keepset 0.14.1 → 0.16.0

package/dist/{categories-B4hhikbr.d.ts → categories-BYezdHCk.d.ts}

@@ -98,6 +98,8 @@ declare const ArchiveSchema: z.ZodDiscriminatedUnion<[z.ZodObject<{
   }, z.core.$strict>]>;
 }, z.core.$strict>], "kind">;
 type Archive = z.infer<typeof ArchiveSchema>;
+/** A column of a table as the table parser names it: the header with every run of characters outside letters, digits and underscore replaced by one underscore, so it is a valid segment of a twindex path. */
+declare const ColumnNameSchema: z.ZodString;
 /** The parser that reads the unpacked file into the dataset's native JSON shape, with the options that format needs. Every option is stated in the entry, so a recipe reads the same way wherever it runs. */
 declare const ParserSchema: z.ZodDiscriminatedUnion<[z.ZodObject<{
   format: z.ZodLiteral<"edge-list">;
@@ -1163,4 +1165,4 @@ declare function listCategories(): readonly Category[];
  */
 declare function countTarget(id: CategoryId, target: CategoryTarget): ReadonlyMap<string, number>;
 //#endregion
-export { ParserSchema as $, LineageSchema as A, Citation as B, CatalogueIndex as C, NamespaceNode as D, DatasetQuery as E, AliasSchema as F, EntrySchema as G, CrosswalkReference as H, Archive as I, Licence as J, HttpUrlSchema as K, ArchiveMember as L, ManifestSchema as M, SemVerSchema as N, createCatalogue as O, parseManifest as P, Parser as Q, ArchiveMemberSchema as R, Catalogue as S, Dataset as T, CrosswalkReferenceSchema as U, CitationSchema as V, Entry as W, Namespace as X, LicenceSchema as Y, NamespaceSchema as Z, GraphEdgeSchema as _, CategoryTargets as a, SnapshotSchema as at, GraphV1 as b, getCategory as c, WaybackTimestampSchema as ct, listCategories as d, RESERVED_META_KEYS as et, OrNode as f, GraphEdge as g, OrV1Schema as h, CategoryTarget as i, Snapshot as it, Manifest as j, Lineage as k, getTarget as l, parseEntry as lt, OrV1 as m, CategoryId as n, RecipeSchema as nt, CountedKey as o, Source as ot, OrNodeSchema as p, IdSchema as q, CategoryIdSchema as r, RelativePathSchema as rt, countTarget as s, SourceSchema as st, Category as t, Recipe as tt, kindOf as u, waybackUrl as ut, GraphNode as v, CatalogueIndexSchema as w, GraphV1Schema as x, GraphNodeSchema as y, ArchiveSchema as z };
+export { Parser as $, LineageSchema as A, Citation as B, CatalogueIndex as C, NamespaceNode as D, DatasetQuery as E, AliasSchema as F, Entry as G, ColumnNameSchema as H, Archive as I, IdSchema as J, EntrySchema as K, ArchiveMember as L, ManifestSchema as M, SemVerSchema as N, createCatalogue as O, parseManifest as P, NamespaceSchema as Q, ArchiveMemberSchema as R, Catalogue as S, Dataset as T, CrosswalkReference as U, CitationSchema as V, CrosswalkReferenceSchema as W, LicenceSchema as X, Licence as Y, Namespace as Z, GraphEdgeSchema as _, CategoryTargets as a, Snapshot as at, GraphV1 as b, getCategory as c, SourceSchema as ct, listCategories as d, waybackUrl as dt, ParserSchema as et, OrNode as f, GraphEdge as g, OrV1Schema as h, CategoryTarget as i, RelativePathSchema as it, Manifest as j, Lineage as k, getTarget as l, WaybackTimestampSchema as lt, OrV1 as m, CategoryId as n, Recipe as nt, CountedKey as o, SnapshotSchema as ot, OrNodeSchema as p, HttpUrlSchema as q, CategoryIdSchema as r, RecipeSchema as rt, countTarget as s, Source as st, Category as t, RESERVED_META_KEYS as tt, kindOf as u, parseEntry as ut, GraphNode as v, CatalogueIndexSchema as w, GraphV1Schema as x, GraphNodeSchema as y, ArchiveSchema as z };

