jats-utils 1.0.7

package/dist/index.d.ts

@@ -0,0 +1,3 @@
+export * from './utils.js';
+export * from './serialize.js';
+//# sourceMappingURL=index.d.ts.map

package/dist/index.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../src/index.ts"],"names":[],"mappings":"AAAA,cAAc,YAAY,CAAC;AAC3B,cAAc,gBAAgB,CAAC"}

package/dist/index.js

@@ -0,0 +1,2 @@
+export * from './utils.js';
+export * from './serialize.js';

package/dist/serialize.d.ts

@@ -0,0 +1,12 @@
+import { type Element } from 'xml-js';
+export type SerializationOptions = {
+    /**
+     * When 'pretty', the default, the xml be formatted in a custom, opinionated way
+     * When 'flat', the xml will be on a single line
+     * When `0`, the XML will be on different lines with 0 spaces.
+     * When any other value (e.g. `2` or `\t`) the XML will be indented at the start of the line by that amount.
+     */
+    format?: number | 'flat' | 'pretty' | '\t';
+};
+export declare function serializeJatsXml(element: Element, opts?: SerializationOptions): string;
+//# sourceMappingURL=serialize.d.ts.map

package/dist/serialize.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"serialize.d.ts","sourceRoot":"","sources":["../src/serialize.ts"],"names":[],"mappings":"AAAA,OAAO,EAAU,KAAK,OAAO,EAAE,MAAM,QAAQ,CAAC;AAG9C,MAAM,MAAM,oBAAoB,GAAG;IACjC;;;;;OAKG;IACH,MAAM,CAAC,EAAE,MAAM,GAAG,MAAM,GAAG,QAAQ,GAAG,IAAI,CAAC;CAC5C,CAAC;AA+EF,wBAAgB,gBAAgB,CAAC,OAAO,EAAE,OAAO,EAAE,IAAI,CAAC,EAAE,oBAAoB,UAgB7E"}

package/dist/serialize.js

@@ -0,0 +1,95 @@
+import { js2xml } from 'xml-js';
+import { escapeForXML } from './utils.js';
+const both = [
+    '\\?xml',
+    'article',
+    'sub-article',
+    'front',
+    'front-stub',
+    'journal-meta',
+    'journal-title-group',
+    'publisher',
+    'article-meta',
+    'article-categories',
+    'title-group',
+    'contrib-group',
+    'contrib',
+    'institution-wrap',
+    'aff',
+    'permissions',
+    'license',
+    'kwd-group',
+    'history',
+    'self-uri',
+    'funding-group',
+    'award-group',
+    'principal-award-recipient',
+    'custom-meta-group',
+    'date',
+    'pub-date',
+    'abstract',
+    'counts',
+    'body',
+    'sec',
+    'fig',
+    'disp-formula',
+    'table-wrap',
+    'caption',
+    'table',
+    'thead',
+    'ref-list',
+    'ref',
+    'back',
+];
+const first = [
+    'journal-id',
+    'journal-title',
+    'issn',
+    'publisher-name',
+    'publisher-loc',
+    'article-id',
+    'article-title',
+    'alt-title',
+    'subtitle',
+    'kwd',
+    'name',
+    'email',
+    'contrib-id',
+    'role',
+    'institution',
+    'institution-id',
+    'award-id',
+    'meta-name',
+    'meta-value',
+    'title',
+    'p',
+    'tr',
+    'label',
+    'graphic',
+    'mixed-citation',
+];
+function indentXML(xml) {
+    return xml
+        .replace(RegExp(`<(\\/)?(${both.join('|')})( [^>]*)?>`, 'g'), '<$1$2$3>\n')
+        .replace(RegExp(`([^\n])<(\\/)?(${both.join('|')})( [^>]*)?>`, 'g'), '$1\n<$2$3$4>')
+        .replace(RegExp(`([^\n])<(${first.join('|')})( [^>]*)?>`, 'g'), '$1\n<$2$3>')
+        .replace(RegExp(`<\\/(${first.join('|')})(\\s*)>([^\n])`, 'g'), '</$1>\n$3');
+}
+export function serializeJatsXml(element, opts) {
+    const { format } = { format: 'pretty', ...opts };
+    const xml = js2xml(element, {
+        compact: false,
+        //  No way to write XML with new lines, but no indentation with js2xml.
+        // If you use 0 or '', you get a single line.
+        spaces: format === 'flat' || format === 'pretty' ? 0 : format || 1,
+        attributeValueFn: escapeForXML,
+    });
+    if (format === 0) {
+        // either `0` or `''`
+        return xml.replace(/\n(\s*)</g, '\n<');
+    }
+    else if (format === 'pretty') {
+        return indentXML(xml);
+    }
+    return xml;
+}

