aboutsummaryrefslogtreecommitdiff
path: root/src/server/ApiManagers/DownloadManager.ts
diff options
context:
space:
mode:
Diffstat (limited to 'src/server/ApiManagers/DownloadManager.ts')
-rw-r--r--src/server/ApiManagers/DownloadManager.ts267
1 files changed, 267 insertions, 0 deletions
diff --git a/src/server/ApiManagers/DownloadManager.ts b/src/server/ApiManagers/DownloadManager.ts
new file mode 100644
index 000000000..5bad46eda
--- /dev/null
+++ b/src/server/ApiManagers/DownloadManager.ts
@@ -0,0 +1,267 @@
+import ApiManager, { Registration } from "./ApiManager";
+import { Method } from "../RouteManager";
+import RouteSubscriber from "../RouteSubscriber";
+import * as Archiver from 'archiver';
+import * as express from 'express';
+import { Database } from "../database";
+import * as path from "path";
+import { DashUploadUtils, SizeSuffix } from "../DashUploadUtils";
+import { publicDirectory } from "..";
+
+export type Hierarchy = { [id: string]: string | Hierarchy };
+export type ZipMutator = (file: Archiver.Archiver) => void | Promise<void>;
+export interface DocumentElements {
+ data: string | any[];
+ title: string;
+}
+
+export default class DownloadManager extends ApiManager {
+
+ protected initialize(register: Registration): void {
+
+ /**
+ * Let's say someone's using Dash to organize images in collections.
+ * This lets them export the hierarchy they've built to their
+ * own file system in a useful format.
+ *
+ * This handler starts with a single document id (interesting only
+ * if it's that of a collection). It traverses the database, captures
+ * the nesting of only nested images or collections, writes
+ * that to a zip file and returns it to the client for download.
+ */
+ register({
+ method: Method.GET,
+ subscription: new RouteSubscriber("imageHierarchyExport").add('docId'),
+ onValidation: async ({ req, res }) => {
+ const id = req.params.docId;
+ const hierarchy: Hierarchy = {};
+ await buildHierarchyRecursive(id, hierarchy);
+ return BuildAndDispatchZip(res, zip => writeHierarchyRecursive(zip, hierarchy));
+ }
+ });
+
+ register({
+ method: Method.GET,
+ subscription: new RouteSubscriber("downloadId").add("docId"),
+ onValidation: async ({ req, res }) => {
+ return BuildAndDispatchZip(res, async zip => {
+ const { id, docs, files } = await getDocs(req.params.docId);
+ const docString = JSON.stringify({ id, docs });
+ zip.append(docString, { name: "doc.json" });
+ files.forEach(val => {
+ zip.file(publicDirectory + val, { name: val.substring(1) });
+ });
+ });
+ }
+ });
+
+ register({
+ method: Method.GET,
+ subscription: new RouteSubscriber("/serializeDoc").add("docId"),
+ onValidation: async ({ req, res }) => {
+ const { docs, files } = await getDocs(req.params.docId);
+ res.send({ docs, files: Array.from(files) });
+ }
+ });
+
+
+ }
+
+}
+
+async function getDocs(id: string) {
+ const files = new Set<string>();
+ const docs: { [id: string]: any } = {};
+ const fn = (doc: any): string[] => {
+ const id = doc.id;
+ if (typeof id === "string" && id.endsWith("Proto")) {
+ //Skip protos
+ return [];
+ }
+ const ids: string[] = [];
+ for (const key in doc.fields) {
+ if (!doc.fields.hasOwnProperty(key)) {
+ continue;
+ }
+ const field = doc.fields[key];
+ if (field === undefined || field === null) {
+ continue;
+ }
+
+ if (field.__type === "proxy" || field.__type === "prefetch_proxy") {
+ ids.push(field.fieldId);
+ } else if (field.__type === "script" || field.__type === "computed") {
+ if (field.captures) {
+ ids.push(field.captures.fieldId);
+ }
+ } else if (field.__type === "list") {
+ ids.push(...fn(field));
+ } else if (typeof field === "string") {
+ const re = /"(?:dataD|d)ocumentId"\s*:\s*"([\w\-]*)"/g;
+ let match: string[] | null;
+ while ((match = re.exec(field)) !== null) {
+ ids.push(match[1]);
+ }
+ } else if (field.__type === "RichTextField") {
+ const re = /"href"\s*:\s*"(.*?)"/g;
+ let match: string[] | null;
+ while ((match = re.exec(field.Data)) !== null) {
+ const urlString = match[1];
+ const split = new URL(urlString).pathname.split("doc/");
+ if (split.length > 1) {
+ ids.push(split[split.length - 1]);
+ }
+ }
+ const re2 = /"src"\s*:\s*"(.*?)"/g;
+ while ((match = re2.exec(field.Data)) !== null) {
+ const urlString = match[1];
+ const pathname = new URL(urlString).pathname;
+ files.add(pathname);
+ }
+ } else if (["audio", "image", "video", "pdf", "web"].includes(field.__type)) {
+ const url = new URL(field.url);
+ const pathname = url.pathname;
+ files.add(pathname);
+ }
+ }
+
+ if (doc.id) {
+ docs[doc.id] = doc;
+ }
+ return ids;
+ };
+ await Database.Instance.visit([id], fn);
+ return { id, docs, files };
+}
+
+/**
+ * This utility function factors out the process
+ * of creating a zip file and sending it back to the client
+ * by piping it into a response.
+ *
+ * Learn more about piping and readable / writable streams here!
