eduaguilera/whep-polities

Stars: 0
Forks: 0
Open beginner issues: 0
Indexed issues: 44
Avg merge: 1h 7m
Merged PRs (30d): 346

Dominant language: Python
License: No license data
Last GitHub push: Sep 14, 2026
Latest indexed: Sep 15, 2026
Contributing guide: No contributing guide
Code of conduct: No code of conduct
Beginner labels: No beginner labels indexed

Open All

44 open issues indexed

The cell over-claim that is NOT a shared polygon: 331,429 km2 across 188 coexisting pairs, and it needs a decision not a gate Open

eduaguilera/whep-polities#143 · 5 comments · 0 reactions · 0 assignees · Indexed Sep 15, 2026
Half the declared polygon areas are read off their own geometry, so the check comparing them is tautological Open

eduaguilera/whep-polities#195 · 19 comments · 0 reactions · 0 assignees · Indexed Sep 15, 2026
Three national rows span an enclave's absorption date, so no single subtraction is correct: one decision, three instances Open

eduaguilera/whep-polities#197 · 4 comments · 0 reactions · 0 assignees · Indexed Sep 15, 2026
Re-spanning orphans banked verdicts (15.7% of 466 assertion-level, all span drift) and the row figure depends on which record is the authority: 9,481 against a regenerated queue, 38,056 against the tracked one Open

eduaguilera/whep-polities#308 · 12 comments · 0 reactions · 0 assignees · Indexed Sep 15, 2026
9,210 valued rows with no year are assigned to year-bounded polities, half of them where several spans were possible Open

eduaguilera/whep-polities#310 · 11 comments · 0 reactions · 0 assignees · Indexed Sep 15, 2026
IIA's "indonesia" mixes the Dutch East Indies with a 7% subset of it: eight archipelago labels collapsed into one Open

eduaguilera/whep-polities#312 · 13 comments · 0 reactions · 0 assignees · Indexed Sep 15, 2026
25 IIA assertions carry a territory the label does not name: one executable reroute, eight decisions, and a Levant double count Open

eduaguilera/whep-polities#315 · 22 comments · 0 reactions · 0 assignees · Indexed Sep 15, 2026
Four polities receive both a label and a label it contains, and the composition registry cannot express it Open

eduaguilera/whep-polities#355 · 13 comments · 0 reactions · 0 assignees · Indexed Sep 15, 2026
iia wine: the 1.0659x factor is a documented hectolitre-density difference (0.097 vs the extract's 0.091), production-side only Open

eduaguilera/whep-polities#358 · 10 comments · 0 reactions · 0 assignees · Indexed Sep 15, 2026
1,619 valued rows repeat one value across years; for the iia half this is the late volumes' 100-unit grid, not a filled gap Open

eduaguilera/whep-polities#366 · 13 comments · 0 reactions · 0 assignees · Indexed Sep 15, 2026
3,081 fao1952 rows are indistinguishable within the panel, and in 239 groups a TOTAL sits beside its own parts — summing them double-counts exactly Open

eduaguilera/whep-polities#367 · 20 comments · 0 reactions · 0 assignees · Indexed Sep 15, 2026
11 iia labels mix territories at the ITEM level, so a label-level reroute cannot fix them — and #315's Ghana reading is wrong Open

eduaguilera/whep-polities#372 · 12 comments · 0 reactions · 0 assignees · Indexed Sep 15, 2026
iia "wheat" is spelt and meslin: the source has no wheat production rows at all — found because cross-source validation IS possible at polity level Open

eduaguilera/whep-polities#375 · 16 comments · 0 reactions · 0 assignees · Indexed Sep 15, 2026
Nine blockers are one decision: does this database admit polities for reporting units that were not sovereign states? Open

eduaguilera/whep-polities#400 · 22 comments · 0 reactions · 0 assignees · Indexed Sep 15, 2026
DEU-1920-1938 holds the 1937 Reich total AND its post-war subdivisions across 22 items, so the published figure is a median 74% of the Reich Open

