keepset 0.18.0 → 0.20.0

package/dist/{bundled-BiuG_xLd.js → bundled-BmqpHjOH.js}

@@ -410,7 +410,7 @@ const CitationSchema = z.object({
 const RelativePathSchema = z.string().refine((path) => path.split("/").every((segment) => segment !== "" && segment !== "." && segment !== ".."), { message: "must be a relative path of non-empty segments with no '.' or '..'" });
 /** Which file of an archive to read: a member path, matched against a member's whole path or its trailing segments (so `net.gml` finds `net/net.gml`), or the member whose name ends with one of a list of extensions (a format can have several, such as Pajek's `.net` and `.paj`). Either must match exactly one member when the recipe runs; macOS resource-fork members (`__MACOSX/` and `._` files) never match. */
 const ArchiveMemberSchema = z.union([z.object({ path: RelativePathSchema }).strict(), z.object({ extensions: z.array(z.string().regex(/^\.[A-Za-z0-9][A-Za-z0-9.]*$/, { message: "must be a file extension such as .mtx" })).min(1) }).strict()]);
-/** How the downloaded bytes unpack into the one file the parser reads. */
+/** How the downloaded bytes unpack into the one file the parser reads. A workbook (`xlsx`) is read as the CSV of its named sheet, with dates as ISO 8601 text, for the `csv-table` parser. */
 const ArchiveSchema = z.discriminatedUnion("kind", [
 	z.object({ kind: z.literal("none") }).strict(),
 	z.object({ kind: z.literal("gzip") }).strict(),
@@ -425,6 +425,10 @@ const ArchiveSchema = z.discriminatedUnion("kind", [
 	z.object({
 		kind: z.literal("tar.gz"),
 		member: ArchiveMemberSchema
+	}).strict(),
+	z.object({
+		kind: z.literal("xlsx"),
+		sheet: z.string().min(1)
 	}).strict()
 ]);
 const ColumnSchema = z.int().min(0);
@@ -485,12 +489,41 @@ const ParserSchema = z.discriminatedUnion("format", [
 	}).refine((parser) => parser.columns === void 0 || new Set(parser.columns).size === parser.columns.length, {
 		message: "a column may be listed once",
 		path: ["columns"]
+	}),
+	z.object({
+		format: z.literal("geojson"),
+		encoding: z.literal("utf-8"),
+		idColumn: ColumnNameSchema.nullable(),
+		numericColumns: z.array(z.object({
+			column: ColumnNameSchema,
+			factor: z.number().positive()
+		}).strict()),
+		dateColumns: z.array(ColumnNameSchema),
+		columns: z.array(ColumnNameSchema).min(1)
+	}).strict().refine((parser) => new Set(parser.columns).size === parser.columns.length, {
+		message: "a column may be listed once",
+		path: ["columns"]
+	}).refine((parser) => parser.idColumn === null || !parser.columns.includes(parser.idColumn), {
+		message: "the id column is the row's id, so it is not listed among the columns",
+		path: ["columns"]
+	}).refine((parser) => parser.numericColumns.every((numeric) => parser.columns.includes(numeric.column)) && parser.dateColumns.every((date) => parser.columns.includes(date)), {
+		message: "a numeric or date column must be among the columns kept",
+		path: ["columns"]
+	}).refine((parser) => (/* @__PURE__ */ new Set([...parser.numericColumns.map((numeric) => numeric.column), ...parser.dateColumns])).size === parser.numericColumns.length + parser.dateColumns.length, {
+		message: "a column may be numeric or a date, not both, and once",
+		path: ["dateColumns"]
 	})
 ]);
+/** Where a source is a paged query and not one file: an ArcGIS feature layer, whose query endpoint returns a limited number of features per request. The source URL is the layer's, the download reads every page in `orderBy` order and the parser reads the one GeoJSON collection they make. */
+const PagingSchema = z.discriminatedUnion("kind", [z.object({
+	kind: z.literal("arcgis-query"),
+	orderBy: z.string().regex(/^[A-Za-z_][A-Za-z0-9_]*$/, { message: "must be a field name of the layer" })
+}).strict()]);
 /** How to obtain the dataset: a fetchable recipe naming the source (by index into `sources`), the archive layout and the parser, or a statement that no declarative recipe exists yet and why. */
 const RecipeSchema = z.discriminatedUnion("status", [z.object({
 	status: z.literal("fetchable"),
 	source: z.int().min(0),
+	paging: PagingSchema.optional(),
 	archive: ArchiveSchema,
 	parser: ParserSchema
 }).strict(), z.object({
@@ -860,7 +893,7 @@ var catalogue_default = {
 		"id": "keepset",
 		"name": "keepset",
 		"description": "The keepset catalogue: datasets with machine-readable recipes that fetch each one and normalise it into its category's target shape.",
-		"version": "1.6.0",
+		"version": "1.8.0",
 		"repository": "https://github.com/Mearman/keepset",
 		"namespaceAliases": {
 			"networkrepository": ["nr"],
@@ -1805,17 +1838,99 @@ var catalogue_default = {
 				"retrieved": "2026-10-19",
 				"status": "live"
 			}, {
-				"url": "https://utility.arcgis.com/usrsvcs/servers/53520af18235499d9ecd9c96f624eeae/rest/services/AM/Leak_Repairs_Live/FeatureServer/0/query?where=1%3D1&outFields=*&returnGeometry=false&f=json",
+				"url": "https://utility.arcgis.com/usrsvcs/servers/53520af18235499d9ecd9c96f624eeae/rest/services/AM/Leak_Repairs_Live/FeatureServer/0",
 				"retrieved": "2026-10-19",
 				"status": "live"
 			}],
 			"licence": { "kind": "unrecorded" },
 			"citations": [],
 			"recipe": {
-				"status": "not-fetchable",
-				"reason": "The source is an ArcGIS feature service that returns Esri JSON in pages of at most 2000 features, and keepset has no ArcGIS JSON parser or paging yet."