package/dist/index.d.ts

@@ -1,4 +1,4 @@
-import { $ as ParserSchema, A as LineageSchema, B as Citation, C as CatalogueIndex, D as NamespaceNode, E as DatasetQuery, F as AliasSchema, G as EntrySchema, H as CrosswalkReference, I as Archive, J as Licence, K as HttpUrlSchema, L as ArchiveMember, M as ManifestSchema, N as SemVerSchema, O as createCatalogue, P as parseManifest, Q as Parser, R as ArchiveMemberSchema, S as Catalogue, T as Dataset, U as CrosswalkReferenceSchema, V as CitationSchema, W as Entry, X as Namespace, Y as LicenceSchema, Z as NamespaceSchema, _ as GraphEdgeSchema, a as CategoryTargets, at as SnapshotSchema, b as GraphV1, c as getCategory, ct as WaybackTimestampSchema, d as listCategories, et as RESERVED_META_KEYS, f as OrNode, g as GraphEdge, h as OrV1Schema, i as CategoryTarget, it as Snapshot, j as Manifest, k as Lineage, l as getTarget, lt as parseEntry, m as OrV1, n as CategoryId, nt as RecipeSchema, o as CountedKey, ot as Source, p as OrNodeSchema, q as IdSchema, r as CategoryIdSchema, rt as RelativePathSchema, s as countTarget, st as SourceSchema, t as Category, tt as Recipe, u as kindOf, ut as waybackUrl, v as GraphNode, w as CatalogueIndexSchema, x as GraphV1Schema, y as GraphNodeSchema, z as ArchiveSchema } from "./categories-B4hhikbr.js";
+import { $ as Parser, A as LineageSchema, B as Citation, C as CatalogueIndex, D as NamespaceNode, E as DatasetQuery, F as AliasSchema, G as Entry, H as ColumnNameSchema, I as Archive, J as IdSchema, K as EntrySchema, L as ArchiveMember, M as ManifestSchema, N as SemVerSchema, O as createCatalogue, P as parseManifest, Q as NamespaceSchema, R as ArchiveMemberSchema, S as Catalogue, T as Dataset, U as CrosswalkReference, V as CitationSchema, W as CrosswalkReferenceSchema, X as LicenceSchema, Y as Licence, Z as Namespace, _ as GraphEdgeSchema, a as CategoryTargets, at as Snapshot, b as GraphV1, c as getCategory, ct as SourceSchema, d as listCategories, dt as waybackUrl, et as ParserSchema, f as OrNode, g as GraphEdge, h as OrV1Schema, i as CategoryTarget, it as RelativePathSchema, j as Manifest, k as Lineage, l as getTarget, lt as WaybackTimestampSchema, m as OrV1, n as CategoryId, nt as Recipe, o as CountedKey, ot as SnapshotSchema, p as OrNodeSchema, q as HttpUrlSchema, r as CategoryIdSchema, rt as RecipeSchema, s as countTarget, st as Source, t as Category, tt as RESERVED_META_KEYS, u as kindOf, ut as parseEntry, v as GraphNode, w as CatalogueIndexSchema, x as GraphV1Schema, y as GraphNodeSchema, z as ArchiveSchema } from "./categories-BYezdHCk.js";
 import { Crosswalk } from "twindex";
 import { z } from "zod";
 //#region src/errors.d.ts
@@ -123,4 +123,4 @@ type JsonScalar = z.infer<typeof JsonScalarSchema>;
  */
 export declare function validateDataset(dataset: Dataset): SyntaxIssue[];
 //#endregion
-export { AliasSchema, type Archive, type ArchiveMember, ArchiveMemberSchema, ArchiveSchema, type Catalogue, type CatalogueIndex, CatalogueIndexSchema, type Category, type CategoryId, CategoryIdSchema, type CategoryTarget, type CategoryTargets, type Citation, CitationSchema, type CountedKey, type CrosswalkReference, CrosswalkReferenceSchema, type Dataset, type DatasetQuery, type Entry, type EntryNames, EntrySchema, type GraphEdge, GraphEdgeSchema, type GraphNode, GraphNodeSchema, type GraphV1, GraphV1Schema, HttpUrlSchema, IdSchema, type JsonScalar, type Licence, LicenceSchema, type Lineage, LineageSchema, type Manifest, ManifestSchema, type Namespace, type NamespaceNode, NamespaceSchema, type OrNode, OrNodeSchema, type OrV1, OrV1Schema, type Parser, ParserSchema, RESERVED_META_KEYS, type Recipe, RecipeSchema, type RecipeStep, RelativePathSchema, type SegmentAliases, SemVerSchema, type Snapshot, SnapshotSchema, type Source, SourceSchema, type SyntaxIssue, WaybackTimestampSchema, countTarget, createCatalogue, getCategory, getTarget, kindOf, listCategories, parseEntry, parseManifest, waybackUrl };
+export { AliasSchema, type Archive, type ArchiveMember, ArchiveMemberSchema, ArchiveSchema, type Catalogue, type CatalogueIndex, CatalogueIndexSchema, type Category, type CategoryId, CategoryIdSchema, type CategoryTarget, type CategoryTargets, type Citation, CitationSchema, ColumnNameSchema, type CountedKey, type CrosswalkReference, CrosswalkReferenceSchema, type Dataset, type DatasetQuery, type Entry, type EntryNames, EntrySchema, type GraphEdge, GraphEdgeSchema, type GraphNode, GraphNodeSchema, type GraphV1, GraphV1Schema, HttpUrlSchema, IdSchema, type JsonScalar, type Licence, LicenceSchema, type Lineage, LineageSchema, type Manifest, ManifestSchema, type Namespace, type NamespaceNode, NamespaceSchema, type OrNode, OrNodeSchema, type OrV1, OrV1Schema, type Parser, ParserSchema, RESERVED_META_KEYS, type Recipe, RecipeSchema, type RecipeStep, RelativePathSchema, type SegmentAliases, SemVerSchema, type Snapshot, SnapshotSchema, type Source, SourceSchema, type SyntaxIssue, WaybackTimestampSchema, countTarget, createCatalogue, getCategory, getTarget, kindOf, listCategories, parseEntry, parseManifest, waybackUrl };

