"""Generate offline test doubles so the UI can be exercised without network. These files are TEST FIXTURES. The numbers are generated, not observed, and the dashboard never loads them unless CONSERVATION_HD_FIXTURES points here. """ from __future__ import annotations import json from pathlib import Path import numpy as np import pandas as pd import sys sys.path.insert(0, str(Path(__file__).resolve().parent)) import sources as src # noqa: E402 # Everything lives in one flat folder; the prefix keeps the test doubles obvious. OUT = Path(__file__).resolve().parent PREFIX = "fixture_" rng = np.random.default_rng(11) COUNTRIES = { "SEN": ("Senegal", 14.5, -14.5), "KEN": ("Kenya", 0.5, 37.9), "TZA": ("Tanzania", -6.4, 34.9), "ETH": ("Ethiopia", 9.1, 40.5), "ZAF": ("South Africa", -30.6, 22.9), "NGA": ("Nigeria", 9.1, 8.7), "ZMB": ("Zambia", -13.1, 27.8), "MOZ": ("Mozambique", -18.7, 35.5), "GHA": ("Ghana", 7.9, -1.0), "MAR": ("Morocco", 31.8, -7.1), "COD": ("Democratic Republic of the Congo", -4.0, 21.8), "BWA": ("Botswana", -22.3, 24.7), } IUCN = ["Ia", "II", "IV", "V", "VI", ""] # --- protected areas ------------------------------------------------------- rows = [] for iso3, (name, lat0, lon0) in COUNTRIES.items(): for i in range(rng.integers(12, 30)): rows.append({ "pa_id": f"Q{rng.integers(100000, 999999)}", "name": f"{name} Reserve {i + 1}", "iso3": iso3, "country": name, "lat": float(lat0 + rng.normal(0, 2.2)), "lon": float(lon0 + rng.normal(0, 2.2)), "area_km2": float(np.exp(rng.normal(6.0, 1.6))), "established": int(rng.integers(1930, 2020)), "iucn_category": str(rng.choice(IUCN)), "source": "fixture", }) pd.DataFrame(rows).to_parquet(OUT / (PREFIX + "protected_areas.parquet"), index=False) # --- world bank style indicator panel ------------------------------------- years = list(range(1990, 2025)) records = [] for iso3, (name, _, _) in COUNTRIES.items(): for code, meta in src.INDICATORS.items(): base = {"Education": 45, "Health": 58, "Income": 30, "Population": 60, "Conservation": 14}[meta["group"]] level = base * rng.uniform(0.5, 1.8) drift = np.linspace(1.0, rng.uniform(1.1, 1.9), len(years)) noise = 1 + rng.normal(0, 0.012, len(years)).cumsum() series = level * drift * noise if code == "SP.POP.TOTL": series = series * 5e5 for y, v in zip(years, series): if rng.random() < 0.08: # realistic gaps continue records.append({"iso3": iso3, "country": name, "indicator": code, "year": y, "value": float(v)}) pd.DataFrame(records).to_parquet(OUT / (PREFIX + "indicators_1990_2024.parquet"), index=False) # --- UN World Population Prospects style panel, with projections ----------- un = [] un_years = list(range(1950, 2101)) for iso3, (name, _, _) in COUNTRIES.items(): for code, meta in src.UN_INDICATORS.items(): base = {"UN:49": 2e7, "UN:51": 2.4, "UN:54": 70, "UN:67": 19, "UN:46": 3e6, "UN:41": 6e6, "UN:83": 72, "UN:84": 6, "UN:85": 15, "UN:86": 78}[code] level = base * rng.uniform(0.6, 1.5) drift = np.linspace(1.0, rng.uniform(0.7, 2.0), len(un_years)) series = level * drift * (1 + rng.normal(0, 0.008, len(un_years)).cumsum()) for y, v in zip(un_years, series): un.append({"iso3": iso3, "country": name, "indicator": code, "year": y, "value": float(abs(v)), "series": "Estimate" if y <= src.UN_LAST_ESTIMATE_YEAR else "Projection"}) pd.DataFrame(un).to_parquet(OUT / (PREFIX + "un_population.parquet"), index=False) # --- subnational hdi ------------------------------------------------------- sub = [] for iso3, (name, _, _) in COUNTRIES.items(): for r in range(4): for y in range(1995, 2023, 3): sub.append({"iso3": iso3, "country": name, "region": f"{name} region {r + 1}", "gdl_code": f"{iso3}r{r + 1}", "level": "Subnat", "year": y, "shdi": float(rng.uniform(0.3, 0.72)), "health_index": float(rng.uniform(0.3, 0.8)), "income_index": float(rng.uniform(0.25, 0.7)), "education_index": float(rng.uniform(0.2, 0.65))}) pd.DataFrame(sub).to_parquet(OUT / (PREFIX + "subnational_hdi.parquet"), index=False) # --- simple ADM0 polygons -------------------------------------------------- features = [] for iso3, (name, lat0, lon0) in COUNTRIES.items(): d = 3.0 features.append({ "type": "Feature", "properties": {"iso3": iso3, "name": name, "shapeName": name}, "geometry": {"type": "Polygon", "coordinates": [[ [lon0 - d, lat0 - d], [lon0 + d, lat0 - d], [lon0 + d, lat0 + d], [lon0 - d, lat0 + d], [lon0 - d, lat0 - d], ]]}, }) (OUT / (PREFIX + "boundaries_adm0.json")).write_text( json.dumps({"type": "FeatureCollection", "features": features})) # --- an effectiveness upload to exercise the join -------------------------- pas = pd.read_parquet(OUT / (PREFIX + "protected_areas.parquet")) sample = pas.sample(60, random_state=3) pd.DataFrame({ "name": sample["name"].values, "iso3": sample["iso3"].values, "score": np.round(rng.uniform(1.5, 4.8, len(sample)), 1), }).to_csv(OUT / (PREFIX + "effectiveness_upload.csv"), index=False) print("fixtures written to", OUT) for f in sorted(OUT.glob(PREFIX + "*")): print(" ", f.name, f.stat().st_size, "bytes")