@ezs/basics 2.3.0 → 2.4.0

package/CHANGELOG.md

@@ -3,6 +3,17 @@
 All notable changes to this project will be documented in this file.
 See [Conventional Commits](https://conventionalcommits.org) for commit guidelines.
 
+# [2.4.0](https://github.com/Inist-CNRS/ezs/compare/@ezs/basics@2.3.0...@ezs/basics@2.4.0) (2023-07-17)
+
+
+### Features
+
+* 🎸 add [TARDump] ([d3ff2a9](https://github.com/Inist-CNRS/ezs/commit/d3ff2a99650fc7c9eacd2c1ad015d9907addeebb))
+
+
+
+
+
 # [2.3.0](https://github.com/Inist-CNRS/ezs/compare/@ezs/basics@2.2.0...@ezs/basics@2.3.0) (2023-06-23)
 
 

package/README.md

@@ -27,6 +27,7 @@ npm install @ezs/basics
 -   [OBJFlatten](#objflatten)
 -   [OBJNamespaces](#objnamespaces)
 -   [OBJStandardize](#objstandardize)
+-   [TARDump](#tardump)
 -   [TARExtract](#tarextract)
 -   [TXTConcat](#txtconcat)
 -   [TXTInflection](#txtinflection)
@@ -537,6 +538,22 @@ Output:
 
 Returns **[Object](https://developer.mozilla.org/docs/Web/JavaScript/Reference/Global_Objects/Object)** 
 
+### TARDump
+
+Take all recevied objects and build a tar file
+
+```json
+{
+}
+```
+
+#### Parameters
+
+-   `manifest` **[String](https://developer.mozilla.org/docs/Web/JavaScript/Reference/Global_Objects/String)?** Location path to store files in the tarball
+-   `location` **[String](https://developer.mozilla.org/docs/Web/JavaScript/Reference/Global_Objects/String)** Location path to store files in the tarball (optional, default `data`)
+-   `json` **[String](https://developer.mozilla.org/docs/Web/JavaScript/Reference/Global_Objects/String)** Parse as JSON the content of each file (optional, default `true`)
+-   `compress` **[Boolean](https://developer.mozilla.org/docs/Web/JavaScript/Reference/Global_Objects/Boolean)** Enable gzip compression (optional, default `false`)
+
 ### TARExtract
 
 Take the content of a tar file, extract some files.

package/lib/fetch.js

@@ -0,0 +1,74 @@
+"use strict";
+
+Object.defineProperty(exports, "__esModule", {
+  value: true
+});
+exports.default = fetch;
+
+var _crossFetch = _interopRequireDefault(require("cross-fetch"));
+
+var _proxyFromEnv = require("proxy-from-env");
+
+var _http = _interopRequireDefault(require("http"));
+
+var _https = _interopRequireDefault(require("https"));
+
+var _betterHttpsProxyAgent = require("better-https-proxy-agent");
+
+function _interopRequireDefault(obj) { return obj && obj.__esModule ? obj : { default: obj }; }
+
+const DefaultOptions = {
+  keepAlive: true,
+  timeout: 1000,
+  keepAliveMsecs: 500,
+  maxSockets: 200,
+  maxFreeSockets: 5,
+  maxCachedSessions: 500
+};
+
+const chooseAgent = (parsedURL, options) => {
+  const proxyurl = (0, _proxyFromEnv.getProxyForUrl)(parsedURL.href);
+
+  if (proxyurl) {
+    const proxyRequestOptions = new URL(proxyurl);
+    return new _betterHttpsProxyAgent.Agent(options, proxyRequestOptions);
+  }
+
+  if (parsedURL.protocol === 'https:') {
+    return new _https.default.Agent(options);
+  }
+
+  return new _http.default.Agent(options);
+};
+
+function fetch(url, options) {
+  const opts = options || {};
+  const {
+    keepAlive,
+    timeout,
+    keepAliveMsecs,
+    maxSockets,
+    maxFreeSockets,
+    maxCachedSessions
+  } = { ...options,
+    ...DefaultOptions
+  };
+  let agent = chooseAgent(new URL(url), {
+    keepAlive,
+    timeout,
+    keepAliveMsecs,
+    maxSockets,
+    maxFreeSockets,
+    maxCachedSessions
+  });
+  opts.agent = agent;
+
+  if (opts.signal) {
+    opts.signal.addEventListener('abort', () => {
+      agent.destroy();
+      agent = null;
+    });
+  }
+
+  return (0, _crossFetch.default)(url, options);
+}

package/lib/index.js

@@ -61,6 +61,8 @@ var _zipExtract = _interopRequireDefault(require("./zip-extract"));
 
 var _tarExtract = _interopRequireDefault(require("./tar-extract"));
 