package/dist/utils.d.ts

@@ -0,0 +1,12 @@
+import type { GenericNode, GenericParent } from 'myst-common';
+import type { Element } from 'xml-js';
+import type { Contributor } from 'myst-frontmatter';
+export declare function convertToUnist(node: Element): GenericNode | GenericParent | undefined;
+export declare function convertToXml(node: GenericNode): Element;
+export declare function escapeForXML(text: string): string;
+export declare function toDate(date?: GenericParent): Date | undefined;
+export declare function formatDate(date?: Date): string | undefined;
+export type PubIdTypes = 'doi' | 'pmc' | 'pmid' | 'publisher-id' | string;
+export declare function findArticleId(node: GenericParent | undefined, pubIdType?: PubIdTypes): string | undefined;
+export declare function authorAndAffiliation(node: GenericParent, article: GenericParent): Contributor;
+//# sourceMappingURL=utils.d.ts.map

package/dist/utils.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"utils.d.ts","sourceRoot":"","sources":["../src/utils.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,WAAW,EAAE,aAAa,EAAE,MAAM,aAAa,CAAC;AAE9D,OAAO,KAAK,EAAE,OAAO,EAAE,MAAM,QAAQ,CAAC;AAKtC,OAAO,KAAK,EAAE,WAAW,EAAE,MAAM,kBAAkB,CAAC;AAEpD,wBAAgB,cAAc,CAAC,IAAI,EAAE,OAAO,GAAG,WAAW,GAAG,aAAa,GAAG,SAAS,CA2CrF;AAED,wBAAgB,YAAY,CAAC,IAAI,EAAE,WAAW,GAAG,OAAO,CAwBvD;AAED,wBAAgB,YAAY,CAAC,IAAI,EAAE,MAAM,UAExC;AA6BD,wBAAgB,MAAM,CAAC,IAAI,CAAC,EAAE,aAAa,GAAG,IAAI,GAAG,SAAS,CAa7D;AAED,wBAAgB,UAAU,CAAC,IAAI,CAAC,EAAE,IAAI,sBAGrC;AAED,MAAM,MAAM,UAAU,GAAG,KAAK,GAAG,KAAK,GAAG,MAAM,GAAG,cAAc,GAAG,MAAM,CAAC;AAE1E,wBAAgB,aAAa,CAC3B,IAAI,EAAE,aAAa,GAAG,SAAS,EAC/B,SAAS,GAAE,UAAkB,GAC5B,MAAM,GAAG,SAAS,CAQpB;AAED,wBAAgB,oBAAoB,CAAC,IAAI,EAAE,aAAa,EAAE,OAAO,EAAE,aAAa,GAAG,WAAW,CAyC7F"}