+ * https://www.freecodecamp.org/news/node-js-streams-everything-you-need-to-know-c9141306be93/
+ *
+ * @param res the writable stream response object that will transfer the generated zip file
+ * @param mutator the callback function used to actually modify and insert information into the zip instance
+ */
+export async function BuildAndDispatchZip(res: express.Response, mutator: ZipMutator): Promise<void> {
+ res.set('Content-disposition', `attachment;`);
+ res.set('Content-Type', "application/zip");
+ const zip = Archiver('zip');
+ zip.pipe(res);
+ await mutator(zip);
+ return zip.finalize();
+}
+
+/**
+ * This function starts with a single document id as a seed,
+ * typically that of a collection, and then descends the entire tree
+ * of image or collection documents that are reachable from that seed.
+ * @param seedId the id of the root of the subtree we're trying to capture, interesting only if it's a collection
+ * @param hierarchy the data structure we're going to use to record the nesting of the collections and images as we descend
+ */
+
+/*
+Below is an example of the JSON hierarchy built from two images contained inside a collection titled 'a nested collection',
+following the general recursive structure shown immediately below
+{
+ "parent folder name":{
+ "first child's fild name":"first child's url"
+ ...
+ "nth child's fild name":"nth child's url"
+ }
+}
+{
+ "a nested collection (865c4734-c036-4d67-a588-c71bb43d1440)":{
+ "an image of a cat (ace99ffd-8ed8-4026-a5d5-a353fff57bdd).jpg":"https://upload.wikimedia.org/wikipedia/commons/thumb/3/3a/Cat03.jpg/1200px-Cat03.jpg",
+ "1*SGJw31T5Q9Zfsk24l2yirg.gif (9321cc9b-9b3e-4cb6-b99c-b7e667340f05).gif":"https://cdn-media-1.freecodecamp.org/images/1*SGJw31T5Q9Zfsk24l2yirg.gif"
+ }
+}
+*/
+async function buildHierarchyRecursive(seedId: string, hierarchy: Hierarchy): Promise<void> {
+ const { title, data } = await getData(seedId);
+ const label = `${title} (${seedId})`;
+ // is the document a collection?
+ if (Array.isArray(data)) {
+ // recurse over all documents in the collection.
+ const local: Hierarchy = {}; // create a child hierarchy for this level, which will get passed in as the parent of the recursive call
+ hierarchy[label] = local; // store it at the index in the parent, so we'll end up with a map of maps of maps
+ await Promise.all(data.map(proxy => buildHierarchyRecursive(proxy.fieldId, local)));
+ } else {
+ // now, data can only be a string, namely the url of the image
+ const filename = label + path.extname(data); // this is the file name under which the output image will be stored
+ hierarchy[filename] = data;
+ }
+}
+
+/**
+ * This is a very specific utility method to help traverse the database
+ * to parse data and titles out of images and collections alone.
+ *
+ * We don't know if the document id given to is corresponds to a view document or a data
+ * document. If it's a data document, the response from the database will have
+ * a data field. If not, call recursively on the proto, and resolve with *its* data
+ *
+ * @param targetId the id of the Dash document whose data is being requests
+ * @returns the data of the document, as well as its title
+ */
+async function getData(targetId: string): Promise<DocumentElements> {
+ return new Promise<DocumentElements>((resolve, reject) => {
+ Database.Instance.getDocument(targetId, async (result: any) => {
+ const { data, proto, title } = result.fields;
+ if (data) {
+ if (data.url) {
+ resolve({ data: data.url, title });
+ } else if (data.fields) {
+ resolve({ data: data.fields, title });
+ } else {
+ reject();
+ }
+ } else if (proto) {
+ getData(proto.fieldId).then(resolve, reject);
+ } else {
+ reject();
+ }
+ });
+ });
+}
+
+/**
+ *
+ * @param file the zip file to which we write the files
+ * @param hierarchy the data structure from which we read, defining the nesting of the documents in the zip
+ * @param prefix lets us create nested folders in the zip file by continually appending to the end
+ * of the prefix with each layer of recursion.
+ *
+ * Function Call #1 => "Dash Export"
+ * Function Call #2 => "Dash Export/a nested collection"
+ * Function Call #3 => "Dash Export/a nested collection/lowest level collection"
+ * ...
+ */
+async function writeHierarchyRecursive(file: Archiver.Archiver, hierarchy: Hierarchy, prefix = "Dash Export"): Promise<void> {
+ for (const documentTitle of Object.keys(hierarchy)) {
+ const result = hierarchy[documentTitle];
+ // base case or leaf node, we've hit a url (image)
+ if (typeof result === "string") {
+ let path: string;
+ let matches: RegExpExecArray | null;
+ if ((matches = /\:1050\/files\/(upload\_[\da-z]{32}.*)/g.exec(result)) !== null) {
+ // image already exists on our server
+ path = `${__dirname}/public/files/${matches[1]}`;
+ } else {
+ // the image doesn't already exist on our server (may have been dragged
+ // and dropped in the browser and thus hosted remotely) so we upload it
+ // to our server and point the zip file to it, so it can bundle up the bytes
+ const information = await DashUploadUtils.UploadImage(result);
+ path = information.serverAccessPaths[SizeSuffix.Original];
+ }
+ // write the file specified by the path to the directory in the
+ // zip file given by the prefix.
+ file.file(path, { name: documentTitle, prefix });
+ } else {
+ // we've hit a collection, so we have to recurse
+ await writeHierarchyRecursive(file, result, `${prefix}/${documentTitle}`);
+ }
+ }
+} \ No newline at end of file
generated by cgit v1.2.3-70-g09d2 (git 2.46.0) at 2025-10-06 19:06:13 +0000