+var _tarDump = _interopRequireDefault(require("./tar-dump"));
+
 var _iniString = _interopRequireDefault(require("./ini-string"));
 
 var _fileSave = _interopRequireDefault(require("./file-save"));
@@ -98,6 +100,7 @@ const funcs = {
   TXTZip: _txtZip.default,
   ZIPExtract: _zipExtract.default,
   TARExtract: _tarExtract.default,
+  TARDump: _tarDump.default,
   INIString: _iniString.default,
   FILESave: _fileSave.default,
   FILELoad: _fileLoad.default,

package/lib/obj-columns.js

@@ -0,0 +1,53 @@
+"use strict";
+
+Object.defineProperty(exports, "__esModule", {
+  value: true
+});
+exports.default = OBJColumns;
+
+/**
+ * Take an `Object` and flatten it to get only one level of keys.
+ *
+ * <caption>Input:</caption>
+ *
+ * ```json
+ * [{
+ *   "foo": {
+ *     "hello": "world"
+ *   },
+ *   "bar": "anything else",
+ *   "baz": 1
+ * }]
+ * ```
+ *
+ * <caption>Output:</caption>
+ *
+ * ```json
+ * [{
+ *   "foo": "{\"hello\":\"world\"}",
+ *   "bar": "anything else",
+ *   "baz": 1
+ * }]
+ * ```
+ *
+ * @name OBJColumns
+ * @alias flatten
+ * @param {undefined} none
+ * @returns {Object}
+ */
+function OBJColumns(data, feed) {
+  if (this.isLast()) {
+    feed.close();
+    return;
+  }
+
+  const obj = {};
+  Object.keys(data).sort((x, y) => x.localeCompare(y)).forEach(key => {
+    if (typeof data[key] === 'object') {
+      obj[key] = JSON.stringify(data[key]);
+    } else {
+      obj[key] = data[key];
+    }
+  });
+  feed.send(obj);
+}

package/lib/skos-object.js

@@ -0,0 +1,73 @@
+"use strict";
+
+Object.defineProperty(exports, "__esModule", {
+  value: true
+});
+exports.default = void 0;
+
+function Concept(data, feed) {
+  const obj = {};
+  Object.keys(data).forEach(key => {
+    const newkey = key.replace('skos$', '');
+
+    if (Array.isArray(data[key])) {
+      data[key].filter(x => x.xml$lang).forEach(item => {
+        const localkey = newkey.concat('@').concat(item.xml$lang);
+        obj[localkey] = item.$t;
+      });
+    } else if (data[key].rdf$resource && !obj[newkey]) {
+      obj[newkey] = data[key].rdf$resource;
+    } else if (data[key].rdf$resource && obj[newkey]) {
+      obj[newkey] = [obj[newkey], data[key].rdf$resource];
+    } else if (data[key].$t && data[key].xml$lang) {
+      const localkey = newkey.concat('@').concat(data[key].xml$lang);
+      obj[localkey] = data[key].$t;
+    } else if (data[key].$t && Array.isArray(obj[newkey])) {
+      obj[newkey].push(data[key].$t);
+    } else if (data[key].$t && obj[newkey]) {
+      obj[newkey] = [obj[newkey], data[key].$t];
+    } else if (data[key].$t && !obj[newkey]) {
+      obj[newkey] = data[key].$t;
+    } else if (typeof data[key] === 'object') {
+      obj[newkey] = (this.getIndex().toString(36) + Math.random().toString(36).substr(2, 5)).toUpperCase();
+      let counter = 0;
+      Object.keys(data[key]).forEach(key2 => {
+        if (typeof data[key][key2] === 'object') {
+          data[key][key2].rdf$about = obj[newkey];
+          Concept.call(this, data[key][key2], feed);
+          counter += 1;
+        }
+      });
+
+      if (counter === 0) {
+        delete obj[newkey];
+      }
+    } else {
+      obj[newkey] = data[key];
+    }
+  });
+  feed.write(obj);
+}
+
+function SKOSObject(data, feed) {
+  if (this.isLast()) {
+    feed.close();
+  } else {
+    Concept.call(this, data, feed);
+    feed.end();
+  }
+}
+/**
+ * Take `Object` generated by XMLMapping & SKOS data and
+ * create a new basic object with only keys & values
+ *
+ * @name SKOSObject
+ * @param {undefined} none
+ * @returns {Object}
+ */
+
+
+var _default = {
+  SKOSObject
+};
+exports.default = _default;