+				"status": "fetchable",
+				"source": 1,
+				"paging": {
+					"kind": "arcgis-query",
+					"orderBy": "OBJECTID"
+				},
+				"archive": { "kind": "none" },
+				"parser": {
+					"format": "geojson",
+					"encoding": "utf-8",
+					"idColumn": "OBJECTID",
+					"numericColumns": [{
+						"column": "geometry_latitude",
+						"factor": 1
+					}, {
+						"column": "geometry_longitude",
+						"factor": 1
+					}],
+					"dateColumns": [
+						"StartDateActual",
+						"StopDateActual",
+						"StopDateEstimated"
+					],
+					"columns": [
+						"StopDateActual",
+						"StopDateEstimated",
+						"geometry_latitude",
+						"geometry_longitude",
+						"StartDateActual",
+						"Activity",
+						"Status"
+					]
+				}
+			},
+			"crosswalk": {
+				"source": "geojson",
+				"target": "or/v1",
+				"rules": [
+					{
+						"op": "move",
+						"from": "meta",
+						"to": "meta"
+					},
+					{
+						"op": "move",
+						"from": "nodes[].id",
+						"to": "nodes[].id"
+					},
+					{
+						"op": "move",
+						"from": "nodes[].StopDateActual",
+						"to": "nodes[].closed"
+					},
+					{
+						"op": "move",
+						"from": "nodes[].StopDateEstimated",
+						"to": "nodes[].deadline"
+					},
+					{
+						"op": "move",
+						"from": "nodes[].geometry_latitude",
+						"to": "nodes[].lat"
+					},
+					{
+						"op": "move",
+						"from": "nodes[].geometry_longitude",
+						"to": "nodes[].lon"
+					},
+					{
+						"op": "move",
+						"from": "nodes[].StartDateActual",
+						"to": "nodes[].opened"
+					},
+					{
+						"op": "move",
+						"from": "nodes[].Activity",
+						"to": "nodes[].skill"
+					},
+					{
+						"op": "move",
+						"from": "nodes[].Status",
+						"to": "nodes[].status"
+					}
+				],
+				"oneWay": []
 			},
-			"crosswalk": { "file": "crosswalks/or-v1/nodes.json" },
 			"expectedCounts": {},
 			"properties": { "surveyId": "or-113" }
 		},
@@ -3481,18 +3596,74 @@ var catalogue_default = {
 				"url": "https://gis.ccc.govt.nz/server/rest/services/OpenData/WaterSupply/FeatureServer/3",
 				"retrieved": "2026-10-19",
 				"status": "live"
-			}, {
-				"url": "https://gis.ccc.govt.nz/server/rest/services/OpenData/WaterSupply/FeatureServer/3/query?where=1%3D1&outFields=*&returnGeometry=false&f=json",
-				"retrieved": "2026-10-19",
-				"status": "live"
 			}],
 			"licence": { "kind": "unrecorded" },
 			"citations": [],
 			"recipe": {
-				"status": "not-fetchable",
-				"reason": "The source is an ArcGIS feature service that returns Esri JSON in pages of at most 2000 features, and keepset has no ArcGIS JSON parser or paging yet."
+				"status": "fetchable",
+				"source": 0,
+				"paging": {
+					"kind": "arcgis-query",
+					"orderBy": "WaterShutOffID"
+				},
+				"archive": { "kind": "none" },
+				"parser": {
+					"format": "geojson",
+					"encoding": "utf-8",
+					"idColumn": "WaterShutOffID",
+					"numericColumns": [],
+					"dateColumns": ["PlannedFinishTime", "CreateDate"],
+					"columns": [
+						"PlannedFinishTime",
+						"CreateDate",
+						"IsPlanned",
+						"WorkStatus",
+						"LocationSuburb"
+					]
+				}
+			},
+			"crosswalk": {
+				"source": "geojson",
+				"target": "or/v1",
+				"rules": [
+					{
+						"op": "move",
+						"from": "meta",
+						"to": "meta"
+					},
+					{
+						"op": "move",
+						"from": "nodes[].id",
+						"to": "nodes[].id"
+					},
+					{
+						"op": "move",
+						"from": "nodes[].PlannedFinishTime",
+						"to": "nodes[].deadline"
+					},
+					{
+						"op": "move",
+						"from": "nodes[].CreateDate",
+						"to": "nodes[].opened"
+					},
+					{
+						"op": "move",
+						"from": "nodes[].IsPlanned",
+						"to": "nodes[].skill"
+					},
+					{
+						"op": "move",
+						"from": "nodes[].WorkStatus",
+						"to": "nodes[].status"
+					},
+					{
+						"op": "move",
+						"from": "nodes[].LocationSuburb",
+						"to": "nodes[].territory"
+					}
+				],
+				"oneWay": []
 			},
-			"crosswalk": { "file": "crosswalks/or-v1/nodes.json" },
 			"expectedCounts": {},
 			"properties": { "surveyId": "or-114" }
 		},
@@ -4583,6 +4754,10 @@ var catalogue_default = {
 				"url": "https://hub.arcgis.com/datasets/297fcd81118a4d5cbf2fd07ebdee628c",
 				"retrieved": "2026-10-19",
 				"status": "live"
+			}, {
+				"url": "https://gis2.co.dakota.mn.us/arcgis/rest/services/DCGIS_OL_PropertyInformation/MapServer/80",
+				"retrieved": "2026-10-19",
+				"status": "live"
 			}],
 			"licence": {
 				"kind": "custom",
@@ -4591,10 +4766,83 @@ var catalogue_default = {
 			},
 			"citations": [],
 			"recipe": {
-				"status": "not-fetchable",
-				"reason": "The source is an ArcGIS map service that returns GeoJSON and keepset has no GeoJSON parser yet."
