Source ingestion
Crawl gradually, preserve provenance, and fail closed.
Japanese akiya and property sale data is fragmented across older municipal and marketplace pages. The pipeline keeps crawl permissions, source URLs, Japanese labels, classification, and review status visible from the start.
Queue
Configured sources
Normalization
Japanese listing fields to preserve
| Japanese label | Normalized field | Buyer-facing use |
|---|---|---|
| 価格 / 売買価格 | priceJpy | Comparable purchase price and price filters |
| 所在地 / 住所 | address, prefecture, municipality | Map placement and local-service context |
| 土地面積 / 建物面積 | landArea, buildingArea | Renovation scope and total-cost estimate |
| 間取り | layout | Room-count comparison |
| 権利 / 所有権 | ownershipRights | Legal review priority |
| 備考 | sourceNotes | Unstructured caveats for human review |
| 物件種目 / タイトル | marketSegment, propertyClass | Akiya, resale, new construction, empty lot, condo, and commercial labels |