jats-xml 0.0.1 → 0.0.2

package/dist/cjs/cli/index.js

@@ -0,0 +1,14 @@
+#!/usr/bin/env node
+"use strict";
+var __importDefault = (this && this.__importDefault) || function (mod) {
+    return (mod && mod.__esModule) ? mod : { "default": mod };
+};
+Object.defineProperty(exports, "__esModule", { value: true });
+const commander_1 = __importDefault(require("commander"));
+const version_1 = __importDefault(require("../version"));
+const parse_1 = require("./parse");
+const program = new commander_1.default.Command();
+(0, parse_1.addDownloadCLI)(program);
+program.version(`v${version_1.default}`, '-v, --version', 'Print the current version of jtex');
+program.option('-d, --debug', 'Log out any errors to the console.');
+program.parse(process.argv);

package/dist/cjs/cli/parse.js

@@ -0,0 +1,199 @@
+"use strict";
+var __awaiter = (this && this.__awaiter) || function (thisArg, _arguments, P, generator) {
+    function adopt(value) { return value instanceof P ? value : new P(function (resolve) { resolve(value); }); }
+    return new (P || (P = Promise))(function (resolve, reject) {
+        function fulfilled(value) { try { step(generator.next(value)); } catch (e) { reject(e); } }
+        function rejected(value) { try { step(generator["throw"](value)); } catch (e) { reject(e); } }
+        function step(result) { result.done ? resolve(result.value) : adopt(result.value).then(fulfilled, rejected); }
+        step((generator = generator.apply(thisArg, _arguments || [])).next());
+    });
+};
+var __importDefault = (this && this.__importDefault) || function (mod) {
+    return (mod && mod.__esModule) ? mod : { "default": mod };
+};
+Object.defineProperty(exports, "__esModule", { value: true });
+exports.addDownloadCLI = void 0;
+const commander_1 = require("commander");
+const fs_1 = __importDefault(require("fs"));
+const path_1 = require("path");
+const myst_cli_utils_1 = require("myst-cli-utils");
+const doi_utils_1 = __importDefault(require("doi-utils"));
+const chalk_1 = __importDefault(require("chalk"));
+const session_1 = require("../session");
+const types_1 = require("../types");
+const jats_1 = require("../jats");
+const myst_common_1 = require("myst-common");
+const unist_util_select_1 = require("unist-util-select");
+const download_1 = require("../download");
+const resolvers_1 = require("../resolvers");
+const utils_1 = require("../utils");
+function hasValidExtension(output) {
+    return ['.xml', '.jats'].includes((0, path_1.extname)(output).toLowerCase());
+}
+function downloadJats(session, urlOrDoi, output) {
+    return __awaiter(this, void 0, void 0, function* () {
+        if (fs_1.default.existsSync(urlOrDoi)) {
+            throw new Error(`File "${urlOrDoi}" is local and cannot be downloaded!`);
+        }
+        if (!(doi_utils_1.default.validate(urlOrDoi) || (0, myst_cli_utils_1.isUrl)(urlOrDoi))) {
+            throw new Error(`Path must be a URL or DOI, not "${urlOrDoi}"`);
+        }
+        if (!hasValidExtension(output)) {
+            session.log.warn(`The extension ${(0, path_1.extname)(output)} is not a valid extension for JATS, try using ".xml" or ".jats"`);
+        }
+        const data = yield (0, download_1.downloadJatsFromUrl)(session, urlOrDoi, resolvers_1.DEFAULT_RESOLVERS);
+        (0, myst_cli_utils_1.writeFileToFolder)(output, data);
+        return data;
+    });
+}
+function parseJats(session, file) {
+    return __awaiter(this, void 0, void 0, function* () {
+        const toc = (0, myst_cli_utils_1.tic)();
+        if (fs_1.default.existsSync(file)) {
+            session.log.debug(`Found ${file} locally, parsing`);
+            const data = fs_1.default.readFileSync(file).toString();
