Aulvem/pet-airlines

Pet Air Travel Dataset (Japan Outbound)

A structured, machine-readable dataset of pet travel policies for major international airlines, focused on flights departing from Japan. Designed to be directly consumable by AI agents and downstream applications without scraping.

Overview

The web is full of human-readable pet travel guides; almost none are usable by an AI agent or a typed application without per-airline scraping. pet_airlines provides the missing structured layer: cabin / baggage / cargo policies, brachycephalic-breed restrictions, weight and carrier-size limits, required documents, fees, and reservation procedures — all under one schema, with English as the master and Japanese as a localized view.

Every record cites a single official source URL and carries an explicit last_checked_at date plus a confidence rating. Unknown values are marked unknown / null / [] rather than guessed.

What's in v0.1

Metric	Value
Records	10
Geographic focus	Japan outbound
Language coverage	English (master) + Japanese (view)
Confidence: high	6
Confidence: medium	4
Last checked	2026-04-27
Schema version	0.1
License	CC-BY 4.0

The 10 airlines covered:

1. Japan Airlines (JAL)
2. All Nippon Airways (ANA)
3. ZIPAIR
4. Korean Air
5. Asiana Airlines
6. Singapore Airlines
7. Cathay Pacific
8. EVA Air
9. Lufthansa
10. United Airlines

How to use

Python (datasets library)

from datasets import load_dataset

ds = load_dataset("Aulvem/pet-airlines")
print(ds["train"][0])
# {'id': 'pet_airline_001', 'airline_name_en': 'Japan Airlines', ...}

Direct JSON fetch (preserves nested structure)

The CSV is the default split, but the JSON file preserves the original nested objects (weight_limit.scope, brachycephalic_restriction.applies, etc.) and is the recommended source if your downstream code needs typed objects rather than flattened columns.

import json, urllib.request

url = "https://huggingface.co/datasets/Aulvem/pet-airlines/resolve/main/pet_airlines_v0.1.json"
data = json.load(urllib.request.urlopen(url))

for r in data["records"]:
    print(r["id"], r["airline_name_en"], r["confidence"])

Schema validation

import json, urllib.request, jsonschema  # pip install jsonschema

base = "https://huggingface.co/datasets/Aulvem/pet-airlines/resolve/main"
schema  = json.load(urllib.request.urlopen(f"{base}/schema.json"))
dataset = json.load(urllib.request.urlopen(f"{base}/pet_airlines_v0.1.json"))
jsonschema.validate(dataset, schema)   # raises on first failure

Schema

The full JSON Schema (Draft 2020-12) ships in this dataset as schema.json. Top-level fields per record:

Field	Type	Notes
id	string	Stable identifier, e.g. pet_airline_001
airline_name_en / airline_name_ja	string	English master + Japanese view
official_url	string (URI)	Single source page; required for every record
source_language	enum	en / ja / zh / ko / de / multi
pet_type	array	dog, cat, bird, rabbit, small_mammal, service_animal, emotional_support_animal, other
cabin_allowed / checked_baggage_allowed / cargo_allowed	enum	yes / no / conditional / unknown
brachycephalic_restriction	object	{ applies, details_en, details_ja }
weight_limit	object	{ value, unit (kg/lb), scope, notes_en, notes_ja }
carrier_size_limit	object	{ length_cm, width_cm, height_cm, sum_cm, notes_en, notes_ja }
reservation_method_en/ja	string	Free text
required_documents_en/ja	array of string	One document per element
fee_info_en/ja	string	Original currency preserved verbatim
notes_en/ja	string	Seasonal / route carve-outs (ISO date ranges)
last_checked_at	string (ISO date)	Verification date
confidence	enum	high / medium / low / unknown

The CSV at pet_airlines_v0.1.csv is a flattened 33-column view of the same data (nested objects expanded with prefixed column names; arrays joined with |). Both files are kept bit-identical with the canonical sources in the GitHub repository.

Important notice

Reference only. Always verify with official sources before travel.

Pet travel rules change frequently and are subject to per-route, per-aircraft, and per-season conditions, and may differ between an airline's published policy and the experience at the check-in counter. Confirm directly with the airline, the airport, and the animal-quarantine authority of every country on your route before booking. The dataset maintainers accept no liability for travel outcomes derived from this data; see the source repository's TERMS.md for the full disclaimer.

Update cadence

The dataset is automatically re-fetched and re-extracted weekly (Mondays 03:00 UTC) by a GitHub Actions workflow in the source repository. Each run produces a pull request with a structured diff against the prior snapshot. Updates are merged into the canonical dataset only after human review, and the merged result is mirrored here on Hugging Face.

You can pin to a specific revision using the standard HF dataset versioning, e.g. load_dataset("Aulvem/pet-airlines", revision="<commit-sha>").

Roadmap

Version	Focus
v0.1 (current)	10 airlines, Japan outbound, JSON + CSV
v0.2	Schema normalization (unit / currency harmonization), 25-30 airlines
v0.3	Country-level quarantine and import rules
v1.0	Public REST API with route / breed / season query semantics
v2.0	MCP server for direct AI-agent integration (Claude, ChatGPT, Cursor, Claude Code)

Schema changes are additive — v0.1-pinned consumers will not be broken by later versions.

Source repository

Canonical source, weekly extraction pipeline, schema definition, and contribution process all live on GitHub:

https://github.com/aulvem/pet_airlines

Issues, pull requests, and inaccuracy reports should go there. This Hugging Face dataset is a downstream mirror.

How to cite

If you use this dataset in research, products, or downstream tooling, please cite as:

pet_airlines contributors (2026). pet_airlines: AI-readable Pet Air Travel Dataset (v0.1). Available at https://huggingface.co/datasets/Aulvem/pet-airlines (mirror) and https://github.com/aulvem/pet_airlines (canonical), licensed under CC-BY 4.0.

CC-BY 4.0 requires attribution; the citation above — or any substantively equivalent form that names the dataset, version, and license — satisfies that requirement.

License

Released under Creative Commons Attribution 4.0 International (CC-BY 4.0). You are free to share and adapt, including for commercial use, provided you give appropriate credit.

The underlying airline policy text remains the property of each airline. This dataset extracts factual information (numerical limits, document lists, yes / no flags) under principles applicable to factual extraction; it does not redistribute the airlines' original prose. If you republish raw airline-authored text obtained via the source URLs, observe each airline's own terms.

Contact

Questions, bug reports, and corrections via GitHub issues:

https://github.com/aulvem/pet_airlines/issues

Please cite the official source page when reporting an inaccuracy.