCF_A.xls"
+raw_path = eia861_raw_path / monthly_heatcontent_file


Bug: the annual heat-content download writes into the monthly filename (raw_path = ... / monthly_heatcontent_file). This overwrites the monthly file and also means the annual file is never cached under its own name. Use annul_heatcontent_file (or rename the variable to annual_heatcontent_file) when constructing raw_path.

Suggested change

raw_path = eia861_raw_path / monthly_heatcontent_file

raw_path = eia861_raw_path / annul_heatcontent_file

Copilot · 2026-04-22T01:26:12Z

+from resstockpostproc.baseline_validation.generation.plot_generator import generate_plots
+from resstockpostproc.baseline_validation.schema.workflow_schema import workflow
+
+
+def main() -> int:
+    """Main entry point for ResStock comparison plot generation."""
+    parser = argparse.ArgumentParser(
+        description=(
+            "Generate comparison graphics and data between a ResStock baseline and other data sources (EIA, RECS, LRD)."
+        ),
+    )
+
+    parser.add_argument(
+        "--config",
+        type=str,
+        default=str(Path(__file__).parent / "workflow.yaml"),
+        help="Path to workflow configuration YAML file (default: workflow.yaml in script directory)",
+    )
+
+    args = parser.parse_args()
+
+    config_path = Path(args.config)
+    if not config_path.exists():
+        print(f"Error: Configuration file not found at {config_path}")
+        return 1
+
+    workflow.ensure_resstock_data_files()
+    generate_plots()


--config is validated for existence but never actually used to load a workflow; the CLI always uses the module-level workflow imported from schema.workflow_schema, and generate_plots() is invoked without the parsed config. This makes --config misleading and prevents running multiple configs without editing the default YAML. Load the config from config_path (e.g., WorkflowConfig.from_yaml) and either pass it into generate_plots(...)/downstream code or set the active workflow based on the CLI arg before calling ensure_resstock_data_files().

Copilot · 2026-04-22T01:26:12Z

+output:
+  output_dir: /Users/radhikar/Documents/buildstock2025/
+  run_name: baseline_val_test_2024_2025_final_test


The checked-in default config hard-codes a user-specific absolute output_dir ("/Users/..."), which will not exist for most users/CI and makes the example config non-portable. Consider using a relative path, a placeholder (e.g. ./test_output), or documenting that users must override this field (and keeping the committed value generic).

Copilot · 2026-04-22T01:26:12Z

+QUALITATIVE_SERIES = [  # from light series in https://sronpersonalpages.nl/~pault/
+    "#77AADD",  # light blude


Typo in comment: “light blude” should be “light blue”.

- main.py: drop --config flag; workflow.yaml next to main.py is the single source of config. README Running-the-Generator section and tests/test_main.py updated (the latter deleted; its removed-flags regression check isn't worth carrying for a first-release tool). - data_scraping/scrap_eia176.py: fix the annual-heat-content download writing into the monthly filename (Copilot #16). Also rename the 'annul_' variable to 'annual_'. - workflow.yaml: replace hardcoded /Users/radhikar/... output_dir with a portable ~/baseline_validation_outputs placeholder (Copilot #18). - shared_utils/colors.py: 'light blude' -> 'light blue' (Copilot #19). - shared_utils/mapping.py: drop unused 'from typing import Dict' (Copilot #13). - shared_utils/db_column_names.py: fix OEDI-schema column-name typo 'out.outdoor_air_dryblub_temp.c' -> 'drybulb' to match the enum name OUTDOOR_DRYBULB_TEMP and the sibling OEDI_NEW entry.

Verified against live Athena (workgroup=rescore): the column in buildstock_sdr.resstock_2024_amy2018_release_2_by_state_vu is literally named 'out.outdoor_air_dryblub_temp.c' (upstream typo). The previous commit corrected the spelling to match the enum name, which would have broken LRD plots against OEDI_VU tables — good catch from Rajendra. Added a comment flagging that the typo tracks an upstream schema choice so future reviewers (Copilot, humans) don't 'fix' it again.