package/dist/index.js

@@ -1,4 +1,4 @@
-import { $ as RecipeError, A as waybackUrl, B as listCategories, C as RecipeSchema, D as WaybackTimestampSchema, E as SourceSchema, F as CategoryIdSchema, G as GraphV1Schema, H as OrV1Schema, I as countTarget, J as CatalogueSyntaxError, K as JsonScalarSchema, L as getCategory, M as groupsOf, N as namesOf, O as parseEntry, Q as NativeParseError, R as getTarget, S as RESERVED_META_KEYS, T as SnapshotSchema, U as GraphEdgeSchema, V as OrNodeSchema, W as GraphNodeSchema, X as DatasetNotFoundError, Y as DatasetAmbiguousError, Z as KeepsetError, _ as HttpUrlSchema, a as CatalogueIndexSchema, b as NamespaceSchema, c as ManifestSchema, d as AliasSchema, et as UnknownCategoryError, f as ArchiveMemberSchema, g as EntrySchema, h as CrosswalkReferenceSchema, i as listNamespaces, j as NAME_SEPARATOR, k as toIssues, l as SemVerSchema, m as CitationSchema, n as getManifest, o as createCatalogue, p as ArchiveSchema, q as ArchiveError, r as listDatasets, s as LineageSchema, t as getDataset, u as parseManifest, v as IdSchema, w as RelativePathSchema, x as ParserSchema, y as LicenceSchema, z as kindOf } from "./bundled-D9eVvy1i.js";
+import { $ as NativeParseError, A as toIssues, B as kindOf, C as RESERVED_META_KEYS, D as SourceSchema, E as SnapshotSchema, G as GraphNodeSchema, H as OrNodeSchema, I as CategoryIdSchema, J as ArchiveError, K as GraphV1Schema, L as countTarget, M as NAME_SEPARATOR, N as groupsOf, O as WaybackTimestampSchema, P as namesOf, Q as KeepsetError, R as getCategory, S as ParserSchema, T as RelativePathSchema, U as OrV1Schema, V as listCategories, W as GraphEdgeSchema, X as DatasetAmbiguousError, Y as CatalogueSyntaxError, Z as DatasetNotFoundError, _ as EntrySchema, a as CatalogueIndexSchema, b as LicenceSchema, c as ManifestSchema, d as AliasSchema, et as RecipeError, f as ArchiveMemberSchema, g as CrosswalkReferenceSchema, h as ColumnNameSchema, i as listNamespaces, j as waybackUrl, k as parseEntry, l as SemVerSchema, m as CitationSchema, n as getManifest, o as createCatalogue, p as ArchiveSchema, q as JsonScalarSchema, r as listDatasets, s as LineageSchema, t as getDataset, tt as UnknownCategoryError, u as parseManifest, v as HttpUrlSchema, w as RecipeSchema, x as NamespaceSchema, y as IdSchema, z as getTarget } from "./bundled-DPEAZdj9.js";
 import { validateCrosswalk } from "twindex";
 //#region src/serialise.ts
 /**
@@ -57,4 +57,4 @@ function validateDataset(dataset) {
 	return issues;
 }
 //#endregion
-export { AliasSchema, ArchiveError, ArchiveMemberSchema, ArchiveSchema, CatalogueIndexSchema, CatalogueSyntaxError, CategoryIdSchema, CitationSchema, CrosswalkReferenceSchema, DatasetAmbiguousError, DatasetNotFoundError, EntrySchema, GraphEdgeSchema, GraphNodeSchema, GraphV1Schema, HttpUrlSchema, IdSchema, JsonScalarSchema, KeepsetError, LicenceSchema, LineageSchema, ManifestSchema, NAME_SEPARATOR, NamespaceSchema, NativeParseError, OrNodeSchema, OrV1Schema, ParserSchema, RESERVED_META_KEYS, RecipeError, RecipeSchema, RelativePathSchema, SemVerSchema, SnapshotSchema, SourceSchema, UnknownCategoryError, WaybackTimestampSchema, countTarget, createCatalogue, getCategory, getDataset, getManifest, getTarget, groupsOf, kindOf, listCategories, listDatasets, listNamespaces, namesOf, parseEntry, parseManifest, serialise, validateDataset, waybackUrl };
+export { AliasSchema, ArchiveError, ArchiveMemberSchema, ArchiveSchema, CatalogueIndexSchema, CatalogueSyntaxError, CategoryIdSchema, CitationSchema, ColumnNameSchema, CrosswalkReferenceSchema, DatasetAmbiguousError, DatasetNotFoundError, EntrySchema, GraphEdgeSchema, GraphNodeSchema, GraphV1Schema, HttpUrlSchema, IdSchema, JsonScalarSchema, KeepsetError, LicenceSchema, LineageSchema, ManifestSchema, NAME_SEPARATOR, NamespaceSchema, NativeParseError, OrNodeSchema, OrV1Schema, ParserSchema, RESERVED_META_KEYS, RecipeError, RecipeSchema, RelativePathSchema, SemVerSchema, SnapshotSchema, SourceSchema, UnknownCategoryError, WaybackTimestampSchema, countTarget, createCatalogue, getCategory, getDataset, getManifest, getTarget, groupsOf, kindOf, listCategories, listDatasets, listNamespaces, namesOf, parseEntry, parseManifest, serialise, validateDataset, waybackUrl };

package/dist/node/fetch-dataset.d.ts

@@ -1,4 +1,4 @@
-import { T as Dataset, W as Entry, i as CategoryTarget, it as Snapshot, ot as Source } from "../categories-B4hhikbr.js";
+import { G as Entry, T as Dataset, at as Snapshot, i as CategoryTarget, st as Source } from "../categories-BYezdHCk.js";
 //#region src/node/fetch-dataset.d.ts
 /** The network port a recipe downloads through: `downloadRaw` by default, which returns the bytes exactly as served, or a stand-in under test. */
 export type FetchPort = (url: string, init: {