+            session.log.debug(toc(`Parsed JATS file from disk in %s`));
+            return new jats_1.Jats(data);
+        }
+        if (doi_utils_1.default.validate(file) || (0, myst_cli_utils_1.isUrl)(file)) {
+            const data = yield (0, download_1.downloadJatsFromUrl)(session, file, resolvers_1.DEFAULT_RESOLVERS);
+            session.log.debug(toc(`Downloaded and parsed JATS file in %s`));
+            return new jats_1.Jats(data);
+        }
+        throw new Error(`Could not find ${file} locally, and it doesn't look like a URL or DOI`);
+    });
+}
+function formatLongString(data, offset = 0, length = 88 - offset) {
+    const out = [data.slice(0, length)];
+    let left = data.slice(length);
+    while (left.length > length) {
+        out.push(left.slice(0, length).trim());
+        left = left.slice(length);
+    }
+    if (left)
+        out.push(left.trim());
+    return out.join(`\n${' '.repeat(offset)}`);
+}
+function formatDictionary(dict, opts) {
+    const maxLabel = Object.keys(dict).reduce((a, b) => Math.max(a, b.length), 0);
+    return Object.entries(dict)
+        .map(([k, t]) => {
+        var _a;
+        let value = t;
+        let color = chalk_1.default.yellow.bold;
+        if (t && typeof t === 'object') {
+            if (!t.value)
+                return null;
+            color = (_a = t.label) !== null && _a !== void 0 ? _a : color;
+            value = t.value;
+        }
+        const wrapped = (opts === null || opts === void 0 ? void 0 : opts.wrap) === false
+            ? String(value)
+            : formatLongString(String(value), maxLabel + 2, opts === null || opts === void 0 ? void 0 : opts.wrap);
+        return `${color(k)}:${' '.repeat(maxLabel - k.length + 1)}${wrapped}`;
+    })
+        .filter((o) => !!o)
+        .join('\n');
+}
+function jatsSummaryCLI(session, file) {
+    var _a, _b;
+    return __awaiter(this, void 0, void 0, function* () {
+        const jats = yield parseJats(session, file);
+        const summary = {
+            DOI: jats.doi ? doi_utils_1.default.buildUrl(jats.doi) : null,
+            Title: (_a = (0, myst_common_1.toText)(jats.articleTitle)) === null || _a === void 0 ? void 0 : _a.replace(/\n/g, ' '),
+            Date: (0, utils_1.formatDate)((0, utils_1.toDate)(jats.publicationDate)),
+            Authors: jats.articleAuthors
+                .map((a) => `${(0, myst_common_1.toText)((0, unist_util_select_1.select)(types_1.Tags.givenNames, a))} ${(0, myst_common_1.toText)((0, unist_util_select_1.select)(types_1.Tags.surname, a))}`)
+                .join(', '),
+            Abstract: (_b = (0, myst_common_1.toText)(jats.abstract)) === null || _b === void 0 ? void 0 : _b.replace(/\n/g, ' '),
+            Keywords: jats.keywords.map((k) => (0, myst_common_1.toText)(k)).join(', '),
+            License: jats.license['xlink:href'],
+            Figures: { label: chalk_1.default.blue.bold, value: String((0, unist_util_select_1.selectAll)(types_1.Tags.fig, jats.body).length) },
+            Equations: {
+                label: chalk_1.default.blue.bold,
+                value: String((0, unist_util_select_1.selectAll)(types_1.Tags.dispFormula, jats.body).length),
+            },
+            Tables: {
+                label: chalk_1.default.blue.bold,
+                value: String((0, unist_util_select_1.selectAll)(types_1.Tags.table, jats.body).length),
+            },
+            Code: {
+                label: chalk_1.default.blue.bold,
+                value: String((0, unist_util_select_1.selectAll)(types_1.Tags.code, jats.body).length),
+            },
+            Sections: {