+				"status": "fetchable",
+				"source": 1,
+				"paging": {
+					"kind": "arcgis-query",
+					"orderBy": "OBJECTID"
+				},
+				"archive": { "kind": "none" },
+				"parser": {
+					"format": "geojson",
+					"encoding": "utf-8",
+					"idColumn": "OBJECTID",
+					"numericColumns": [
+						{
+							"column": "geometry_latitude",
+							"factor": 1
+						},
+						{
+							"column": "geometry_longitude",
+							"factor": 1
+						},
+						{
+							"column": "SaleAmount",
+							"factor": 1
+						}
+					],
+					"dateColumns": ["SaleDate"],
+					"columns": [
+						"SaleDate",
+						"geometry_latitude",
+						"geometry_longitude",
+						"SaleAmount",
+						"CITYNAME"
+					]
+				}
+			},
+			"crosswalk": {
+				"source": "geojson",
+				"target": "or/v1",
+				"rules": [
+					{
+						"op": "move",
+						"from": "meta",
+						"to": "meta"
+					},
+					{
+						"op": "move",
+						"from": "nodes[].id",
+						"to": "nodes[].id"
+					},
+					{
+						"op": "move",
+						"from": "nodes[].SaleDate",
+						"to": "nodes[].deadline"
+					},
+					{
+						"op": "move",
+						"from": "nodes[].geometry_latitude",
+						"to": "nodes[].lat"
+					},
+					{
+						"op": "move",
+						"from": "nodes[].geometry_longitude",
+						"to": "nodes[].lon"
+					},
+					{
+						"op": "move",
+						"from": "nodes[].SaleAmount",
+						"to": "nodes[].prize"
+					},
+					{
+						"op": "move",
+						"from": "nodes[].CITYNAME",
+						"to": "nodes[].territory"
+					}
+				],
+				"oneWay": []
 			},
-			"crosswalk": { "file": "crosswalks/or-v1/nodes.json" },
 			"expectedCounts": {},
 			"properties": { "surveyId": "or-46" }
 		},
@@ -4947,6 +5195,10 @@ var catalogue_default = {
 				"url": "https://environment.data.gov.uk/dataset/9c7728fc-68e2-46e1-addb-63790b6b5313",
 				"retrieved": "2026-10-19",
 				"status": "live"
+			}, {
+				"url": "https://environment.data.gov.uk/api/file/download?fileDataSetId=e06036f5-7f41-46a3-93ff-fbd0fa1a83cc&fileName=Environment_Agency_Prosecutions.xlsx",
+				"retrieved": "2026-10-19",
+				"status": "live"
 			}],
 			"licence": {
 				"kind": "spdx",
@@ -4954,10 +5206,73 @@ var catalogue_default = {
 			},
 			"citations": [],
 			"recipe": {
-				"status": "not-fetchable",
-				"reason": "The download is an xlsx workbook with an explanation sheet and a data sheet, and keepset has no spreadsheet parser yet."
+				"status": "fetchable",
+				"source": 1,
+				"archive": {
+					"kind": "xlsx",
+					"sheet": "Environment Agency Prosecutions"
+				},
+				"parser": {
+					"format": "csv-table",
+					"encoding": "utf-8",
+					"delimiter": "comma",
+					"skipLines": 0,
+					"idColumn": null,
+					"numericColumns": [{
+						"column": "Fine",
+						"factor": 1
+					}],
+					"columns": [
+						"Date_of_Action",
+						"Fine",
+						"Incident_Type",
+						"Verdict",
+						"Incident_Area"
+					]
+				}
+			},
+			"crosswalk": {
+				"source": "csv-table",
+				"target": "or/v1",
+				"rules": [
+					{
+						"op": "move",
+						"from": "meta",
+						"to": "meta"
+					},
+					{
+						"op": "move",
+						"from": "nodes[].id",
+						"to": "nodes[].id"
+					},
+					{
+						"op": "move",
+						"from": "nodes[].Date_of_Action",
+						"to": "nodes[].closed"
+					},
+					{
+						"op": "move",
+						"from": "nodes[].Fine",
+						"to": "nodes[].prize"
+					},
+					{
+						"op": "move",
+						"from": "nodes[].Incident_Type",
+						"to": "nodes[].skill"
+					},
+					{
+						"op": "move",
+						"from": "nodes[].Verdict",
+						"to": "nodes[].status"
+					},
+					{
+						"op": "move",
+						"from": "nodes[].Incident_Area",
+						"to": "nodes[].territory"
+					}
+				],
+				"oneWay": []
 			},
-			"crosswalk": { "file": "crosswalks/or-v1/nodes.json" },
 			"expectedCounts": {},
 			"properties": { "surveyId": "or-50" }
 		},
@@ -5201,10 +5516,61 @@ var catalogue_default = {
 			"licence": { "kind": "unrecorded" },
 			"citations": [],
 			"recipe": {
-				"status": "not-fetchable",
-				"reason": "The source is an ArcGIS feature service that returns GeoJSON in pages of a limited number of features and keepset has no GeoJSON parser or paging yet."