package/lib/tar-dump.js

@@ -0,0 +1,59 @@
+"use strict";
+
+Object.defineProperty(exports, "__esModule", {
+  value: true
+});
+exports.default = TARDump;
+
+var _tarStream = _interopRequireDefault(require("tar-stream"));
+
+var _zlib = require("zlib");
+
+var _lodash = _interopRequireDefault(require("lodash.merge"));
+
+function _interopRequireDefault(obj) { return obj && obj.__esModule ? obj : { default: obj }; }
+
+const eol = '\n';
+/**
+ * Take all recevied objects and build a tar file
+ *
+ * ```json
+ * {
+ * }
+ * ```
+ *
+ * @name TARDump
+ * @param {String} [manifest] Location path to store files in the tarball
+ * @param {String} [location=data] Location path to store files in the tarball
+ * @param {String} [json=true] Parse as JSON the content of each file
+ * @param {Boolean} [compress=false] Enable gzip compression
+ */
+
+function TARDump(data, feed) {
+  if (!this.pack) {
+    this.pack = _tarStream.default.pack();
+    const compress = this.getParam('compress', false);
+    const d = new Date();
+    const manifest = (0, _lodash.default)({
+      created: d.toUTCString()
+    }, [].concat(this.getParam('manifest', [])).filter(Boolean));
+    this.pack.entry({
+      name: 'manifest.json'
+    }, JSON.stringify(manifest, null, '  '));
+    const stream = compress ? this.pack.pipe((0, _zlib.createGzip)()) : this.pack;
+    feed.flow(stream).finally(() => feed.close());
+  }
+
+  if (this.isLast()) {
+    this.pack.finalize();
+    return;
+  }
+
+  const id = this.getIndex().toString().padStart(10, '0');
+  const value = JSON.stringify(data).concat(eol);
+  const location = this.getParam('location', 'data');
+  this.pack.entry({
+    name: `${location}/f${id}.json`
+  }, value);
+  feed.end();
+}

package/lib/url-pager.js

@@ -0,0 +1,179 @@
+"use strict";
+
+Object.defineProperty(exports, "__esModule", {
+  value: true
+});
+exports.default = URLPager;
+
+var _debug = _interopRequireDefault(require("debug"));
+
+var _url = require("url");
+
+var _nodeAbortController = _interopRequireDefault(require("node-abort-controller"));
+
+var _lodash = _interopRequireDefault(require("lodash.get"));
+
+var _parseHeaders = _interopRequireDefault(require("parse-headers"));
+
+var _asyncRetry = _interopRequireDefault(require("async-retry"));
+
+var _request = _interopRequireDefault(require("./request"));
+
+function _interopRequireDefault(obj) { return obj && obj.__esModule ? obj : { default: obj }; }
+
+/**
+ * Take `Object` as parameters of URL, throw each chunk from the result
+ *
+ *
+ * Input:
+ *
+ * ```json
+ * [{"q": "a"}]
+ * ```
+ *
+ * Script:
+ *
+ * ```ini
+ * [URLPager]
+ * url = https://api.search.net
+ * path = total
+ * ```
+ *
+ * Output:
+ *
+ * ```json
+ * [
+ *      {
+ *          "q": "a",
+ *          "total": 22
+ *          "offset": 0,
+ *          "pageNumber": 1,
+ *          "totalPages", 3,
+ *          "maxPages": 1000,
+ *          "limit": 10
+ *      },
+ *      {
+ *          "q": "a",
+ *          "total": 22
+ *          "offset": 10,
+ *          "pageNumber": 2,
+ *          "totalPages", 3,
+ *          "maxPages": 1000,
+ *          "limit": 10
+ *      },
+ *      {
+ *          "q": "a",
+ *          "total": 22
+ *          "offset": 20,
+ *          "pageNumber": 3,
+ *          "totalPages", 3,
+ *          "maxPages": 1000,
+ *          "limit": 10
+ *      }
+ *  ]
+ * ```
+ *
+ * #### Example with URLs
+ *
+ * Input:
+ *
+ * ```json
+ * [
+ *   "https://httpbin.org/get?a=a",
+ *   "https://httpbin.org/get?a=b",
+ *   "https://httpbin.org/get?a=c"
+ * ]
+ * ```
+ *
+ * Script:
+ *
+ * ```ini
+ * [URLPager]
+ * path = .args
+ * ```
+ *
+ * Output:
+ *
+ * ```json
+ * [{"a": "a"}, {"a": "b"}, {"a": "c" }]
+ * ```
+ *
+ * @name URLPager
+ * @param {String} [url] URL to fetch (by default input string is taken)
+ * @param {String} [path=total] choose the path to find the number of result
+ * @param {Number} [timeout=1000] Timeout in milliseconds
+ * @param {Boolean} [noerror=false] Ignore all errors, the target field will remain undefined
+ * @param {Number} [retries=5] The maximum amount of times to retry the connection
+ * @returns {Object}
+ */
+async function URLPager(data, feed) {
+  if (this.isLast()) {
+    return feed.close();
+  }
+
+  const url = this.getParam('url');
+  const path = this.getParam('path', 'total');
+  const limit = Number(this.getParam('limit', 10));
+  const maxPages = Number(this.getParam('maxPages', 1000));
+  const retries = Number(this.getParam('retries', 5));
+  const noerror = Boolean(this.getParam('noerror', false));
+  const timeout = Number(this.getParam('timeout')) || 1000;
+  const headers = (0, _parseHeaders.default)([].concat(this.getParam('header')).filter(Boolean).join('\n'));
+  const cURL = new _url.URL(url || data);
+  const controller = new _nodeAbortController.default();
+  const parameters = {
+    timeout,
+    headers,
+    signal: controller.signal
+  };
+  const options = {
+    retries
+  };
+  cURL.search = new _url.URLSearchParams(data);
+
+  const onError = e => {
+    controller.abort();
+
+    if (noerror) {
+      (0, _debug.default)('ezs')(`Ignore item #${this.getIndex()} [URLPager] <${e}>`);
+      return feed.send(data);
+    }
+
+    (0, _debug.default)('ezs')(`Break item #${this.getIndex()} [URLPager] <${e}>`);
+    return feed.send(e);
+  };
+
+  try {
+    const response = await (0, _asyncRetry.default)((0, _request.default)(cURL.href, parameters), options);
+    const json = await response.json();
+    const total = (0, _lodash.default)(json, path);
+
+    if (total === 0) {
+      return onError(new Error('No result.'));
+    }
+
+    if (total === undefined) {
+      return onError(new Error('Unexpected response.'));
+    }
+
+    let totalPages = Math.ceil(json.total / limit);
+
+    if (totalPages > maxPages) {
+      totalPages = maxPages;
+    }
+
+    for (let pageNumber = 1; pageNumber <= totalPages; pageNumber += 1) {
+      feed.write({ ...data,
+        offset: (pageNumber - 1) * limit,
+        pageNumber,
+        totalPages,
+        maxPages,
+        limit
+      });
+    }
+
+    feed.end();
+  } catch (e) {
+    onError(e);
+  }
+}