+                label: chalk_1.default.blue.bold,
+                value: String((0, unist_util_select_1.selectAll)(types_1.Tags.sec, jats.body).length),
+            },
+            Paragraphs: {
+                label: chalk_1.default.blue.bold,
+                value: String((0, unist_util_select_1.selectAll)(types_1.Tags.p, jats.body).length),
+            },
+            Citations: { label: chalk_1.default.blue.bold, value: String(jats.references.length) },
+            'Cross-References': {
+                label: chalk_1.default.blue.bold,
+                value: String((0, unist_util_select_1.selectAll)(types_1.Tags.xref, jats.body).length),
+            },
+            'Sub Articles': { label: chalk_1.default.blue.bold, value: String(jats.subArticles.length) },
+        };
+        session.log.info(formatDictionary(summary));
+    });
+}
+function jatsReferencesCLI(session, file) {
+    return __awaiter(this, void 0, void 0, function* () {
+        const jats = yield parseJats(session, file);
+        const sorted = jats.references
+            .map((ref) => {
+            const doi = (0, utils_1.findDoi)(ref);
+            const title = (0, myst_common_1.toText)((0, unist_util_select_1.select)(types_1.Tags.articleTitle, ref));
+            const year = (0, myst_common_1.toText)((0, unist_util_select_1.select)(types_1.Tags.year, ref));
+            const surnames = (0, unist_util_select_1.selectAll)(types_1.Tags.surname, ref);
+            const short = surnames.length > 2
+                ? (0, myst_common_1.toText)(surnames[0]) + ' et al.'
+                : surnames.length === 2
+                    ? (0, myst_common_1.toText)(surnames[0]) + ' and ' + (0, myst_common_1.toText)(surnames[1])
+                    : (0, myst_common_1.toText)(surnames[0]);
+            const s = (0, unist_util_select_1.selectAll)(`[rid=${ref.id}]`, jats.body);
+            return {
+                Citation: `${short} (${year})`,
+                Title: title,
+                DOI: doi ? doi_utils_1.default.buildUrl(doi) : null,
+                Count: s.length,
+            };
+        })
+            .sort((a, b) => b.Count - a.Count);
+        sorted.forEach((r) => {
+            session.log.info(formatDictionary(r, { wrap: false }) + '\n');
+        });
+    });
+}
+function makeSummaryCLI(program) {
+    const command = new commander_1.Command('summary')
+        .description('Parse a JATS file and provide a summary')
+        .argument('<jats>', 'The JATS file or remote URL to be parsed')
+        .action((0, myst_cli_utils_1.clirun)(jatsSummaryCLI, { program, getSession: session_1.getSession }));
+    return command;
+}
+function makeReferencesCLI(program) {
+    const command = new commander_1.Command('refs')
+        .alias('references')
+        .description('Parse a JATS file and provide a summary')
+        .argument('<jats>', 'The JATS file or remote URL to be parsed')
+        .action((0, myst_cli_utils_1.clirun)(jatsReferencesCLI, { program, getSession: session_1.getSession }));
+    return command;
+}
+function makeDownloadCLI(program) {
+    const command = new commander_1.Command('download')
+        .description('Parse a JATS file and provide a summary')
+        .argument('<url>', 'The JATS url or a DOI')
+        .argument('<output>', 'The JATS url or a DOI')
+        .action((0, myst_cli_utils_1.clirun)(downloadJats, { program, getSession: session_1.getSession }));
+    return command;
+}
+function addDownloadCLI(program) {
+    program.addCommand(makeDownloadCLI(program));
+    program.addCommand(makeSummaryCLI(program));
+    program.addCommand(makeReferencesCLI(program));
+}
+exports.addDownloadCLI = addDownloadCLI;