package/dist/node/fetch-dataset.js

@@ -1,4 +1,4 @@
-import { $ as RecipeError, I as countTarget, P as getOwn, Q as NativeParseError, R as getTarget, Z as KeepsetError, q as ArchiveError, t as getDataset } from "../bundled-D9eVvy1i.js";
+import { $ as NativeParseError, F as getOwn, J as ArchiveError, L as countTarget, Q as KeepsetError, et as RecipeError, t as getDataset, z as getTarget } from "../bundled-DPEAZdj9.js";
 import { RoundTripMismatchError, assertRoundTrip, encode } from "twindex";
 import { createHash } from "node:crypto";
 import { gunzipSync } from "node:zlib";
@@ -359,16 +359,36 @@ const DECIMAL = /^[+-]?(?:\d+\.?\d*|\.\d+)(?:[eE][+-]?\d+)?$/;
 /** The byte order mark some exporters put before the first header. */
 const BYTE_ORDER_MARK = "";
 /**
-* The name the parser gives a column: the header with every run of characters outside letters, digits and underscore replaced by one underscore, leading and trailing underscores dropped, and a name starting with a digit prefixed with `c_`. The rule depends on the header alone, so a crosswalk written against a file's headers reads the same wherever it runs.
+* The name the parser gives a column: the header with every run of characters outside letters, digits and underscore replaced by one underscore, leading and trailing underscores dropped, and a name starting with a digit prefixed with `c_`. A header with nothing left (an empty cell, as in a table exported with an unnamed index column) is `column_` and its 1-based position. The rule depends on the header and its position alone, so a crosswalk written against a file's headers reads the same wherever it runs.
 * @param header - A header cell.
-* @returns The column name, or undefined when nothing of the header is left.
+* @param position - The cell's 1-based position among the header cells.
+* @returns The column name.
 */
-function columnName(header) {
+function columnName(header, position) {
 	const name = header.trim().replaceAll(/[^A-Za-z0-9_]+/g, "_").replaceAll(/^_+|_+$/g, "");
-	if (name === "") return;
+	if (name === "") return `column_${String(position)}`;
 	return /^\d/.test(name) ? `c_${name}` : name;
 }
 /**
+* The names the parser gives a header row's cells, one per cell in order: each cell's `columnName`, except that `id` is the row's id, so a cell of that name which is not the id column is named by its position, as an unnamed one is. The id column is among them; it is only left out of a row's other columns.
+* @param cells - The header cells.
+* @param idColumn - The recipe's id column, or null for the row number.
+* @param line - The 1-based line of the header, for the error.
+* @returns The names.
+* @throws NativeParseError when two cells get the same name.
+*/
+function columnNames(cells, idColumn, line) {
+	const names = [];
+	for (const [position, cell] of cells.entries()) {
+		const named = columnName(cell, position + 1);
+		const name = named === "id" && idColumn !== "id" ? `column_${String(position + 1)}` : named;
+		const clash = names.indexOf(name);
+		if (clash !== -1) throw new NativeParseError("csv-table", line, `header cells ${String(clash + 1)} and ${String(position + 1)} both become the column ${name}`);
+		names.push(name);
+	}
+	return names;
+}
+/**
 * Splits text into records and fields by RFC 4180: a field may be quoted, a quoted field may hold the delimiter, line breaks and doubled quotes, and a record ends at an unquoted LF, CRLF or CR.
 * @param text - The file text, after the lines the recipe skips.
 * @param delimiter - The field delimiter.
@@ -454,26 +474,28 @@ function numeric$1(cell, factor, column, line) {
 	return Number(cell) * factor;
 }
 /**
-* Parses a delimited table into rows: the first record after `skipLines` lines is the header, each later record is a row, and every row carries every column but the id column, an empty cell being null. Cells are trimmed. A column the recipe declares numeric is read as a decimal number times its factor and any other column stays a string. A row's id is the cell of the recipe's id column, which is then not also a column, or its 1-based row number when the recipe names none. Blank lines are skipped; a row with a different number of fields than the header, a header cell with no name, two columns with the same name, and an id cell that is empty are errors.
+* The records of a table file after the lines the recipe skips and without its blank lines: the header first, then the rows.
+* @param text - The file text, which may begin with a byte order mark.
+* @param options - The recipe's table parser, of which the delimiter and the lines to skip are read.
+* @returns The records.
+* @throws NativeParseError when a quote is never closed or opens inside an unquoted field, or the file has fewer lines than the recipe skips.
+*/
+function tableRecords(text, options) {
+	return records(afterLines(text.startsWith(BYTE_ORDER_MARK) ? text.slice(1) : text, options.skipLines), DELIMITERS[options.delimiter], options.skipLines + 1).filter((record) => record.fields.length > 1 || record.fields[0]?.trim() !== "");
+}
+/**
+* Parses a delimited table into rows: the first record after `skipLines` lines is the header, each later record is a row, and every row carries every column but the id column, an empty cell being null. Cells are trimmed. A column the recipe declares numeric is read as a decimal number times its factor and any other column stays a string. `id` is the row's id, so a column of that name which is not the id column is named by its position. A row's id is the cell of the recipe's id column, which is then not also a column, or its 1-based row number when the recipe names none. Blank lines are skipped; a row with a different number of fields than the header, two columns with the same name, and an id cell that is empty are errors.
 * @param text - The file text.
 * @param options - The recipe's table parser.
 * @returns The rows, in file order.
 * @throws NativeParseError naming the line of the first problem.
 */
 function parseCsvTable(text, options) {
-	const [header, ...rows] = records(afterLines(text.startsWith(BYTE_ORDER_MARK) ? text.slice(1) : text, options.skipLines), DELIMITERS[options.delimiter], options.skipLines + 1).filter((record) => record.fields.length > 1 || record.fields[0]?.trim() !== "");
+	const [header, ...rows] = tableRecords(text, options);
 	if (header === void 0) throw new NativeParseError("csv-table", options.skipLines + 1, "the file has no header row");
-	const names = [];
-	for (const [position, cell] of header.fields.entries()) {
-		const name = columnName(cell);
-		if (name === void 0) throw new NativeParseError("csv-table", header.line, `header cell ${String(position + 1)} (${JSON.stringify(cell)}) has no name left once its characters other than letters, digits and underscores are removed`);
-		const clash = names.indexOf(name);
-		if (clash !== -1) throw new NativeParseError("csv-table", header.line, `header cells ${String(clash + 1)} and ${String(position + 1)} both become the column ${name}`);
-		names.push(name);
-	}
+	const names = columnNames(header.fields, options.idColumn, header.line);
 	const { idColumn, numericColumns } = options;
 	for (const wanted of [...idColumn === null ? [] : [idColumn], ...numericColumns.map((entry) => entry.column)]) if (!names.includes(wanted)) throw new NativeParseError("csv-table", header.line, `the recipe names the column ${wanted}, which the header does not have`);
-	if (idColumn === null && names.includes("id")) throw new NativeParseError("csv-table", header.line, "the file has a column named id, which would be shadowed by the row number; the recipe names it as its idColumn");
 	const factors = new Map(numericColumns.map((entry) => [entry.column, entry.factor]));
 	return { nodes: rows.map((record, rowIndex) => {
 		if (record.fields.length !== names.length) throw new NativeParseError("csv-table", record.line, `the row has ${String(record.fields.length)} fields, the header has ${String(names.length)}`);

package/dist-cli/keepset.mjs

@@ -4,7 +4,7 @@ import { KeepsetError, NAME_SEPARATOR, countTarget, getDataset, listDatasets, li
 import { fetchDataset, roundTripDataset } from "keepset/node";
 import { parseArgs } from "node:util";
 //#region package.json
-var version = "0.14.1";
+var version = "0.16.0";
 const usage = `Usage: keepset <command> [options]
 
