Classic Automotive Ontology — Complete Reference

Domain: Automotive & Motorsport, 1950–2010
Source files: src/models.py, src/queries.py, seeds/dim5_technology.json, seeds/dim6_incidents.json, alembic/versions/001_initial_schema.py

---

1. Canonical Data Model

Every term in the ontology maps to this shape (OntologyTerm in models.py):

Field	Type	Notes
term	str	Human-readable name; whitespace-stripped. Unique together with dimension.
wikidata_qid	str \| None	Bare QID (e.g. Q786820). Full URIs normalised on input.
dimension	Dimension	One of 7 values (see §2).
sub_category	str \| None	Finer classification within a dimension.
era_start	int \| None	Year; validated 1885–2025.
era_end	int \| None	Year; validated 1885–2025.
relevance_score	int	1–5 (5 = core, 1 = marginal). Default 3 for SPARQL-sourced; manual seeds carry their own.
source	Source	wikidata, dbpedia, auto-ontology, or manual.
notes	str \| None	Free-text context. First 200 chars used in embedding text.
embedding	list[float] \| None	384-dim vector (all-MiniLM-L6-v2). Excluded from CSV export.

Embedding text formula (what gets vectorised):

term | sub_category | dimension | notes[:200]

Fields included only when non-null.

Unique constraint: (term, dimension)

---

2. Dimensions

The enum Dimension defines 7 values. CLAUDE.md groups EVENT + CIRCUIT as a combined "Dimension 3" for documentation purposes.

#	Dimension	Source	Sub-categories
1	MANUFACTURER	SPARQL	Automobile Manufacturer
2	MODEL	SPARQL (2 queries)	Road Car, Racing Car
3a	EVENT	SPARQL	Motorsport Championship
3b	CIRCUIT	SPARQL	Racing Circuit
4	PERSON	SPARQL (2 queries)	Racing Driver, Automotive Designer/Engineer
5	TECHNOLOGY	Manual seed only	See §4
6	INCIDENT	SPARQL + manual seed	See §5

---

3. SPARQL Query Coverage

All queries are paginated (LIMIT/OFFSET), use OPTIONAL for sparse Wikidata fields, and enforce a 55 s timeout per page.

3.1 Dimension 1 — MANUFACTURER

• Wikidata class: wd:Q786820 (automobile manufacturer), via wdt:P31/wdt:P279*
• Era filter: founded ≤ 2010, defunct ≥ 1950 (or unset)
• Fields pulled: label, country, founding year, defunct year

3.2 Dimension 2A — Road Car Models

• Wikidata class: wd:Q3231690 (automobile model)
• Era filter: start year ∈ [1950, 2010]
• Fields pulled: label, manufacturer, body style, start year, end year
• Relevance default: 3

3.3 Dimension 2B — Racing Car Models

• Wikidata classes (VALUES): Q223557, Q740308, Q1434981, Q3397737, Q2839680, Q1358119, Q2864737
• Covers: Formula cars, sports-racing prototypes, GT racing cars, etc.
• Era filter: start year ∈ [1948, 2012]
• Relevance default: 4

3.4 Dimension 3A — Motorsport Championships / Series

• Wikidata classes (VALUES): Q13219666, Q27020041, Q5389, Q187294, Q1968, Q181600, Q154469, Q190603, Q172678, Q230937, Q1025657, Q1199713
• Covers: racing series, World Championship, Formula One season, etc.
• Era filter: start year ≥ 1950
• Relevance default: 4

3.5 Dimension 3B — Circuits & Venues

• Wikidata classes (VALUES): Q1777138, Q10512426, Q1378645, Q1543884, Q4721
• Covers: racing circuit, motorsport venue, road course, street circuit, autodrome
• Era filter: opened ≤ era_end (no lower bound)
• Fields pulled: label, country, opened year

3.6 Dimension 4A — Racing Drivers

• Wikidata property: wdt:P106 wd:Q10841764 (occupation = racing driver)
• Era filter: born ≤ 1985, died ≥ 1950 (both optional)
• Fields pulled: label, nationality, birth year, death year

3.7 Dimension 4B — Automotive Designers / Engineers

• Wikidata occupations (VALUES): Q3501779, Q15296811, Q1326886, Q11063, Q81096
• Covers: automotive designer, car designer, mechanical engineer, engineer, scientist
• Era filter: born ∈ [1880, 1970]
• Fields pulled: label, occupation, nationality, birth year
• Relevance default: 4

3.8 Dimension 6 — Incidents (SPARQL portion)

• Wikidata class: wd:Q9687 (accident), via wdt:P31/wdt:P279*
• Era filter: event year ∈ [era_start, era_end] (1950–2010)
• Fields pulled: label, event year, location
• Relevance default: 4

---

4. Manual Seed — Dimension 5: Technology (40 terms)

All sourced as "source": "manual". Sub-categories and sample terms below.

