MOONSUGARmoonsugar.space

DVC · Project Apu

DVC · Project Apu · Build dossier

A pass over the project’s own repository: what is implemented, what is partial, what stayed on paper.

Not a web product: Python 3 on the standard library alone, Google Apps Script (V8) and Earth Engine scripts pasted in by hand. The output is KML/KMZ/GeoJSON/CSV files; there is no server of its own.

2 implemented · 2 partial · 1 on roadmap · 1 not implemented

Concept illustration: DVC · Project Apu. Not a product screenshot

What the code says

  • The pipeline is 12 Python 3 scripts (11 under tools/, 1 under analysis/), standard library only

    Source tools/build_sheet_imports.py

  • gas/Code.gs: 1,902 lines and 96 functions; doGet serves an HTML panel plus the kml, geojson, relationships, ai, networklink and health endpoints

    Source gas/Code.gs

  • Dataset: 74 curated sites, 173 candidates, 119 great circles, 6 hypotheses, 5 research tasks and 33 evidence sources, counted from CSV lines

    Source google_sheets_templates/Sites.csv

  • Precomputed relations: 3,575 circle memberships, 1,200 relationship edges, 1,200 ring intersections and 247 context rows

    Source google_sheets_templates/CircleMembership.csv

  • Static exports in the working tree: the KML holds 2,782 placemarks, 30 folders and 140 StyleMaps; the network GeoJSON has 382 features and the relationship one 2,415

    Source exports/dvc_global_network_google_earth.kml

  • Git: four commits on a clean main branch; data/, exports/ and release/ are excluded, so a fresh clone carries only code, docs and CSV templates

    Source .gitignore

Module state

  1. implemented

    The static export chain

    Raw KML plus two CSVs pass through sheet templates into KML/KMZ/GeoJSON; the generators and their finished outputs sit in the working tree.

    Source tools/build_static_exports.py

  2. implemented

    Separating facts from hypotheses

    Six hypotheses in CSV, each with an error metric, a confidence score, a status and a falsification test.

    Source google_sheets_templates/Hypotheses.csv

  3. partial

    The live Apps Script feed

    The source is complete (96 functions, seven formats) but undeployed: the NetworkLink href is a placeholder and no running instance exists in the repository.

    Source google_earth/dvc_networklink.kml

  4. partial

    External enrichment

    The Wikidata caches are populated; the Wikipedia nearby step returned an empty NearbyContext table, and the large cache is unversioned.

    Source tools/enrich_wikidata.py

  5. on roadmap

    DEM morphology analysis

    Earth Engine scripts must be pasted into the Google editor by hand; no raster output exists in the repository and hypothesis apu_h04 is marked as awaiting measurement.

    Source earth_engine/el_cono_morphology.js

  6. not implemented

    WGS84 recompute, null models, tests and CI

    The analysis script uses spherical formulas and no Monte Carlo script exists. No test files and no workflow under .github were found; the pipeline runs by hand.

    Source docs/project_apu_research_program.md

How this was checked

  • The count of 74 sites is verifiable from the file

    Sites.csv holds 75 lines including the header, and the README refers to all 74 sites.

    Source google_sheets_templates/Sites.csv

  • The 173 candidates are confirmed by the export

    CandidateAnomalies.csv has 174 lines, and the exported candidates GeoJSON contains exactly 173 Feature objects.

    Source exports/dvc_candidate_anomalies.geojson

  • The Giza offset from the 61-degree line is 0.045 km, on a spherical model

    The generated CSV records cross_track -0.045 km, a value produced by the spherical script rather than WGS84 geodesy.

    Source data/generated/apu_el_cono_site_metrics.csv

  • The generator really writes an interactive KML

    The exported KML carries 140 StyleMaps and 2,781 ExtendedData blocks, and the builder emits StyleMap nodes.

    Source tools/build_static_exports.py

Limits of these claims

  • There is no website or app of its own: viewing depends on Google Earth, Google Sheets and a hand-published Apps Script.
  • Live mode is not proven to work: no deployed address exists in the repository and the NetworkLink carries a placeholder.
  • Geodesic metrics use a spherical model anchored on one fixed El Cono coordinate; the WGS84 recompute is not implemented.
  • No tests, no CI and no dependency manifest; the pipeline scripts are run by hand in a fixed order.
  • The alternative historical claims stay hypotheses at candidate status: the code checks geometry, not history.
  • Enrichment reproducibility depends on the external Wikidata and Wikipedia APIs, and the large caches are unversioned.

Why there is no live build

Why there is no live build

There is nothing to publish: the project has no website of its own, and delivery runs through Google Earth and Google Sheets. Live mode is unproven, since no deployed address exists in the repository and the NetworkLink holds a placeholder.

Back to the project overview ↗