 Commands:

package/package.json

@@ -1,7 +1,7 @@
 {
   "name": "keepset",
   "description": "Library, CLI and statically generated catalogue for dataset fetch-and-normalise recipes",
-  "version": "0.14.1",
+  "version": "0.16.0",
   "bin": {
     "keepset": "./dist-cli/keepset.mjs"
   },
@@ -93,8 +93,11 @@
     "lint:check": "eslint . --cache --max-warnings 0",
     "prepare": "husky",
     "prepublishOnly": "pnpm run lint:check && pnpm run build && pnpm run typecheck && pnpm run test && publint && attw --pack --profile esm-only",
-    "seed:check": "tsx scripts/seed/import-phd-catalogue.ts --pipeline scripts/seed/sources/pipeline-catalog.json --graphbox scripts/seed/sources/graphbox-catalog.json --check",
-    "seed:import": "tsx scripts/seed/import-phd-catalogue.ts",
+    "seed:check": "pnpm run seed:check:graph && pnpm run seed:check:or",
+    "seed:check:graph": "tsx scripts/seed/import-phd-catalogue.ts --pipeline scripts/seed/sources/pipeline-catalog.json --graphbox scripts/seed/sources/graphbox-catalog.json --check",
+    "seed:check:or": "tsx scripts/seed/import-or-survey.ts --check",
+    "seed:import:graph": "tsx scripts/seed/import-phd-catalogue.ts",
+    "seed:import:or": "tsx scripts/seed/import-or-survey.ts",
     "site": "tsx site/build.ts",
     "test": "vitest run",
     "typecheck": "tsc -p tsconfig.json && tsc -p tsconfig.node.json && tsc -p tsconfig.site.json && tsc -p tsconfig.sw.json"