Sub-category breakdown

Sub-category	Terms
Chassis/Suspension	Hydropneumatic suspension, Active suspension, Double wishbone suspension, MacPherson strut, De Dion rear axle, Torsion bar suspension
Chassis/Structure	Monocoque chassis, Spaceframe chassis, Unibody construction
Chassis/Materials	Carbon fibre monocoque
Powertrain/Layout	Mid-engine layout, Rear-engine layout
Braking	Disc brakes, Anti-lock braking system (ABS)
Fuel System	Fuel injection, Bosch K-Jetronic, Bosch L-Jetronic, Weber carburettor, Kugelfischer mechanical injection
Forced Induction	Turbocharger, Turbo lag, Intercooler (charge air cooler)
Engine/Valvetrain	Twin-cam engine (DOHC), VTEC (Variable Valve Timing)
Engine/Configuration	V12 engine, Flat-six engine (Boxer), Inline-six engine
Engine/Lubrication	Dry-sump lubrication
Transmission	Sequential gearbox
Drivetrain	Limited-slip differential, All-wheel drive (4WD) rally, Transaxle
Aerodynamics	Ground effect aerodynamics, Aerodynamic downforce
Materials	Kevlar bodywork
Regulation	Group B homologation rules, Group A regulations (WRC), FIA homologation special
Steering	Power steering (hydraulic)

Full term list with era and relevance

Term	era_start	era_end	Score	Wikidata QID
Hydropneumatic suspension	1955	2010	5	Q1542898
Monocoque chassis	1962	2010	5	—
Mid-engine layout	1960	2010	5	—
Disc brakes	1952	2010	5	Q337853
Fuel injection	1954	2010	5	—
Turbocharger	1962	2010	5	Q171173
V12 engine	1950	2010	5	—
Flat-six engine (Boxer)	1963	2010	5	—
Ground effect aerodynamics	1977	1983	5	—
Aerodynamic downforce	1968	2010	5	—
Carbon fibre monocoque	1981	2010	5	—
VTEC (Variable Valve Timing)	1989	2010	5	—
Active suspension	1987	1994	5	—
All-wheel drive (4WD) rally	1980	2010	5	—
Group B homologation rules	1982	1986	5	—
Bosch K-Jetronic	1973	1995	4	—
Bosch L-Jetronic	1967	1993	4	—
Twin-cam engine (DOHC)	1950	2010	4	—
Inline-six engine	1950	2010	4	—
Sequential gearbox	1989	2010	4	—
Limited-slip differential	1950	2010	4	—
Anti-lock braking system (ABS)	1978	2010	4	Q216225
Double wishbone suspension	1950	2010	4	—
MacPherson strut	1950	2010	4	Q1366622
Weber carburettor	1950	1990	4	—
Group A regulations (WRC)	1987	1996	4	—
FIA homologation special	1960	2010	4	—
Spaceframe chassis	1950	1970	4	—
Rear-engine layout	1950	1998	4	—
Unibody construction	1950	2010	4	—
De Dion rear axle	1950	1985	3	—
Kevlar bodywork	1976	2010	3	—
Dry-sump lubrication	1950	2010	3	—
Torsion bar suspension	1950	2000	3	—
Turbo lag	1976	2010	3	—
Intercooler (charge air cooler)	1977	2010	3	—
Power steering (hydraulic)	1956	2010	3	—
Transaxle	1950	2010	3	—
Kugelfischer mechanical injection	1960	1980	3	—
De Dion rear axle	1950	1985	3	—

---

5. Manual Seed — Dimension 6: Incidents (23 terms)

All sourced as "source": "manual". Sub-category taxonomy and full list:

Sub-category taxonomy

Sub-category	Meaning
Accident/Consequential	Crash with lasting rule or industry changes
Accident/Fatal	Fatal crash without major systemic changes
Accident/Technical	Fatal crash with technical/design root cause
Accident/Manufacturer Consequence	Crash that caused a manufacturer to withdraw
Accident/Regulatory Consequence	Same as Consequential, emphasising regulatory outcome
Regulatory Consequence	Rule change or ban triggered by an incident
Incident/Technical Disqualification	Car/team DSQ for technical infringement
Incident/Technical Protest	Post-race technical protest that changed the result
Incident/Regulatory	Finish or result dispute with a regulatory dimension
Incident/Championship Impact	On-track incident that shaped a championship battle
Incident/Disqualification	Broad DQ (e.g. lighting rules)