+				"status": "fetchable",
+				"source": 0,
+				"paging": {
+					"kind": "arcgis-query",
+					"orderBy": "OBJECTID"
+				},
+				"archive": { "kind": "none" },
+				"parser": {
+					"format": "geojson",
+					"encoding": "utf-8",
+					"idColumn": "OBJECTID",
+					"numericColumns": [{
+						"column": "lien_balance",
+						"factor": 1
+					}],
+					"dateColumns": [],
+					"columns": [
+						"sale_or_purchase_date",
+						"lien_balance",
+						"lien_type"
+					]
+				}
+			},
+			"crosswalk": {
+				"source": "geojson",
+				"target": "or/v1",
+				"rules": [
+					{
+						"op": "move",
+						"from": "meta",
+						"to": "meta"
+					},
+					{
+						"op": "move",
+						"from": "nodes[].id",
+						"to": "nodes[].id"
+					},
+					{
+						"op": "move",
+						"from": "nodes[].sale_or_purchase_date",
+						"to": "nodes[].opened"
+					},
+					{
+						"op": "move",
+						"from": "nodes[].lien_balance",
+						"to": "nodes[].prize"
+					},
+					{
+						"op": "move",
+						"from": "nodes[].lien_type",
+						"to": "nodes[].skill"
+					}
+				],
+				"oneWay": []
 			},
-			"crosswalk": { "file": "crosswalks/or-v1/nodes.json" },
 			"expectedCounts": {},
 			"properties": { "surveyId": "or-44" }
 		},
@@ -5226,17 +5592,83 @@ var catalogue_default = {
 				"retrieved": "2026-10-19",
 				"status": "live"
 			}, {
-				"url": "https://services-eu1.arcgis.com/PyGgzM45TvWtSBsM/arcgis/rest/services/DCC_Derelict_Sites_Register_Public_points_view/FeatureServer/0/query?where=1%3D1&outFields=*&f=geojson",
+				"url": "https://services-eu1.arcgis.com/PyGgzM45TvWtSBsM/arcgis/rest/services/DCC_Derelict_Sites_Register_Public_points_view/FeatureServer/0",
 				"retrieved": "2026-10-19",
 				"status": "live"
 			}],
 			"licence": { "kind": "unrecorded" },
 			"citations": [],
 			"recipe": {
-				"status": "not-fetchable",
-				"reason": "The data is an ArcGIS feature service that returns GeoJSON, and keepset has no GeoJSON parser or paging yet."
+				"status": "fetchable",
+				"source": 1,
+				"paging": {
+					"kind": "arcgis-query",
+					"orderBy": "OBJECTID"
+				},
+				"archive": { "kind": "none" },
+				"parser": {
+					"format": "geojson",
+					"encoding": "utf-8",
+					"idColumn": "OBJECTID",
+					"numericColumns": [{
+						"column": "geometry_latitude",
+						"factor": 1
+					}, {
+						"column": "geometry_longitude",
+						"factor": 1
+					}],
+					"dateColumns": [],
+					"columns": [
+						"geometry_latitude",
+						"geometry_longitude",
+						"date_added_to_the_derelict_sites_register",
+						"derelict_site_status",
+						"administrative_area_name"
+					]
+				}
+			},
+			"crosswalk": {
+				"source": "geojson",
+				"target": "or/v1",
+				"rules": [
+					{
+						"op": "move",
+						"from": "meta",
+						"to": "meta"
+					},
+					{
+						"op": "move",
+						"from": "nodes[].id",
+						"to": "nodes[].id"
+					},
+					{
+						"op": "move",
+						"from": "nodes[].geometry_latitude",
+						"to": "nodes[].lat"
+					},
+					{
+						"op": "move",
+						"from": "nodes[].geometry_longitude",
+						"to": "nodes[].lon"
+					},
+					{
+						"op": "move",
+						"from": "nodes[].date_added_to_the_derelict_sites_register",
+						"to": "nodes[].opened"
+					},
+					{
+						"op": "move",
+						"from": "nodes[].derelict_site_status",
+						"to": "nodes[].status"
+					},
+					{
+						"op": "move",
+						"from": "nodes[].administrative_area_name",
+						"to": "nodes[].territory"
+					}
+				],
+				"oneWay": []
 			},
-			"crosswalk": { "file": "crosswalks/or-v1/nodes.json" },
 			"expectedCounts": {},
 			"properties": { "surveyId": "or-90" }
 		},
@@ -5625,14 +6057,66 @@ var catalogue_default = {
 				"url": "https://services1.arcgis.com/A14KNJpxNyBTu19J/arcgis/rest/services/Weed_Abatement_Parcels/FeatureServer",
 				"retrieved": "2026-10-19",
 				"status": "live"
+			}, {
+				"url": "https://services1.arcgis.com/A14KNJpxNyBTu19J/arcgis/rest/services/Weed_Abatement_Parcels/FeatureServer/0",
+				"retrieved": "2026-10-19",
+				"status": "live"
 			}],
 			"licence": { "kind": "unrecorded" },
 			"citations": [],
 			"recipe": {
-				"status": "not-fetchable",
-				"reason": "The data is an ArcGIS feature service that returns GeoJSON of parcel polygons in pages of at most 1000 features, and keepset has no GeoJSON parser or paging yet."
+				"status": "fetchable",
+				"source": 1,
+				"paging": {
+					"kind": "arcgis-query",
+					"orderBy": "OBJECTID"
+				},
+				"archive": { "kind": "none" },
+				"parser": {
+					"format": "geojson",
+					"encoding": "utf-8",
+					"idColumn": "OBJECTID",
+					"numericColumns": [],
+					"dateColumns": [],
+					"columns": [
+						"TYPE",
+						"STATUS",
+						"CITY"
+					]
+				}
+			},
+			"crosswalk": {
+				"source": "geojson",
+				"target": "or/v1",
+				"rules": [
+					{
+						"op": "move",
+						"from": "meta",
+						"to": "meta"
+					},
+					{
+						"op": "move",
+						"from": "nodes[].id",
+						"to": "nodes[].id"
+					},
+					{
+						"op": "move",
+						"from": "nodes[].TYPE",
+						"to": "nodes[].skill"
+					},
+					{
+						"op": "move",
+						"from": "nodes[].STATUS",
+						"to": "nodes[].status"
+					},
+					{
+						"op": "move",
+						"from": "nodes[].CITY",
+						"to": "nodes[].territory"
+					}
+				],
+				"oneWay": []
 			},
-			"crosswalk": { "file": "crosswalks/or-v1/nodes.json" },
 			"expectedCounts": {},
 			"properties": { "surveyId": "or-94" }
 		},
@@ -7195,14 +7679,72 @@ var catalogue_default = {
 				"url": "https://hub.arcgis.com/datasets/a963186c15d845bb93cce1c00022e380",
 				"retrieved": "2026-10-19",
 				"status": "live"
+			}, {
+				"url": "https://services3.arcgis.com/UYCjD74OfK5zL9b8/arcgis/rest/services/Arrears_2025/FeatureServer/0",
+				"retrieved": "2026-10-19",
+				"status": "live"
 			}],
 			"licence": { "kind": "unrecorded" },
 			"citations": [],
 			"recipe": {
-				"status": "not-fetchable",
-				"reason": "The source is an ArcGIS feature service that returns GeoJSON in pages of a limited number of features and keepset has no GeoJSON parser or paging yet."