package/dist/cjs/download.js

@@ -0,0 +1,48 @@
+"use strict";
+var __awaiter = (this && this.__awaiter) || function (thisArg, _arguments, P, generator) {
+    function adopt(value) { return value instanceof P ? value : new P(function (resolve) { resolve(value); }); }
+    return new (P || (P = Promise))(function (resolve, reject) {
+        function fulfilled(value) { try { step(generator.next(value)); } catch (e) { reject(e); } }
+        function rejected(value) { try { step(generator["throw"](value)); } catch (e) { reject(e); } }
+        function step(result) { result.done ? resolve(result.value) : adopt(result.value).then(fulfilled, rejected); }
+        step((generator = generator.apply(thisArg, _arguments || [])).next());
+    });
+};
+var __importDefault = (this && this.__importDefault) || function (mod) {
+    return (mod && mod.__esModule) ? mod : { "default": mod };
+};
+Object.defineProperty(exports, "__esModule", { value: true });
+exports.downloadJatsFromUrl = void 0;
+const doi_utils_1 = __importDefault(require("doi-utils"));
+const node_fetch_1 = __importDefault(require("node-fetch"));
+const myst_cli_utils_1 = require("myst-cli-utils");
+const resolvers_1 = require("./resolvers");
+function dowloadFromUrl(session, jatsUrl) {
+    return __awaiter(this, void 0, void 0, function* () {
+        session.log.debug(`Fetching JATS from ${jatsUrl}`);
+        const jatsResp = yield (0, node_fetch_1.default)(jatsUrl);
+        if (!jatsResp.ok)
+            throw new Error(`Problem fetching JATS from ${jatsUrl}`);
+        const contentType = jatsResp.headers.get('content-type');
+        if (!(contentType === 'application/xml' || (contentType === null || contentType === void 0 ? void 0 : contentType.includes('text/plain')))) {
+            throw new Error(`Expected content-type "application/xml" instead we got ${contentType} for ${jatsUrl}`);
+        }
+        const data = yield jatsResp.text();
+        return data;
+    });
+}
+function downloadJatsFromUrl(session, urlOrDoi, resolvers) {
+    return __awaiter(this, void 0, void 0, function* () {
+        if (doi_utils_1.default.validate(urlOrDoi)) {
+            const jatsUrl = yield (0, resolvers_1.resolveJatsUrlFromDoi)(session, urlOrDoi, resolvers);
+            const data = yield dowloadFromUrl(session, jatsUrl);
+            return data;
+        }
+        if ((0, myst_cli_utils_1.isUrl)(urlOrDoi)) {
+            const data = yield dowloadFromUrl(session, urlOrDoi);
+            return data;
+        }
+        throw new Error(`Could not find ${urlOrDoi} locally, and it doesn't look like a URL or DOI`);
+    });
+}
+exports.downloadJatsFromUrl = downloadJatsFromUrl;

package/dist/cjs/index.js

@@ -13,6 +13,13 @@ var __createBinding = (this && this.__createBinding) || (Object.create ? (functi
 var __exportStar = (this && this.__exportStar) || function(m, exports) {
     for (var p in m) if (p !== "default" && !Object.prototype.hasOwnProperty.call(exports, p)) __createBinding(exports, m, p);
 };
+var __importDefault = (this && this.__importDefault) || function (mod) {
+    return (mod && mod.__esModule) ? mod : { "default": mod };
+};
 Object.defineProperty(exports, "__esModule", { value: true });
+exports.Jats = exports.version = void 0;
+var version_1 = require("./version");
+Object.defineProperty(exports, "version", { enumerable: true, get: function () { return __importDefault(version_1).default; } });
+var jats_1 = require("./jats");
+Object.defineProperty(exports, "Jats", { enumerable: true, get: function () { return jats_1.Jats; } });
 __exportStar(require("./types"), exports);
-//# sourceMappingURL=index.js.map