- save_figure(formats=[FileType.html]) used a mutable list default evaluated once at function-definition time. Switch to the standard None-default + body-side fallback pattern. - Remove three unused helpers from plot_helpers/utils.py: filter_by_season, add_month_column, format_large_number, plus their supporting NUM2MONTH and SEASON2MONTHS dicts. All have zero external references. The local NUM2MONTH used full month names while the canonical shared_utils.mapping.NUM2MONTH used 3-letter abbreviations — the loaded gun is gone.

cache_dir was annotated as Optional with a None default but the body unconditionally used 'cache_dir / s3path', so calling with the default crashed with TypeError. Drop the Optional — every caller already passes a real Path, so make the contract honest. Also validate that full_s3_path starts with 's3://' before stripping; the previous .replace('s3://', '') silently no-op'd on non-s3 inputs and produced a confusing split error downstream. Mirrors what download_s3_file() already does.

- bar_plotter.py: drop dead xtitle/ytitle assignments (4 sites — lint F841/RUF059); dict() call -> {} literal (C408); add noqa to the second_category_title param that's accepted for API uniformity but never read (ARG001, same rationale as tilemap_plotter). - hover_formatting.py: int(math.floor(x)) -> math.floor(x) (RUF046 — math.floor already returns int). - histogram_utils.py: wrap a 127-char line (E501); dict comprehension -> dict.fromkeys (C420). - s3_manager.py: @lru_cache(maxsize=None) -> @functools.cache (UP033); hashlib.md5(usedforsecurity=False) since we're using it as an integrity checksum against S3 ETag, not for security (S324); noqa on the module-level _s3_client singleton global (PLW0603). - timing.py: mutable class attrs annotated with ClassVar (RUF012 — _stats and _pending_events); noqa on the long-lived trace-file open (SIM115 — closed by stop_trace). - tilemap_plotter.py: drop 3 unused signature params (orientation, label_formatter, categories — no callers pass them); noqa the remaining 2 (first_category_title, second_category_title — callers DO pass meaningful values but the body silently hardcodes empty strings; a known design debt, preserving current behavior). - caching.py: noqa the 'shelve file may not exist in read-only mode' try/except/pass (S110) — the comment already explains the intent.

…validation-tests

…dstock_query packaging)

…OOM hangs

…<br> which upload-artifact rejects)

…wrappers)

RECS now ships jackknife replicate weights; the Fay (1-epsilon)^2 denominator was the wrong variance estimator for them. Jackknife uses (R-1)/R * sum of squared log-residuals. CI ribbons widen ~3.84x across 854 RECS plots; point estimates unchanged.

rajeee added 9 commits November 5, 2025 11:52

Baseline validation

7f078fe

Some changes

020dde5

RECS monthly plot working

69f98d5

RECS annual working

7262986

RECS ships and boats plots

751b8ce

RECS plotter working

54fbdf7

In the works

3eeb548

More changes

81cf503

More recs changes

1a9fbc0

joseph-robertson assigned rajeee Dec 10, 2025

rajeee added 9 commits December 19, 2025 18:21

LRD plots and others

adb1c68

All chars working

b93c73d

more fixes

0233aaa

Further tweaks

7b15ca7

Further fixes

4f20b48

more fixes

aac4e45

More refactoring

acf5ca1

Fixes and simplification

8509af3

Resizable figures

b4745d3

rajeee requested a review from Copilot March 18, 2026 15:37

Copilot started reviewing on behalf of rajeee March 18, 2026 15:37 View session

This comment was marked as low quality.