+				"status": "fetchable",
+				"source": 1,
+				"paging": {
+					"kind": "arcgis-query",
+					"orderBy": "FID"
+				},
+				"archive": { "kind": "none" },
+				"parser": {
+					"format": "geojson",
+					"encoding": "utf-8",
+					"idColumn": "FID",
+					"numericColumns": [],
+					"dateColumns": [],
+					"columns": [
+						"ArrearsC_2",
+						"LeaseTypeD",
+						"ArrearsSta",
+						"RegionDesc"
+					]
+				}
+			},
+			"crosswalk": {
+				"source": "geojson",
+				"target": "or/v1",
+				"rules": [
+					{
+						"op": "move",
+						"from": "meta",
+						"to": "meta"
+					},
+					{
+						"op": "move",
+						"from": "nodes[].id",
+						"to": "nodes[].id"
+					},
+					{
+						"op": "move",
+						"from": "nodes[].ArrearsC_2",
+						"to": "nodes[].opened"
+					},
+					{
+						"op": "move",
+						"from": "nodes[].LeaseTypeD",
+						"to": "nodes[].skill"
+					},
+					{
+						"op": "move",
+						"from": "nodes[].ArrearsSta",
+						"to": "nodes[].status"
+					},
+					{
+						"op": "move",
+						"from": "nodes[].RegionDesc",
+						"to": "nodes[].territory"
+					}
+				],
+				"oneWay": []
 			},
-			"crosswalk": { "file": "crosswalks/or-v1/nodes.json" },
 			"expectedCounts": {},
 			"properties": { "surveyId": "or-41" }
 		},
@@ -8349,14 +8891,97 @@ var catalogue_default = {
 				"url": "https://hub.arcgis.com/datasets/1fd891c3301c4c4581b86c338468fbe4",
 				"retrieved": "2026-10-19",
 				"status": "live"
+			}, {
+				"url": "https://services1.arcgis.com/79kfd2K6fskCAkyg/arcgis/rest/services/PM_SiteVisit_Violations/FeatureServer/0",
+				"retrieved": "2026-10-19",
+				"status": "live"
 			}],
 			"licence": { "kind": "unrecorded" },
 			"citations": [],
 			"recipe": {
-				"status": "not-fetchable",
-				"reason": "The source is an ArcGIS feature service that returns GeoJSON in pages of a limited number of features, and keepset has no GeoJSON parser or paging yet."
+				"status": "fetchable",
+				"source": 1,
+				"paging": {
+					"kind": "arcgis-query",
+					"orderBy": "ObjectId"
+				},
+				"archive": { "kind": "none" },
+				"parser": {
+					"format": "geojson",
+					"encoding": "utf-8",
+					"idColumn": "ObjectId",
+					"numericColumns": [
+						{
+							"column": "geometry_latitude",
+							"factor": 1
+						},
+						{
+							"column": "geometry_longitude",
+							"factor": 1
+						},
+						{
+							"column": "CitationAmount",
+							"factor": 1
+						}
+					],
+					"dateColumns": ["G6A_G6_COMPL_DD"],
+					"columns": [
+						"G6A_G6_COMPL_DD",
+						"geometry_latitude",
+						"geometry_longitude",
+						"CitationAmount",
+						"G6A_G6_STATUS",
+						"CouncilDist"
+					]
+				}
+			},
+			"crosswalk": {
+				"source": "geojson",
+				"target": "or/v1",
+				"rules": [
+					{
+						"op": "move",
+						"from": "meta",
+						"to": "meta"
+					},
+					{
+						"op": "move",
+						"from": "nodes[].id",
+						"to": "nodes[].id"
+					},
+					{
+						"op": "move",
+						"from": "nodes[].G6A_G6_COMPL_DD",
+						"to": "nodes[].deadline"
+					},
+					{
+						"op": "move",
+						"from": "nodes[].geometry_latitude",
+						"to": "nodes[].lat"
+					},
+					{
+						"op": "move",
+						"from": "nodes[].geometry_longitude",
+						"to": "nodes[].lon"
+					},
+					{
+						"op": "move",
+						"from": "nodes[].CitationAmount",
+						"to": "nodes[].prize"
+					},
+					{
+						"op": "move",
+						"from": "nodes[].G6A_G6_STATUS",
+						"to": "nodes[].status"
+					},
+					{
+						"op": "move",
+						"from": "nodes[].CouncilDist",
+						"to": "nodes[].territory"
+					}
+				],
+				"oneWay": []
 			},
-			"crosswalk": { "file": "crosswalks/or-v1/nodes.json" },
 			"expectedCounts": {},
 			"properties": { "surveyId": "or-62" }
 		},
@@ -13445,6 +14070,10 @@ var catalogue_default = {
 				"url": "https://hub.arcgis.com/datasets/596fa66b53a948e4b74fea86dbe44690",
 				"retrieved": "2026-10-19",
 				"status": "live"
+			}, {
+				"url": "https://www.maps.north-ayrshire.gov.uk/arcgis/rest/services/AGOL/Building_Warrant_Applications_Weekly/FeatureServer/2",
+				"retrieved": "2026-10-19",
+				"status": "live"
 			}],
 			"licence": {
 				"kind": "custom",
@@ -13453,10 +14082,82 @@ var catalogue_default = {
 			},
 			"citations": [],
 			"recipe": {
-				"status": "not-fetchable",
-				"reason": "The data is an ArcGIS feature service that returns GeoJSON in pages of at most 2000 features, and keepset has no GeoJSON parser or paging yet."