eduaguilera/whep-polities#411 · 9 comments · 0 reactions · 0 assignees · Indexed Sep 15, 2026
IIA reads blank cells as 0 across all volumes, worst by far in iia_1938_39: 770 false zeros there, ~539 reaching layer B, 83 contradicted by another volume Open

eduaguilera/whep-polities#414 · 15 comments · 0 reactions · 0 assignees · Indexed Sep 15, 2026
IIA tobacco and hops are ~100x too large from 1934: 427 rows across 93 labels, 348 convicted per-row, and existing corrections reach 28 Open

eduaguilera/whep-polities#416 · 20 comments · 0 reactions · 0 assignees · Indexed Sep 15, 2026
IIA's Karafuto Prefecture (Japanese Sakhalin) is routed to Russia, and in 1918 it IS the entire Russian record Open

eduaguilera/whep-polities#422 · 8 comments · 0 reactions · 0 assignees · Indexed Sep 15, 2026
A x10 class at 1933 sits below every detector's floor: 14 layer-B cells are ten times too small, and the source prints the right value in another volume Open

eduaguilera/whep-polities#424 · 9 comments · 0 reactions · 0 assignees · Indexed Sep 15, 2026
The atomic-write lesson never propagated: 4 exposed write sites (not 18 tools), fixed in #439 and now gated by an AST sweep that finds 48 sites where a grep found 16 Open

eduaguilera/whep-polities#431 · 14 comments · 0 reactions · 0 assignees · Indexed Sep 15, 2026
Map: six IIA defect classes found on 2026-08-19 share two volumes and one era, and three of their remedies interact Open

eduaguilera/whep-polities#433 · 8 comments · 0 reactions · 0 assignees · Indexed Sep 15, 2026
The triage queue was stale because its generator CRASHED on null spans (fixed in #458); the resync itself is blocked on #310's period-year rule, and re-keys 291 of 1,074 assertions Open

eduaguilera/whep-polities#434 · 13 comments · 0 reactions · 0 assignees · Indexed Sep 15, 2026
`unattributable` hides a splice class: 8 iia series are two raw labels in sequence, and USA sugar publishes a different COMPONENT in each of four eras (never a national total) Open

eduaguilera/whep-polities#443 · 10 comments · 0 reactions · 0 assignees · Indexed Sep 15, 2026
The panel has no channel for value precision: 89,938 non-zero values (47.6%) sit on a 1000-grid and read as exact Open

eduaguilera/whep-polities#446 · 9 comments · 0 reactions · 0 assignees · Indexed Sep 15, 2026
FAO-1952's China whole and parts land on one polity: the six shared cells are AVERAGED, four publishing 52-65% of the national total, and two where the whole itself is wrong Open

eduaguilera/whep-polities#449 · 8 comments · 0 reactions · 0 assignees · Indexed Sep 15, 2026
1,995 collapse groups publish a blended value: 1,577 true duplicate keys and 418 item-code collisions, and only 3.0% are reachable by a label-pair screen Open

eduaguilera/whep-polities#451 · 15 comments · 0 reactions · 0 assignees · Indexed Sep 15, 2026
Layer-B iia `china, mainland` is 10.6% Kwantung (28 cells, 27 distinct values), spliced at 1933/34 — routing already flagged `redirected`, quantity and seam are new Open

eduaguilera/whep-polities#483 · 9 comments · 0 reactions · 0 assignees · Indexed Sep 15, 2026
Eight iia fertilizer series publish a MINOR material as the nutrient category: spain/p is phosphate rock at 5,400 t while the same page prints 999,607 t of superphosphate Open

eduaguilera/whep-polities#490 · 5 comments · 0 reactions · 0 assignees · Indexed Sep 15, 2026
The IIA extract has 93 corrupted country labels carrying 1,237 valued rows, and ~509 production/area cells reach no polity — 340 of them under one label, `[error] inde` Open