Full term list

Term	era	Score	QID
Le Mans 1955 disaster	1955	5	Q1060047
Mercedes motorsport withdrawal 1955	1955	5	—
1986 Portugal Rally crowd fatalities	1986	5	—
1986 Tour de Corse — Toivonen/Cresto fatal crash	1986	5	—
Group B ban (FIA 1986)	1986	5	—
1957 Mille Miglia — de Portago crash	1957	5	—
Mille Miglia cancellation 1957	1957	5	—
1976 German GP — Niki Lauda Nürburgring fire	1976	5	Q1058987
1982 Belgian GP — Gilles Villeneuve fatal crash	1982	5	Q918046
1994 San Marino GP — Senna and Ratzenberger	1994	5	Q924037
1999 Le Mans — Mercedes CLR triple flip	1999	5	—
1965 Monte Carlo Rally — Mini Cooper S exclusion	1965	5	—
Group B spectator safety revolution (1983-1986)	1983–1986	5	—
Nürburgring retirement from F1 (1977)	1976–1977	4	—
1994 Benetton B194 — launch control controversy	1994	4	—
1976 Japanese GP — McLaren brake adjuster protest	1976	4	—
1966 Le Mans — Ford GT40 photo-finish controversy	1966	4	—
1988 Spa — Senna vs Prost first lap collision	1988	4	—
1958 German GP — Peter Collins fatal crash	1958	4	—
1967 Monaco GP — Lorenzo Bandini fatal crash	1967	4	—
1968 French GP — Jo Schlesser fatal crash	1968	4	—
1977 South African GP — Tom Pryce fatal crash	1977	4	—
1995 Rally Catalunya — Ford Escort subframe dispute	1995	3	—

---

6. Database Schema

CREATE TABLE ontology_terms (
    id              SERIAL PRIMARY KEY,
    term            TEXT        NOT NULL,
    wikidata_qid    TEXT,
    dimension       TEXT        NOT NULL,     -- matches Dimension enum values
    sub_category    TEXT,
    era_start       INTEGER,
    era_end         INTEGER,
    relevance_score INTEGER     NOT NULL DEFAULT 3
                                CHECK (relevance_score BETWEEN 1 AND 5),
    source          TEXT        NOT NULL DEFAULT 'wikidata',
    notes           TEXT,
    embedding       vector(384),              -- all-MiniLM-L6-v2, cosine similarity
    created_at      TIMESTAMPTZ NOT NULL DEFAULT NOW(),
    updated_at      TIMESTAMPTZ NOT NULL DEFAULT NOW(),
    UNIQUE (term, dimension)
);

Indexes: - idx_ot_dimension on (dimension) — B-tree - idx_ot_era on (era_start, era_end) — B-tree - idx_ot_embedding using ivfflat (embedding vector_cosine_ops) with lists = 100

---

7. Sources & Coverage Summary

Dimension	Source	Expected volume
MANUFACTURER	SPARQL (Wikidata Q786820)	Hundreds
MODEL — Road Car	SPARQL (Q3231690, era 1950–2010)	Thousands
MODEL — Racing Car	SPARQL (7 car classes, era 1948–2012)	Hundreds
EVENT — Championships	SPARQL (12 series classes)	Hundreds
CIRCUIT	SPARQL (5 venue classes)	Hundreds
PERSON — Drivers	SPARQL (occupation Q10841764)	Thousands
PERSON — Designers	SPARQL (5 engineering occupations)	Hundreds
TECHNOLOGY	Manual seed (dim5_technology.json)	40 terms
INCIDENT	SPARQL + manual seed (dim6_incidents.json)	SPARQL: variable; manual: 23 terms

---

8. Notes on Irrelevant Files

The following files in the repo folder are not part of the automotive ontology and should be ignored or removed:

File	What it actually is
detailed_summary.md	Summary of a research document on Agentic AI in Financial Services
plan.md / plan 2.md	Plans for summarising the Agentic AI document above
user_pasted_clipboard_long_content_as_file_##_Expanded_validation_verdict_...txt	AI guru ranking validation report (completely unrelated)

These are artefacts of separate research sessions and have no bearing on the ontology build.