+				"status": "fetchable",
+				"source": 1,
+				"paging": {
+					"kind": "arcgis-query",
+					"orderBy": "OBJECTID"
+				},
+				"archive": { "kind": "none" },
+				"parser": {
+					"format": "geojson",
+					"encoding": "utf-8",
+					"idColumn": "OBJECTID",
+					"numericColumns": [{
+						"column": "geometry_latitude",
+						"factor": 1
+					}, {
+						"column": "geometry_longitude",
+						"factor": 1
+					}],
+					"dateColumns": ["RECPTD", "DECIDD"],
+					"columns": [
+						"DECIDD",
+						"geometry_latitude",
+						"geometry_longitude",
+						"RECPTD",
+						"APPTYP",
+						"DECISION"
+					]
+				}
+			},
+			"crosswalk": {
+				"source": "geojson",
+				"target": "or/v1",
+				"rules": [
+					{
+						"op": "move",
+						"from": "meta",
+						"to": "meta"
+					},
+					{
+						"op": "move",
+						"from": "nodes[].id",
+						"to": "nodes[].id"
+					},
+					{
+						"op": "move",
+						"from": "nodes[].DECIDD",
+						"to": "nodes[].closed"
+					},
+					{
+						"op": "move",
+						"from": "nodes[].geometry_latitude",
+						"to": "nodes[].lat"
+					},
+					{
+						"op": "move",
+						"from": "nodes[].geometry_longitude",
+						"to": "nodes[].lon"
+					},
+					{
+						"op": "move",
+						"from": "nodes[].RECPTD",
+						"to": "nodes[].opened"
+					},
+					{
+						"op": "move",
+						"from": "nodes[].APPTYP",
+						"to": "nodes[].skill"
+					},
+					{
+						"op": "move",
+						"from": "nodes[].DECISION",
+						"to": "nodes[].status"
+					}
+				],
+				"oneWay": []
 			},
-			"crosswalk": { "file": "crosswalks/or-v1/nodes.json" },
 			"expectedCounts": {},
 			"properties": { "surveyId": "or-78" }
 		},
@@ -15201,14 +15902,84 @@ var catalogue_default = {
 				"url": "https://utility.arcgis.com/usrsvcs/servers/931a175b8eaa4abebf1379743e6d71c1/rest/services/Environmental_Resources/Weed_Abatement/FeatureServer",
 				"retrieved": "2026-10-19",
 				"status": "live"
+			}, {
+				"url": "https://utility.arcgis.com/usrsvcs/servers/931a175b8eaa4abebf1379743e6d71c1/rest/services/Environmental_Resources/Weed_Abatement/FeatureServer/0",
+				"retrieved": "2026-10-19",
+				"status": "live"
 			}],
 			"licence": { "kind": "unrecorded" },
 			"citations": [],
 			"recipe": {
-				"status": "not-fetchable",
-				"reason": "The data is an ArcGIS feature service that returns GeoJSON in pages of at most 15000 features, and keepset has no GeoJSON parser or paging yet."
+				"status": "fetchable",
+				"source": 1,
+				"paging": {
+					"kind": "arcgis-query",
+					"orderBy": "OBJECTID"
+				},
+				"archive": { "kind": "none" },
+				"parser": {
+					"format": "geojson",
+					"encoding": "utf-8",
+					"idColumn": "OBJECTID",
+					"numericColumns": [{
+						"column": "geometry_latitude",
+						"factor": 1
+					}, {
+						"column": "geometry_longitude",
+						"factor": 1
+					}],
+					"dateColumns": ["created_date"],
+					"columns": [
+						"geometry_latitude",
+						"geometry_longitude",
+						"created_date",
+						"status",
+						"district"
+					]
+				}
+			},
+			"crosswalk": {
+				"source": "geojson",
+				"target": "or/v1",
+				"rules": [
+					{
+						"op": "move",
+						"from": "meta",
+						"to": "meta"
+					},
+					{
+						"op": "move",
+						"from": "nodes[].id",
+						"to": "nodes[].id"
+					},
+					{
+						"op": "move",
+						"from": "nodes[].geometry_latitude",
+						"to": "nodes[].lat"
+					},
+					{
+						"op": "move",
+						"from": "nodes[].geometry_longitude",
+						"to": "nodes[].lon"
+					},
+					{
+						"op": "move",
+						"from": "nodes[].created_date",
+						"to": "nodes[].opened"
+					},
+					{
+						"op": "move",
+						"from": "nodes[].status",
+						"to": "nodes[].status"
+					},
+					{
+						"op": "move",
+						"from": "nodes[].district",
+						"to": "nodes[].territory"
+					}
+				],
+				"oneWay": []
 			},
-			"crosswalk": { "file": "crosswalks/or-v1/nodes.json" },
 			"expectedCounts": {},
 			"properties": { "surveyId": "or-92" }
 		},
@@ -18798,14 +19569,84 @@ var catalogue_default = {
 				"url": "https://services3.arcgis.com/vtilYbC3YeF6EBJ1/arcgis/rest/services/service_05f1e63ba1494703b42f0dd606859f0b/FeatureServer",
 				"retrieved": "2026-10-19",
 				"status": "live"
+			}, {
+				"url": "https://services3.arcgis.com/vtilYbC3YeF6EBJ1/arcgis/rest/services/service_05f1e63ba1494703b42f0dd606859f0b/FeatureServer/0",
+				"retrieved": "2026-10-19",
+				"status": "live"
 			}],
 			"licence": { "kind": "unrecorded" },
 			"citations": [],
 			"recipe": {
-				"status": "not-fetchable",
-				"reason": "The data is an ArcGIS feature service that returns GeoJSON in pages of at most 1000 features, and keepset has no GeoJSON parser or paging yet."