package/schemas/entry.schema.json

@@ -1 +1 @@
-{"$defs":{"Crosswalk":{"$ref":"https://cdn.jsdelivr.net/npm/twindex@0.1.4/schemas/twindex.schema.json"},"Entry":{"additionalProperties":false,"properties":{"aliases":{"items":{"pattern":"^[A-Za-z0-9][A-Za-z0-9._-]*(?:\\/[A-Za-z0-9][A-Za-z0-9._-]*)*$","type":"string"},"minItems":1,"type":"array"},"category":{"enum":["graph/v1","or/v1"],"type":"string"},"citations":{"items":{"additionalProperties":false,"properties":{"doi":{"pattern":"^10\\.\\d{4,9}\\/\\S+$","type":"string"},"text":{"minLength":1,"type":"string"},"url":{"format":"uri","type":"string"}},"required":["text"],"type":"object"},"type":"array"},"crosswalk":{"anyOf":[{"$ref":"#/$defs/Crosswalk"},{"additionalProperties":false,"properties":{"file":{"type":"string"}},"required":["file"],"type":"object"}]},"description":{"minLength":1,"type":"string"},"domain":{"minLength":1,"type":"string"},"expectedCounts":{"additionalProperties":{"maximum":9007199254740991,"minimum":0,"type":"integer"},"propertyNames":{"pattern":"^[A-Za-z][A-Za-z0-9]*$","type":"string"},"type":"object"},"id":{"pattern":"^[A-Za-z0-9][A-Za-z0-9._-]*$","type":"string"},"licence":{"oneOf":[{"additionalProperties":false,"properties":{"id":{"pattern":"^[A-Za-z0-9][A-Za-z0-9.+-]*$","type":"string"},"kind":{"const":"spdx","type":"string"}},"required":["kind","id"],"type":"object"},{"additionalProperties":false,"properties":{"kind":{"const":"custom","type":"string"},"name":{"minLength":1,"type":"string"},"url":{"format":"uri","type":"string"}},"required":["kind","name","url"],"type":"object"},{"additionalProperties":false,"properties":{"kind":{"const":"unrecorded","type":"string"}},"required":["kind"],"type":"object"}]},"name":{"minLength":1,"type":"string"},"namespace":{"items":{"pattern":"^[A-Za-z0-9][A-Za-z0-9._-]*$","type":"string"},"minItems":1,"type":"array"},"properties":{"additionalProperties":{"type":["string","number","boolean","null"]},"propertyNames":{"pattern":"^[A-Za-z][A-Za-z0-9]*$","type":"string"},"type":"object"},"recipe":{"oneOf":[{"additionalProperties":false,"properties":{"archive":{"oneOf":[{"additionalProperties":false,"properties":{"kind":{"const":"none","type":"string"}},"required":["kind"],"type":"object"},{"additionalProperties":false,"properties":{"kind":{"const":"gzip","type":"string"}},"required":["kind"],"type":"object"},{"additionalProperties":false,"properties":{"kind":{"const":"zip","type":"string"},"member":{"anyOf":[{"additionalProperties":false,"properties":{"path":{"type":"string"}},"required":["path"],"type":"object"},{"additionalProperties":false,"properties":{"extensions":{"items":{"pattern":"^\\.[A-Za-z0-9][A-Za-z0-9.]*$","type":"string"},"minItems":1,"type":"array"}},"required":["extensions"],"type":"object"}]}},"required":["kind","member"],"type":"object"},{"additionalProperties":false,"properties":{"kind":{"const":"tar","type":"string"},"member":{"anyOf":[{"additionalProperties":false,"properties":{"path":{"type":"string"}},"required":["path"],"type":"object"},{"additionalProperties":false,"properties":{"extensions":{"items":{"pattern":"^\\.[A-Za-z0-9][A-Za-z0-9.]*$","type":"string"},"minItems":1,"type":"array"}},"required":["extensions"],"type":"object"}]}},"required":["kind","member"],"type":"object"},{"additionalProperties":false,"properties":{"kind":{"const":"tar.gz","type":"string"},"member":{"anyOf":[{"additionalProperties":false,"properties":{"path":{"type":"string"}},"required":["path"],"type":"object"},{"additionalProperties":false,"properties":{"extensions":{"items":{"pattern":"^\\.[A-Za-z0-9][A-Za-z0-9.]*$","type":"string"},"minItems":1,"type":"array"}},"required":["extensions"],"type":"object"}]}},"required":["kind","member"],"type":"object"}]},"parser":{"oneOf":[{"additionalProperties":false,"properties":{"commentPrefixes":{"items":{"minLength":1,"type":"string"},"type":"array"},"delimiter":{"enum":["detect","whitespace","tab","comma","space"],"type":"string"},"encoding":{"enum":["utf-8","latin1"],"type":"string"},"format":{"const":"edge-list","type":"string"},"skipLines":{"maximum":9007199254740991,"minimum":0,"type":"integer"},"sourceColumn":{"maximum":9007199254740991,"minimum":0,"type":"integer"},"targetColumn":{"maximum":9007199254740991,"minimum":0,"type":"integer"}},"required":["format","encoding","delimiter","skipLines","commentPrefixes","sourceColumn","targetColumn"],"type":"object"},{"additionalProperties":false,"properties":{"encoding":{"enum":["utf-8","latin1"],"type":"string"},"format":{"const":"matrix-market","type":"string"}},"required":["format","encoding"],"type":"object"},{"additionalProperties":false,"properties":{"encoding":{"enum":["utf-8","latin1"],"type":"string"},"format":{"const":"gml","type":"string"}},"required":["format","encoding"],"type":"object"},{"additionalProperties":false,"properties":{"encoding":{"enum":["utf-8","latin1"],"type":"string"},"format":{"const":"pajek","type":"string"}},"required":["format","encoding"],"type":"object"},{"additionalProperties":false,"properties":{"delimiter":{"enum":["comma","tab","semicolon"],"type":"string"},"encoding":{"enum":["utf-8","latin1"],"type":"string"},"format":{"const":"csv-table","type":"string"},"idColumn":{"anyOf":[{"pattern":"^[A-Za-z](?:[A-Za-z0-9_]*[A-Za-z0-9])?$","type":"string"},{"type":"null"}]},"numericColumns":{"items":{"additionalProperties":false,"properties":{"column":{"pattern":"^[A-Za-z](?:[A-Za-z0-9_]*[A-Za-z0-9])?$","type":"string"},"factor":{"exclusiveMinimum":0,"type":"number"}},"required":["column","factor"],"type":"object"},"type":"array"},"skipLines":{"maximum":9007199254740991,"minimum":0,"type":"integer"}},"required":["format","encoding","delimiter","skipLines","idColumn","numericColumns"],"type":"object"}]},"source":{"maximum":9007199254740991,"minimum":0,"type":"integer"},"status":{"const":"fetchable","type":"string"}},"required":["status","source","archive","parser"],"type":"object"},{"additionalProperties":false,"properties":{"reason":{"minLength":1,"type":"string"},"status":{"const":"not-fetchable","type":"string"}},"required":["status","reason"],"type":"object"}]},"sources":{"items":{"additionalProperties":false,"properties":{"retrieved":{"format":"date","pattern":"^(?:(?:\\d\\d[2468][048]|\\d\\d[13579][26]|\\d\\d0[48]|[02468][048]00|[13579][26]00)-02-29|\\d{4}-(?:(?:0[13578]|1[02])-(?:0[1-9]|[12]\\d|3[01])|(?:0[469]|11)-(?:0[1-9]|[12]\\d|30)|(?:02)-(?:0[1-9]|1\\d|2[0-8])))$","type":"string"},"snapshot":{"additionalProperties":false,"properties":{"timestamp":{"pattern":"^\\d{8}(?:\\d{6})?$","type":"string"},"url":{"format":"uri","type":"string"}},"required":["url","timestamp"],"type":"object"},"status":{"enum":["live","dead"],"type":"string"},"url":{"format":"uri","type":"string"}},"required":["url","retrieved","status"],"type":"object"},"minItems":1,"type":"array"},"tags":{"items":{"minLength":1,"type":"string"},"type":"array"}},"required":["id","name","tags","category","sources","licence","citations","recipe","crosswalk","expectedCounts","properties"],"type":"object"}},"$id":"https://cdn.jsdelivr.net/npm/keepset@0.14.1/schemas/entry.schema.json","$ref":"#/$defs/Entry","$schema":"https://json-schema.org/draft/2020-12/schema"}