package/dist/cjs/jats.js

@@ -0,0 +1,87 @@
+"use strict";
+Object.defineProperty(exports, "__esModule", { value: true });
+exports.Jats = void 0;
+const xml_js_1 = require("xml-js");
+const utils_1 = require("./utils");
+const unist_util_select_1 = require("unist-util-select");
+const types_1 = require("./types");
+class Jats {
+    constructor(data) {
+        this.raw = (0, xml_js_1.xml2js)(data, { compact: false });
+        const { declaration, elements } = this.raw;
+        this.declaration = declaration === null || declaration === void 0 ? void 0 : declaration.attributes;
+        if (!((elements === null || elements === void 0 ? void 0 : elements.length) === 2 && elements[0].type === 'doctype' && elements[1].name === 'article')) {
+            throw new Error('article is not the only element of the JATS');
+        }
+        this.doctype = elements[0].doctype;
+        this.tree = (0, utils_1.convertToUnist)(elements[1]);
+    }
+    get front() {
+        return (0, unist_util_select_1.select)(types_1.Tags.front, this.tree);
+    }
+    get premissions() {
+        return (0, unist_util_select_1.select)(types_1.Tags.permissions, this.front);
+    }
+    get doi() {
+        return (0, utils_1.findDoi)(this.front);
+    }
+    get publicationDates() {
+        return (0, unist_util_select_1.selectAll)(types_1.Tags.pubDate, this.front);
+    }
+    get publicationDate() {
+        return this.publicationDates.find((d) => !!(0, unist_util_select_1.select)(types_1.Tags.day, d));
+    }
+    get license() {
+        return (0, unist_util_select_1.select)(types_1.Tags.license, this.premissions);
+    }
+    get keywordGroup() {
+        return (0, unist_util_select_1.select)(types_1.Tags.kwdGroup, this.front);
+    }
+    /** The first keywords */
+    get keywords() {
+        return (0, unist_util_select_1.selectAll)(types_1.Tags.kwd, this.keywordGroup);
+    }
+    get keywordGroups() {
+        return (0, unist_util_select_1.selectAll)(types_1.Tags.kwdGroup, this.front);
+    }
+    get titleGroup() {
+        return (0, unist_util_select_1.select)(types_1.Tags.titleGroup, this.front);
+    }
+    get articleTitle() {
+        return (0, unist_util_select_1.select)(types_1.Tags.articleTitle, this.titleGroup);
+    }
+    get articleSubtitle() {
+        return (0, unist_util_select_1.select)(types_1.Tags.subtitle, this.titleGroup);
+    }
+    get abstract() {
+        return (0, unist_util_select_1.select)(types_1.Tags.abstract, this.front);
+    }
+    get abstracts() {
+        return (0, unist_util_select_1.selectAll)(types_1.Tags.abstract, this.front);
+    }
+    get contribGroup() {
+        return (0, unist_util_select_1.select)(types_1.Tags.contribGroup, this.front);
+    }
+    get contribGroups() {
+        return (0, unist_util_select_1.selectAll)(types_1.Tags.contribGroup, this.front);
+    }
+    get articleAuthors() {
+        return (0, unist_util_select_1.selectAll)(types_1.Tags.contrib, this.contribGroup);
+    }
+    get body() {
+        return (0, unist_util_select_1.select)(types_1.Tags.body, this.tree);
+    }
+    get back() {
+        return (0, unist_util_select_1.select)(types_1.Tags.back, this.tree);
+    }
+    get subArticles() {
+        return (0, unist_util_select_1.selectAll)(types_1.Tags.subArticle, this.tree);
+    }
+    get refList() {
+        return (0, unist_util_select_1.select)(types_1.Tags.refList, this.back);
+    }
+    get references() {
+        return (0, unist_util_select_1.selectAll)(types_1.Tags.ref, this.refList);
+    }
+}
+exports.Jats = Jats;