+				"status": "fetchable",
+				"source": 1,
+				"paging": {
+					"kind": "arcgis-query",
+					"orderBy": "objectid"
+				},
+				"archive": { "kind": "none" },
+				"parser": {
+					"format": "geojson",
+					"encoding": "utf-8",
+					"idColumn": "objectid",
+					"numericColumns": [{
+						"column": "geometry_latitude",
+						"factor": 1
+					}, {
+						"column": "geometry_longitude",
+						"factor": 1
+					}],
+					"dateColumns": ["date_towed"],
+					"columns": [
+						"geometry_latitude",
+						"geometry_longitude",
+						"date_towed",
+						"reason",
+						"vehicle_city"
+					]
+				}
+			},
+			"crosswalk": {
+				"source": "geojson",
+				"target": "or/v1",
+				"rules": [
+					{
+						"op": "move",
+						"from": "meta",
+						"to": "meta"
+					},
+					{
+						"op": "move",
+						"from": "nodes[].id",
+						"to": "nodes[].id"
+					},
+					{
+						"op": "move",
+						"from": "nodes[].geometry_latitude",
+						"to": "nodes[].lat"
+					},
+					{
+						"op": "move",
+						"from": "nodes[].geometry_longitude",
+						"to": "nodes[].lon"
+					},
+					{
+						"op": "move",
+						"from": "nodes[].date_towed",
+						"to": "nodes[].opened"
+					},
+					{
+						"op": "move",
+						"from": "nodes[].reason",
+						"to": "nodes[].skill"
+					},
+					{
+						"op": "move",
+						"from": "nodes[].vehicle_city",
+						"to": "nodes[].territory"
+					}
+				],
+				"oneWay": []
 			},
-			"crosswalk": { "file": "crosswalks/or-v1/nodes.json" },
 			"expectedCounts": {},
 			"properties": { "surveyId": "or-96" }
 		},
@@ -18938,14 +19779,91 @@ var catalogue_default = {
 				"url": "https://hub.arcgis.com/datasets/b8ac65c653b343739be21cb2773faace",
 				"retrieved": "2026-10-19",
 				"status": "live"
+			}, {
+				"url": "https://services.arcgis.com/fLeGjb7u4uXqeF9q/arcgis/rest/services/REAL_ESTATE_TAX_DELINQUENCIES/FeatureServer/0",
+				"retrieved": "2026-10-19",
+				"status": "live"
 			}],
 			"licence": { "kind": "unrecorded" },
 			"citations": [],
 			"recipe": {
-				"status": "not-fetchable",
-				"reason": "The source is an ArcGIS feature service that returns GeoJSON in pages of a limited number of features and keepset has no GeoJSON parser or paging yet."
+				"status": "fetchable",
+				"source": 1,
+				"paging": {
+					"kind": "arcgis-query",
+					"orderBy": "OBJECTID"
+				},
+				"archive": { "kind": "none" },
+				"parser": {
+					"format": "geojson",
+					"encoding": "utf-8",
+					"idColumn": "OBJECTID",
+					"numericColumns": [
+						{
+							"column": "geometry_latitude",
+							"factor": 1
+						},
+						{
+							"column": "geometry_longitude",
+							"factor": 1
+						},
+						{
+							"column": "TOTAL_DUE",
+							"factor": 1
+						}
+					],
+					"dateColumns": [],
+					"columns": [
+						"geometry_latitude",
+						"geometry_longitude",
+						"TOTAL_DUE",
+						"BUILDING_CATEGORY",
+						"ZIP_CODE"
+					]
+				}
+			},
+			"crosswalk": {
+				"source": "geojson",
+				"target": "or/v1",
+				"rules": [
+					{
+						"op": "move",
+						"from": "meta",
+						"to": "meta"
+					},
+					{
+						"op": "move",
+						"from": "nodes[].id",
+						"to": "nodes[].id"
+					},
+					{
+						"op": "move",
+						"from": "nodes[].geometry_latitude",
+						"to": "nodes[].lat"
+					},
+					{
+						"op": "move",
+						"from": "nodes[].geometry_longitude",
+						"to": "nodes[].lon"
+					},
+					{
+						"op": "move",
+						"from": "nodes[].TOTAL_DUE",
+						"to": "nodes[].prize"
+					},
+					{
+						"op": "move",
+						"from": "nodes[].BUILDING_CATEGORY",
+						"to": "nodes[].skill"
+					},
+					{
+						"op": "move",
+						"from": "nodes[].ZIP_CODE",
+						"to": "nodes[].territory"
+					}
+				],
+				"oneWay": []
 			},
-			"crosswalk": { "file": "crosswalks/or-v1/nodes.json" },
 			"expectedCounts": {},
 			"properties": { "surveyId": "or-40" }
 		},
@@ -52253,14 +53171,78 @@ var catalogue_default = {
 				"url": "https://services3.arcgis.com/IgR6ay8jzuIj7VSF/arcgis/rest/services/Weed_Abatement_2017/FeatureServer",
 				"retrieved": "2026-10-19",
 				"status": "live"
+			}, {
+				"url": "https://services3.arcgis.com/IgR6ay8jzuIj7VSF/arcgis/rest/services/Weed_Abatement_2017/FeatureServer/0",
+				"retrieved": "2026-10-19",
+				"status": "live"
 			}],
 			"licence": { "kind": "unrecorded" },
 			"citations": [],
 			"recipe": {
-				"status": "not-fetchable",
-				"reason": "The data is an ArcGIS feature service that returns GeoJSON in pages of at most 1000 features, and keepset has no GeoJSON parser or paging yet."
