Clean addresses
Load a CSV of addresses as your customers typed them. Download it back with district, khoroo, building, door, floor, unit, the matched building and its coordinates on every row. The page calls POST /v1/clean 100 rows at a time; the key needs the clean:read scope. Example file: clean-example.csv.
The file and the key stay in this page. Addresses go only to the API; dock stores none. The first row is the header. Up to 10,000 rows, 300 characters per address. Every 100 rows count as one call.
Output columns
Every column of yours stays. The dock_ columns are added after them.
| Column | Value |
|---|---|
dock_status | cleaned, empty (no address), too_long (over 300 characters), not_sent (stopped or an error) |
dock_official_status, dock_unofficial | Matched to the state address register or not. For now every row is none, true |
dock_district, dock_district_code, dock_khoroo | District, its code, khoroo number |
dock_microdistrict ... dock_landmark | Parts as typed, several joined by | |
dock_entrance_no, dock_floor_no, dock_unit_no, dock_pavilion_no | As values, joined by ;. A basement is negative: "В1 давхар" is -1 |
dock_building_id, dock_lon, dock_lat | Matched building footprint and a point on it (WGS84) |
dock_method, dock_match_confidence, dock_source | How it was found, that method's share right on test rows, footprint source |
dock_withheld | Restricted site: building and point withheld |
dock_remainder | Text not placed in any field |
dock_choices | Ambiguous address only: up to 3 places it could be, answer first, joined by |. Each: district, nearest area and street, then lon lat. A restricted one: district and (withheld) |
The file must be UTF-8 (Excel: "CSV UTF-8"). Comma, semicolon and tab separated files are read. The output is comma separated with a UTF-8 BOM, so Excel shows the Cyrillic. Text starting with =, +, - or @ gets a leading ', so Excel never runs it as a formula.