+{"$defs":{"Crosswalk":{"$ref":"https://cdn.jsdelivr.net/npm/twindex@0.1.4/schemas/twindex.schema.json"},"Entry":{"additionalProperties":false,"properties":{"aliases":{"items":{"pattern":"^[A-Za-z0-9][A-Za-z0-9._-]*(?:\\/[A-Za-z0-9][A-Za-z0-9._-]*)*$","type":"string"},"minItems":1,"type":"array"},"category":{"enum":["graph/v1","or/v1"],"type":"string"},"citations":{"items":{"additionalProperties":false,"properties":{"doi":{"pattern":"^10\\.\\d{4,9}\\/\\S+$","type":"string"},"text":{"minLength":1,"type":"string"},"url":{"format":"uri","type":"string"}},"required":["text"],"type":"object"},"type":"array"},"crosswalk":{"anyOf":[{"$ref":"#/$defs/Crosswalk"},{"additionalProperties":false,"properties":{"file":{"type":"string"}},"required":["file"],"type":"object"}]},"description":{"minLength":1,"type":"string"},"domain":{"minLength":1,"type":"string"},"expectedCounts":{"additionalProperties":{"maximum":9007199254740991,"minimum":0,"type":"integer"},"propertyNames":{"pattern":"^[A-Za-z][A-Za-z0-9]*$","type":"string"},"type":"object"},"id":{"pattern":"^[A-Za-z0-9][A-Za-z0-9._-]*$","type":"string"},"licence":{"oneOf":[{"additionalProperties":false,"properties":{"id":{"pattern":"^[A-Za-z0-9][A-Za-z0-9.+-]*$","type":"string"},"kind":{"const":"spdx","type":"string"}},"required":["kind","id"],"type":"object"},{"additionalProperties":false,"properties":{"kind":{"const":"custom","type":"string"},"name":{"minLength":1,"type":"string"},"url":{"format":"uri","type":"string"}},"required":["kind","name","url"],"type":"object"},{"additionalProperties":false,"properties":{"kind":{"const":"unrecorded","type":"string"}},"required":["kind"],"type":"object"}]},"name":{"minLength":1,"type":"string"},"namespace":{"items":{"pattern":"^[A-Za-z0-9][A-Za-z0-9._-]*$","type":"string"},"minItems":1,"type":"array"},"properties":{"additionalProperties":{"type":["string","number","boolean","null"]},"propertyNames":{"pattern":"^[A-Za-z][A-Za-z0-9]*$","type":"string"},"type":"object"},"recipe":{"oneOf":[{"additionalProperties":false,"properties":{"archive":{"oneOf":[{"additionalProperties":false,"properties":{"kind":{"const":"none","type":"string"}},"required":["kind"],"type":"object"},{"additionalProperties":false,"properties":{"kind":{"const":"gzip","type":"string"}},"required":["kind"],"type":"object"},{"additionalProperties":false,"properties":{"kind":{"const":"zip","type":"string"},"member":{"anyOf":[{"additionalProperties":false,"properties":{"path":{"type":"string"}},"required":["path"],"type":"object"},{"additionalProperties":false,"properties":{"extensions":{"items":{"pattern":"^\\.[A-Za-z0-9][A-Za-z0-9.]*$","type":"string"},"minItems":1,"type":"array"}},"required":["extensions"],"type":"object"}]}},"required":["kind","member"],"type":"object"},{"additionalProperties":false,"properties":{"kind":{"const":"tar","type":"string"},"member":{"anyOf":[{"additionalProperties":false,"properties":{"path":{"type":"string"}},"required":["path"],"type":"object"},{"additionalProperties":false,"properties":{"extensions":{"items":{"pattern":"^\\.[A-Za-z0-9][A-Za-z0-9.]*$","type":"string"},"minItems":1,"type":"array"}},"required":["extensions"],"type":"object"}]}},"required":["kind","member"],"type":"object"},{"additionalProperties":false,"properties":{"kind":{"const":"tar.gz","type":"string"},"member":{"anyOf":[{"additionalProperties":false,"properties":{"path":{"type":"string"}},"required":["path"],"type":"object"},{"additionalProperties":false,"properties":{"extensions":{"items":{"pattern":"^\\.[A-Za-z0-9][A-Za-z0-9.]*$","type":"string"},"minItems":1,"type":"array"}},"required":["extensions"],"type":"object"}]}},"required":["kind","member"],"type":"object"}]},"parser":{"oneOf":[{"additionalProperties":false,"properties":{"commentPrefixes":{"items":{"minLength":1,"type":"string"},"type":"array"},"delimiter":{"enum":["detect","whitespace","tab","comma","space"],"type":"string"},"encoding":{"enum":["utf-8","latin1"],"type":"string"},"format":{"const":"edge-list","type":"string"},"skipLines":{"maximum":9007199254740991,"minimum":0,"type":"integer"},"sourceColumn":{"maximum":9007199254740991,"minimum":0,"type":"integer"},"targetColumn":{"maximum":9007199254740991,"minimum":0,"type":"integer"}},"required":["format","encoding","delimiter","skipLines","commentPrefixes","sourceColumn","targetColumn"],"type":"object"},{"additionalProperties":false,"properties":{"encoding":{"enum":["utf-8","latin1"],"type":"string"},"format":{"const":"matrix-market","type":"string"}},"required":["format","encoding"],"type":"object"},{"additionalProperties":false,"properties":{"encoding":{"enum":["utf-8","latin1"],"type":"string"},"format":{"const":"gml","type":"string"}},"required":["format","encoding"],"type":"object"},{"additionalProperties":false,"properties":{"encoding":{"enum":["utf-8","latin1"],"type":"string"},"format":{"const":"pajek","type":"string"}},"required":["format","encoding"],"type":"object"},{"additionalProperties":false,"properties":{"delimiter":{"enum":["comma","tab","semicolon"],"type":"string"},"encoding":{"enum":["utf-8","latin1"],"type":"string"},"format":{"const":"csv-table","type":"string"},"idColumn":{"anyOf":[{"pattern":"^[A-Za-z](?:[A-Za-z0-9_]*[A-Za-z0-9])?$","type":"string"},{"type":"null"}]},"numericColumns":{"items":{"additionalProperties":false,"properties":{"column":{"pattern":"^[A-Za-z](?:[A-Za-z0-9_]*[A-Za-z0-9])?$","type":"string"},"factor":{"exclusiveMinimum":0,"type":"number"}},"required":["column","factor"],"type":"object"},"type":"array"},"skipLines":{"maximum":9007199254740991,"minimum":0,"type":"integer"}},"required":["format","encoding","delimiter","skipLines","idColumn","numericColumns"],"type":"object"}]},"source":{"maximum":9007199254740991,"minimum":0,"type":"integer"},"status":{"const":"fetchable","type":"string"}},"required":["status","source","archive","parser"],"type":"object"},{"additionalProperties":false,"properties":{"reason":{"minLength":1,"type":"string"},"status":{"const":"not-fetchable","type":"string"}},"required":["status","reason"],"type":"object"}]},"sources":{"items":{"additionalProperties":false,"properties":{"retrieved":{"format":"date","pattern":"^(?:(?:\\d\\d[2468][048]|\\d\\d[13579][26]|\\d\\d0[48]|[02468][048]00|[13579][26]00)-02-29|\\d{4}-(?:(?:0[13578]|1[02])-(?:0[1-9]|[12]\\d|3[01])|(?:0[469]|11)-(?:0[1-9]|[12]\\d|30)|(?:02)-(?:0[1-9]|1\\d|2[0-8])))$","type":"string"},"snapshot":{"additionalProperties":false,"properties":{"timestamp":{"pattern":"^\\d{8}(?:\\d{6})?$","type":"string"},"url":{"format":"uri","type":"string"}},"required":["url","timestamp"],"type":"object"},"status":{"enum":["live","dead"],"type":"string"},"url":{"format":"uri","type":"string"}},"required":["url","retrieved","status"],"type":"object"},"minItems":1,"type":"array"},"tags":{"items":{"minLength":1,"type":"string"},"type":"array"}},"required":["id","name","tags","category","sources","licence","citations","recipe","crosswalk","expectedCounts","properties"],"type":"object"}},"$id":"https://cdn.jsdelivr.net/npm/keepset@0.16.0/schemas/entry.schema.json","$ref":"#/$defs/Entry","$schema":"https://json-schema.org/draft/2020-12/schema"}