+				"status": "fetchable",
+				"source": 1,
+				"paging": {
+					"kind": "arcgis-query",
+					"orderBy": "OBJECTID"
+				},
+				"archive": { "kind": "none" },
+				"parser": {
+					"format": "geojson",
+					"encoding": "utf-8",
+					"idColumn": "OBJECTID",
+					"numericColumns": [{
+						"column": "geometry_latitude",
+						"factor": 1
+					}, {
+						"column": "geometry_longitude",
+						"factor": 1
+					}],
+					"dateColumns": ["CreationDate"],
+					"columns": [
+						"geometry_latitude",
+						"geometry_longitude",
+						"CreationDate",
+						"ABATE"
+					]
+				}
+			},
+			"crosswalk": {
+				"source": "geojson",
+				"target": "or/v1",
+				"rules": [
+					{
+						"op": "move",
+						"from": "meta",
+						"to": "meta"
+					},
+					{
+						"op": "move",
+						"from": "nodes[].id",
+						"to": "nodes[].id"
+					},
+					{
+						"op": "move",
+						"from": "nodes[].geometry_latitude",
+						"to": "nodes[].lat"
+					},
+					{
+						"op": "move",
+						"from": "nodes[].geometry_longitude",
+						"to": "nodes[].lon"
+					},
+					{
+						"op": "move",
+						"from": "nodes[].CreationDate",
+						"to": "nodes[].opened"
+					},
+					{
+						"op": "move",
+						"from": "nodes[].ABATE",
+						"to": "nodes[].status"
+					}
+				],
+				"oneWay": []
 			},
-			"crosswalk": { "file": "crosswalks/or-v1/nodes.json" },
 			"expectedCounts": {},
 			"properties": { "surveyId": "or-93" }
 		},
@@ -52671,17 +53653,77 @@ var catalogue_default = {
 				"retrieved": "2026-10-19",
 				"status": "live"
 			}, {
-				"url": "https://services1.arcgis.com/eGSDp8lpKe5izqVc/arcgis/rest/services/2024_Sheriff_Evictions_WFL1/FeatureServer/0/query?where=1%3D1&outFields=*&f=json",
+				"url": "https://services1.arcgis.com/eGSDp8lpKe5izqVc/arcgis/rest/services/2024_Sheriff_Evictions_WFL1/FeatureServer/0",
 				"retrieved": "2026-10-19",
 				"status": "live"
 			}],
 			"licence": { "kind": "unrecorded" },
 			"citations": [],
 			"recipe": {
-				"status": "not-fetchable",
-				"reason": "The source is an ArcGIS feature service that returns Esri JSON in pages of at most 2000 features, and keepset has no ArcGIS JSON parser or paging yet."
+				"status": "fetchable",
+				"source": 1,
+				"paging": {
+					"kind": "arcgis-query",
+					"orderBy": "OBJECTID"
+				},
+				"archive": { "kind": "none" },
+				"parser": {
+					"format": "geojson",
+					"encoding": "utf-8",
+					"idColumn": "OBJECTID",
+					"numericColumns": [{
+						"column": "geometry_latitude",
+						"factor": 1
+					}, {
+						"column": "geometry_longitude",
+						"factor": 1
+					}],
+					"dateColumns": [],
+					"columns": [
+						"geometry_latitude",
+						"geometry_longitude",
+						"Date",
+						"Status_1"
+					]
+				}
+			},
+			"crosswalk": {
+				"source": "geojson",
+				"target": "or/v1",
+				"rules": [
+					{
+						"op": "move",
+						"from": "meta",
+						"to": "meta"
+					},
+					{
+						"op": "move",
+						"from": "nodes[].id",
+						"to": "nodes[].id"
+					},
+					{
+						"op": "move",
+						"from": "nodes[].geometry_latitude",
+						"to": "nodes[].lat"
+					},
+					{
+						"op": "move",
+						"from": "nodes[].geometry_longitude",
+						"to": "nodes[].lon"
+					},
+					{
+						"op": "move",
+						"from": "nodes[].Date",
+						"to": "nodes[].opened"
+					},
+					{
+						"op": "move",
+						"from": "nodes[].Status_1",
+						"to": "nodes[].status"
+					}
+				],
+				"oneWay": []
 			},
-			"crosswalk": { "file": "crosswalks/or-v1/nodes.json" },
 			"expectedCounts": {},
 			"properties": { "surveyId": "or-103" }
 		},
@@ -53968,6 +55010,10 @@ var catalogue_default = {
 				"url": "https://hub.arcgis.com/datasets/1b2d03addfaa41bb83c17f9237e1504c",
 				"retrieved": "2026-10-19",
 				"status": "live"
+			}, {
+				"url": "https://services2.arcgis.com/CyVvlIiUfRBmMQuu/arcgis/rest/services/Delinquent_Real_Estate_Taxes_view/FeatureServer/0",
+				"retrieved": "2026-10-19",
+				"status": "live"
 			}],
 			"licence": {
 				"kind": "custom",
@@ -53976,10 +55022,52 @@ var catalogue_default = {
 			},
 			"citations": [],
 			"recipe": {
-				"status": "not-fetchable",
-				"reason": "The source is an ArcGIS feature service that returns GeoJSON in pages of a limited number of features and keepset has no GeoJSON parser or paging yet."
+				"status": "fetchable",
+				"source": 1,
+				"paging": {
+					"kind": "arcgis-query",
+					"orderBy": "OBJECTID"
+				},
+				"archive": { "kind": "none" },
+				"parser": {
+					"format": "geojson",
+					"encoding": "utf-8",
+					"idColumn": "OBJECTID",
+					"numericColumns": [{
+						"column": "Total_Delinquent_Amount_Due",
+						"factor": 1
+					}],
+					"dateColumns": [],
+					"columns": ["Total_Delinquent_Amount_Due", "District"]
+				}
+			},
+			"crosswalk": {
+				"source": "geojson",
+				"target": "or/v1",
+				"rules": [
+					{
+						"op": "move",
+						"from": "meta",
+						"to": "meta"
+					},
+					{
+						"op": "move",
+						"from": "nodes[].id",
+						"to": "nodes[].id"
+					},
+					{
+						"op": "move",
+						"from": "nodes[].Total_Delinquent_Amount_Due",
+						"to": "nodes[].prize"
+					},
+					{
+						"op": "move",
+						"from": "nodes[].District",
+						"to": "nodes[].territory"
+					}
+				],
+				"oneWay": []
 			},
-			"crosswalk": { "file": "crosswalks/or-v1/nodes.json" },
 			"expectedCounts": {},
 			"properties": { "surveyId": "or-47" }
 		},