package/dist/cjs/resolvers.js

@@ -0,0 +1,63 @@
+"use strict";
+var __awaiter = (this && this.__awaiter) || function (thisArg, _arguments, P, generator) {
+    function adopt(value) { return value instanceof P ? value : new P(function (resolve) { resolve(value); }); }
+    return new (P || (P = Promise))(function (resolve, reject) {
+        function fulfilled(value) { try { step(generator.next(value)); } catch (e) { reject(e); } }
+        function rejected(value) { try { step(generator["throw"](value)); } catch (e) { reject(e); } }
+        function step(result) { result.done ? resolve(result.value) : adopt(result.value).then(fulfilled, rejected); }
+        step((generator = generator.apply(thisArg, _arguments || [])).next());
+    });
+};
+var __importDefault = (this && this.__importDefault) || function (mod) {
+    return (mod && mod.__esModule) ? mod : { "default": mod };
+};
+Object.defineProperty(exports, "__esModule", { value: true });
+exports.resolveJatsUrlFromDoi = exports.DEFAULT_RESOLVERS = exports.joss = exports.plos = exports.elife = void 0;
+const doi_utils_1 = __importDefault(require("doi-utils"));
+const node_fetch_1 = __importDefault(require("node-fetch"));
+exports.elife = {
+    test(url) {
+        return new URL(url).hostname === 'elifesciences.org';
+    },
+    jatsUrl(url) {
+        return `${url}.xml`;
+    },
+};
+exports.plos = {
+    test(url) {
+        return new URL(url).hostname === 'journals.plos.org';
+    },
+    jatsUrl(url) {
+        return url.replace('/article?', '/article/file?') + '&type=manuscript';
+    },
+};
+exports.joss = {
+    test(url) {
+        return new URL(url).hostname === 'joss.theoj.org';
+    },
+    jatsUrl(url) {
+        // Probably a better way to do this, the joss papers on on github!
+        const doi = new URL(url).pathname.replace('/papers/', '');
+        const [org, jossId] = doi.split('/');
+        const id = jossId.split('.')[1];
+        return `https://raw.githubusercontent.com/openjournals/joss-papers/master/joss.${id}/${org}.${jossId}.jats`;
+    },
+};
+exports.DEFAULT_RESOLVERS = [exports.elife, exports.plos, exports.joss];
+function resolveJatsUrlFromDoi(session, doi, resolvers = exports.DEFAULT_RESOLVERS) {
+    return __awaiter(this, void 0, void 0, function* () {
+        if (!doi_utils_1.default.validate(doi))
+            throw new Error(`The doi ${doi} is not valid`);
+        const doiUrl = doi_utils_1.default.buildUrl(doi);
+        session.log.debug(`Resolving DOI ${doiUrl}`);
+        const resp = yield (0, node_fetch_1.default)(doiUrl);
+        const articleUrl = resp.url;
+        session.log.debug(`Found resolved URL for DOI at ${articleUrl}`);
+        const resolver = resolvers.find((r) => r.test(articleUrl));
+        if (!resolver)
+            throw new Error(`Could not resolve JATS for ${articleUrl}, no resolver matched`);
+        const jatsUrl = resolver.jatsUrl(articleUrl);
+        return jatsUrl;
+    });
+}
+exports.resolveJatsUrlFromDoi = resolveJatsUrlFromDoi;

package/dist/cjs/session.js

@@ -0,0 +1,16 @@
+"use strict";
+Object.defineProperty(exports, "__esModule", { value: true });
+exports.getSession = exports.Session = void 0;
+const myst_cli_utils_1 = require("myst-cli-utils");
+class Session {
+    constructor(opts) {
+        var _a;
+        this.API_URL = 'https://api.myst.tools';
+        this.log = (_a = opts === null || opts === void 0 ? void 0 : opts.logger) !== null && _a !== void 0 ? _a : (0, myst_cli_utils_1.chalkLogger)(myst_cli_utils_1.LogLevel.debug);
+    }
+}
+exports.Session = Session;
+function getSession(logger) {
+    return new Session({ logger });
+}
+exports.getSession = getSession;