package/schemas/manifest.schema.json

@@ -1 +1 @@
-{"$defs":{"Manifest":{"additionalProperties":false,"properties":{"description":{"minLength":1,"type":"string"},"id":{"pattern":"^[A-Za-z0-9][A-Za-z0-9._-]*$","type":"string"},"lineage":{"additionalProperties":false,"properties":{"catalogue":{"pattern":"^[A-Za-z0-9][A-Za-z0-9._-]*$","type":"string"},"repository":{"format":"uri","type":"string"},"version":{"pattern":"^(?:0|[1-9]\\d*)\\.(?:0|[1-9]\\d*)\\.(?:0|[1-9]\\d*)(?:-[0-9A-Za-z-]+(?:\\.[0-9A-Za-z-]+)*)?(?:\\+[0-9A-Za-z-]+(?:\\.[0-9A-Za-z-]+)*)?$","type":"string"}},"required":["catalogue","version","repository"],"type":"object"},"name":{"minLength":1,"type":"string"},"namespaceAliases":{"additionalProperties":{"items":{"pattern":"^[A-Za-z0-9][A-Za-z0-9._-]*$","type":"string"},"minItems":1,"type":"array"},"propertyNames":{"pattern":"^[A-Za-z0-9][A-Za-z0-9._-]*$","type":"string"},"type":"object"},"repository":{"format":"uri","type":"string"},"version":{"pattern":"^(?:0|[1-9]\\d*)\\.(?:0|[1-9]\\d*)\\.(?:0|[1-9]\\d*)(?:-[0-9A-Za-z-]+(?:\\.[0-9A-Za-z-]+)*)?(?:\\+[0-9A-Za-z-]+(?:\\.[0-9A-Za-z-]+)*)?$","type":"string"}},"required":["id","name","description","version","repository"],"type":"object"}},"$id":"https://cdn.jsdelivr.net/npm/keepset@0.14.1/schemas/manifest.schema.json","$ref":"#/$defs/Manifest","$schema":"https://json-schema.org/draft/2020-12/schema"}
+{"$defs":{"Manifest":{"additionalProperties":false,"properties":{"description":{"minLength":1,"type":"string"},"id":{"pattern":"^[A-Za-z0-9][A-Za-z0-9._-]*$","type":"string"},"lineage":{"additionalProperties":false,"properties":{"catalogue":{"pattern":"^[A-Za-z0-9][A-Za-z0-9._-]*$","type":"string"},"repository":{"format":"uri","type":"string"},"version":{"pattern":"^(?:0|[1-9]\\d*)\\.(?:0|[1-9]\\d*)\\.(?:0|[1-9]\\d*)(?:-[0-9A-Za-z-]+(?:\\.[0-9A-Za-z-]+)*)?(?:\\+[0-9A-Za-z-]+(?:\\.[0-9A-Za-z-]+)*)?$","type":"string"}},"required":["catalogue","version","repository"],"type":"object"},"name":{"minLength":1,"type":"string"},"namespaceAliases":{"additionalProperties":{"items":{"pattern":"^[A-Za-z0-9][A-Za-z0-9._-]*$","type":"string"},"minItems":1,"type":"array"},"propertyNames":{"pattern":"^[A-Za-z0-9][A-Za-z0-9._-]*$","type":"string"},"type":"object"},"repository":{"format":"uri","type":"string"},"version":{"pattern":"^(?:0|[1-9]\\d*)\\.(?:0|[1-9]\\d*)\\.(?:0|[1-9]\\d*)(?:-[0-9A-Za-z-]+(?:\\.[0-9A-Za-z-]+)*)?(?:\\+[0-9A-Za-z-]+(?:\\.[0-9A-Za-z-]+)*)?$","type":"string"}},"required":["id","name","description","version","repository"],"type":"object"}},"$id":"https://cdn.jsdelivr.net/npm/keepset@0.16.0/schemas/manifest.schema.json","$ref":"#/$defs/Manifest","$schema":"https://json-schema.org/draft/2020-12/schema"}