package/dist/utils.js

@@ -0,0 +1,179 @@
+import { toText } from 'myst-common';
+import { doi } from 'doi-utils';
+import { select, selectAll } from 'unist-util-select';
+import { Tags } from 'jats-tags';
+export function convertToUnist(node) {
+    switch (node.type) {
+        case 'element': {
+            const { name, attributes, elements } = node;
+            const children = elements === null || elements === void 0 ? void 0 : elements.map(convertToUnist).filter((n) => !!n);
+            const { type, ...attrs } = attributes !== null && attributes !== void 0 ? attributes : {};
+            if (type !== undefined)
+                attrs._type = type;
+            const next = { type: name !== null && name !== void 0 ? name : 'unknown', ...attrs };
+            if (name === 'code') {
+                next.value = elements === null || elements === void 0 ? void 0 : elements[0].text;
+            }
+            else if (children)
+                next.children = children;
+            return next;
+        }
+        case 'text': {
+            const { attributes, text } = node;
+            return {
+                type: 'text',
+                ...attributes,
+                value: String(text).replace(/\n(\s+)$/, ''),
+            };
+        }
+        case 'cdata': {
+            const { attributes, cdata } = node;
+            return {
+                type: 'cdata',
+                ...attributes,
+                cdata: String(cdata).trim(),
+            };
+        }
+        case 'comment': {
+            const { comment } = node;
+            return { type: 'comment', value: String(comment) };
+        }
+        case 'instruction': {
+            // For example:
+            // <?properties manuscript?> becomes:
+            // { type: 'instruction', name: 'properties', instruction: 'manuscript' }
+            return undefined;
+        }
+        default:
+            console.log(node);
+            throw new Error(`found ${node.type} ${node.name}`);
+    }
+}
+export function convertToXml(node) {
+    const { type, ...rest } = node;
+    switch (type) {
+        case 'text': {
+            const { value, ...attributes } = rest;
+            return { type: 'text', attributes, text: value };
+        }
+        case 'code': {
+            const { value, ...attributes } = rest;
+            return { type: 'element', name: type, attributes, elements: [{ type: 'text', text: value }] };
+        }
+        case 'comment': {
+            return { type: 'comment', comment: rest.value };
+        }
+        case 'cdata': {
+            const { cdata, ...attributes } = rest;
+            return { type: 'cdata', attributes, cdata };
+        }
+        default: {
+            const { children, _type, ...attributes } = rest;
+            if (_type !== undefined)
+                attributes.type = _type;
+            return { type: 'element', name: type, attributes, elements: children === null || children === void 0 ? void 0 : children.map(convertToXml) };
+        }
+    }
+}
+export function escapeForXML(text) {
+    return text.replace(/&(?!amp;)/g, '&amp;').replace(/</g, '&lt;');
+}
+const MonthLookup = {
+    jan: 0,
+    january: 0,
+    feb: 1,
+    february: 1,
+    mar: 2,
+    march: 2,
+    apr: 3,
+    april: 3,
+    may: 4,
+    jun: 5,
+    june: 5,
+    jul: 6,
+    july: 6,
+    aug: 7,
+    august: 7,
+    sep: 8,
+    sept: 8,
+    september: 8,
+    oct: 9,
+    october: 9,
+    nov: 10,
+    november: 10,
+    dec: 11,
+    december: 11,
+};
+export function toDate(date) {
+    if (!date)
+        return;
+    const isoDate = date['iso-8601-date'];
+    if (isoDate)
+        return new Date(isoDate);
+    const year = Number(toText(select('year', date)));
+    if (!year || Number.isNaN(year))
+        return;
+    const monthText = toText(select('month', date));
+    const monthTextNumber = Number(monthText);
+    const month = Number.isNaN(monthTextNumber) ? MonthLookup[monthText] : monthTextNumber - 1;
+    if (month == null)
+        return new Date(Date.UTC(year, 0));
+    const day = Number(toText(select('day', date)));
+    if (!day || Number.isNaN(day))
+        return new Date(Date.UTC(year, month));
+    return new Date(Date.UTC(year, month, day));
+}
+export function formatDate(date) {
+    if (!date)
+        return;
+    return new Intl.DateTimeFormat('en-US', { dateStyle: 'long', timeZone: 'UTC' }).format(date);
+}
+export function findArticleId(node, pubIdType = 'doi') {
+    if (!node)
+        return undefined;
+    const id = select(`[pub-id-type=${pubIdType}]`, node);
+    if (id && toText(id))
+        return toText(id);
+    const doiTag = selectAll(`${Tags.articleId},${Tags.pubId}`, node).find((t) => doi.validate(toText(t)));
+    return toText(doiTag) || undefined;
+}
+export function authorAndAffiliation(node, article) {
+    const author = {
+        name: `${toText(select(Tags.givenNames, node))} ${toText(select(Tags.surname, node))}`,
+    };
+    const orcid = select('[contrib-id-type=orcid]', node);
+    if (orcid) {
+        author.orcid = toText(orcid).replace(/(https?:\/\/)?orcid\.org\//, '');
+    }
+    //
+    /**
+     * For example:
+     *
+     * ```xml
+     * <aff id="aff2">
+     * <label>2</label>
+     * <institution-wrap>
+     * <institution-id institution-id-type="ror">https://ror.org/00t9vx427</institution-id>
+     * <institution>Department of Biochemistry, University of Texas Southwestern Medical Center</institution>
+     * </institution-wrap>
+     * <addr-line>
+     * <named-content content-type="city">Dallas</named-content>
+     * </addr-line>
+     * <country>United States</country>
+     * </aff>
+     * ```
+     */
+    const affiliationRefs = selectAll('xref[ref-type=aff]', node);
+    const affiliations = affiliationRefs.map((xref) => select(`[id=${xref.rid}]`, article));
+    const affiliationText = affiliations
+        .map((aff) => {
+        // TODO: handle rors!
+        const ror = select(`[institution-id-type=ror]`, aff);
+        return toText(select('institution', aff));
+    })
+        .filter((t) => !!t);
+    if (affiliationText.length > 0) {
+        author.affiliations = affiliationText;
+    }
+    return author;
+}

package/package.json

@@ -0,0 +1,46 @@
+{
+  "name": "jats-utils",
+  "version": "1.0.7",
+  "description": "Utility functions for working with JATS in Typescript",
+  "author": "Rowan Cockett <rowan@curvenote.com>",
+  "homepage": "https://github.com/curvenote/jats",
+  "license": "MIT",
+  "sideEffects": false,
+  "type": "module",
+  "exports": "./dist/index.js",
+  "types": "./dist/index.d.ts",
+  "files": [
+    "dist"
+  ],
+  "keywords": [
+    "jats",
+    "open-science",
+    "publishing"
+  ],
+  "publishConfig": {
+    "access": "public"
+  },
+  "repository": {
+    "type": "git",
+    "url": "git+https://github.com/curvenote/jats.git"
+  },
+  "scripts": {
+    "clean": "rm -rf dist",
+    "lint": "eslint \"src/**/*.ts*\" -c ./.eslintrc.cjs",
+    "lint:format": "prettier --check \"src/**/*.{ts,tsx,md}\"",
+    "build:esm": "tsc",
+    "build": "npm-run-all -l clean -p build:esm"
+  },
+  "bugs": {
+    "url": "https://github.com/curvenote/jats/issues"
+  },
+  "dependencies": {},
+  "peerDependencies": {
+    "xml-js": "^1",
+    "jats-tags": "^1"
+  },
+  "devDependencies": {
+    "myst-common": "^1.0.0",
+    "jats-tags": "^1"
+  }
+}