eduaguilera/whep-polities#493 · 9 comments · 0 reactions · 0 assignees · Indexed Sep 15, 2026
IIA's own area revisions are an unused independent test of our period boundaries: they agree in 11 of 20 cases, and 9 sit inside a single polity's span Open

eduaguilera/whep-polities#503 · 6 comments · 0 reactions · 0 assignees · Indexed Sep 15, 2026
Six small-territory polygons are 5-26% larger than their territories, and Liechtenstein rules out the obvious explanation Open

eduaguilera/whep-polities#570 · 3 comments · 0 reactions · 0 assignees · Indexed Sep 15, 2026
04_territory_basis.py --check FAILS after a normal pipeline run: the committed file and its check agree only while stage 02's output is absent Open

eduaguilera/whep-polities#573 · 4 comments · 0 reactions · 0 assignees · Indexed Sep 15, 2026
Nothing says which sovereign a dependent territory belongs to, so downstream reconstructs it two different accidental ways Open

data-quality decision-needed

eduaguilera/whep-polities#598 · 2 comments · 0 reactions · 0 assignees · Indexed Sep 15, 2026
Half the live rows state no area at all, and where they do `polygon_area_km2` means two different things: 33 of 223 disagree with their own polygon by over 5% Open

eduaguilera/whep-polities#600 · 4 comments · 0 reactions · 0 assignees · Indexed Sep 15, 2026
fao1952 gets ZERO cross-source validation because it reports in thousands and the tool keys on the unit string: normalising quadruples the tool's reach (804 → 3,165 cells) Open

eduaguilera/whep-polities#612 · 4 comments · 0 reactions · 0 assignees · Indexed Sep 15, 2026
Five of the ten files declared unregenerable have no gate reading their CONTENT — measured clean today, so this is a gap rather than a defect Open

eduaguilera/whep-polities#633 · 0 comments · 0 reactions · 0 assignees · Indexed Sep 15, 2026
An iia 1934-1938 average against the SAME series' 1928-1932 average is an unused screen: median 1.09 across 1,324 pairs, and 17 pairs break by 30x-925x in BOTH directions Open

eduaguilera/whep-polities#640 · 1 comment · 0 reactions · 0 assignees · Indexed Sep 15, 2026
IDN-1800-1889, IDN-1889-1945 and NNI-1904-1913 carry another country's cow_code, and nothing checks the column Open

eduaguilera/whep-polities#653 · 0 comments · 0 reactions · 0 assignees · Indexed Sep 15, 2026
19 rows hold one polygon across a span the reference says the border moved in, and GUF-1946-2025's moves by 6.4x Open

eduaguilera/whep-polities#654 · 0 comments · 0 reactions · 0 assignees · Indexed Sep 15, 2026
The US area jumps 6.2% in 1959 where no border moved: our era split uses the statehood convention, the polygons either side use two different ones Open

eduaguilera/whep-polities#655 · 0 comments · 0 reactions · 0 assignees · Indexed Sep 15, 2026
SWE-1814-1905 spans the union with Norway but carries Sweden proper's polygon: -41% against the reference, 821 layer-B rows Open

eduaguilera/whep-polities#656 · 0 comments · 0 reactions · 0 assignees · Indexed Sep 15, 2026
Modern-boundary units whose data predates them are recorded 'unroutable', not 'back_cast', so they read as unmatched data Open

eduaguilera/whep-polities#657 · 0 comments · 0 reactions · 0 assignees · Indexed Sep 15, 2026
Idaho, South Dakota, Washington and Oklahoma report data before statehood and the territory that reported it has no polity Open

eduaguilera/whep-polities#658 · 0 comments · 0 reactions · 0 assignees · Indexed Sep 15, 2026
build_database.py is not reproducible: 54 of 1,186 geometries move by up to 4.4 km2 across an identical rebuild Open

eduaguilera/whep-polities#660 · 0 comments · 0 reactions · 0 assignees · Indexed Sep 15, 2026

A short digest of beginner-friendly GitHub issues.