Sign in to view

rajeee added 8 commits March 18, 2026 10:48

Merge branch 'develop' into baseline_validation

2b4855a

Editable HTML plots with SVG/PNG download buttons

24718bf

Human-readable plot titles and labels

29bb2e3

Add test mode for faster plot generation subset

8d9d983

remove dashboard

ae640e1

Ignore cache and history

7a994a9

Undo inadvertent change

bbb920b

GitHub Action and others added 3 commits April 21, 2026 23:00

SDR options analysis results.

c5a9907

rajeee requested a review from asparke2 April 22, 2026 00:03

rajeee requested a review from Copilot April 22, 2026 01:21

Copilot started reviewing on behalf of rajeee April 22, 2026 01:22 View session

Copilot AI reviewed Apr 22, 2026

View reviewed changes

rajeee added 19 commits April 21, 2026 20:34

end of line fix

1bdbc0e

Stabilize Plotly div_ids across output dir moves (hash rel_path)

d0d9c14

Default --test to false; wire --test through baseline_validation main.py

95801bd

Add baseline-validation-tests CI job: runs main.py --test, uploads plots

d3d0537

CI: install buildstock_query + fix artifact upload path for baseline-…

e16d006

…validation-tests

Bundle BSQ db_schema TOMLs locally, pass as dict (decouples from buil…

feec211

…dstock_query packaging)

Add --no-parallel flag to main.py, use it in CI to avoid worker pool …

30c2dfa

…OOM hangs

CI: tar baseline validation outputs before upload (filenames contain …

969dcf4

…<br> which upload-artifact rejects)

Inline _get_plot_data into get_base_data (drop redundant wrapper)

dc4a69a

Remove unused scale_to_eia_customers and its orphan AggregationBy alias

bc7b8d6

Inline _normalize_bounds and _partition_key_dict (single-caller thin …

ecb016d

…wrappers)

Drop 'refactor plan V2 step N.N' breadcrumbs from module docstrings

6f89061

Add changelog entry for baseline_validation tool (PR 1529)

93bf1e5

rajeee changed the title ~~Add Baseline validation tool to resstockpostproc~~ Add baseline validation tool to resstockpostproc Apr 22, 2026

rajeee added 2 commits April 24, 2026 17:43

Add 'uses AMY2018 versions' subtitle to comparison dashboard

df5cda7

Add Google Analytics tag to comparison dashboard

c70f8da

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Add baseline validation tool to resstockpostproc#1529

Add baseline validation tool to resstockpostproc#1529
rajeee wants to merge 226 commits intodevelopfrom
baseline_validation

rajeee commented Nov 5, 2025 •

edited

Loading

Uh oh!

This comment was marked as low quality.

Uh oh!

Copilot AI left a comment •

edited by rajeee

Loading

Uh oh!

Copilot AI Apr 22, 2026

Uh oh!

Copilot AI Apr 22, 2026

Uh oh!

Copilot AI Apr 22, 2026

Uh oh!

Copilot AI Apr 22, 2026

Uh oh!

Copilot AI Apr 22, 2026

Uh oh!

Copilot AI Apr 22, 2026

Uh oh!

Copilot AI Apr 22, 2026

Uh oh!

Copilot AI Apr 22, 2026

Uh oh!

Copilot AI Apr 22, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

	raw_path = eia861_raw_path / monthly_heatcontent_file
	raw_path = eia861_raw_path / annul_heatcontent_file

		QUALITATIVE_SERIES = [ # from light series in https://sronpersonalpages.nl/~pault/
		"#77AADD", # light blude

Conversation

rajeee commented Nov 5, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Pull Request Description

Validation test on CI

Checklist

Required:

Optional (not all items may apply):

Uh oh!

This comment was marked as low quality.

Uh oh!

Copilot AI left a comment • edited by rajeee Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Pull request overview

Reviewed changes

Uh oh!

Copilot AI Apr 22, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Apr 22, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Apr 22, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Apr 22, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Apr 22, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Apr 22, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Apr 22, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Apr 22, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Apr 22, 2026

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

rajeee commented Nov 5, 2025 •

edited

Loading

Copilot AI left a comment •

edited by rajeee

Loading