package/lib/utils.js

@@ -0,0 +1,18 @@
+"use strict";
+
+Object.defineProperty(exports, "__esModule", {
+  value: true
+});
+exports.default = exports.writeTo = void 0;
+
+const writeTo = function writeTo(stream, data, cb) {
+  if (!stream.write(data)) {
+    stream.once('drain', cb);
+  } else {
+    process.nextTick(cb);
+  }
+};
+
+exports.writeTo = writeTo;
+var _default = writeTo;
+exports.default = _default;

package/package.json

@@ -1,7 +1,7 @@
 {
   "name": "@ezs/basics",
   "description": "Basics statements for EZS",
-  "version": "2.3.0",
+  "version": "2.4.0",
   "author": "Nicolas Thouvenin <nthouvenin@gmail.com>",
   "bugs": "https://github.com/Inist-CNRS/ezs/issues",
   "dependencies": {
@@ -20,15 +20,16 @@
     "lodash.get": "4.4.2",
     "lodash.mapkeys": "4.6.0",
     "lodash.mapvalues": "4.6.0",
+    "lodash.merge": "4.6.2",
     "lodash.set": "4.3.2",
     "lodash.zipobject": "4.1.3",
-    "make-dir": "3.1.0",
+    "make-dir": "4.0.0",
     "micromatch": "4.0.4",
     "node-abort-controller": "1.1.0",
     "parse-headers": "2.0.4",
     "path-exists": "4.0.0",
     "stream-write": "1.0.1",
-    "tar-stream": "3.1.4",
+    "tar-stream": "3.1.6",
     "tmp-filepath": "2.0.0",
     "unzipper": "0.10.11",
     "xml-mapping": "1.7.2",
@@ -41,7 +42,7 @@
   "directories": {
     "test": "test"
   },
-  "gitHead": "d363643a24747f3fd2230deb8d3eb68aa313369d",
+  "gitHead": "313d6ef24c4b958533685a198a090b7ecd2b33b9",
   "homepage": "https://github.com/Inist-CNRS/ezs/tree/master/packages/basics#readme",
   "keywords": [
     "ezs"