// ==UserScript==
// @name         DeviantArt - media download helper
// @namespace    deviantart_mediadownloadhelper
// @version      2.1
// @license      GNU AGPLv3
// @description  DeviantArt - Download multi-image deviations. Download largest resampled media file (useful if no original download is available). Fix timeout of original image download. Download filenames are prefixed with artist's name.
// @author       marp
// @homepageURL  https://greasyfork.org/en/users/204542-marp
// @grant        GM_xmlhttpRequest
// @grant        GM_download
// @connect      www.deviantart.com
// @connect      wixmp.com
// @match        https://www.deviantart.com/*/art/*
// @run-at document-end
// ==/UserScript==
// jshint esversion:11
(function() {
'use strict';
// Download symbol, downwards arrow with one horizontal lines below arrow
// the horizontal line is slightly wider and much thicker than original DA download symbol
const extraDownload1Svg = '<svg viewBox="0 0 24 24" xmlns="http://www.w3.org/2000/svg">' +
'<path d="m 19.913,18 c 1.333333,0 1.333333,5 0,5 h -16 c -1.3333328,0 -1.3333328,-5 0,-5 z ' +
'm -4,-17 c 0.507356,-4.4041e-4 0.934639,0.379128 0.994,0.883 L 16.913,2 v 6 h 3.414 ' +
'c 0.853894,-5.62e-5 1.315031,1.0010965 0.76,1.65 l -0.088,0.09 -8.413,7.648 ' +
'c -0.345621,0.314527 -0.862993,0.347745 -1.246,0.08 L 11.24,17.388 2.828,9.74 ' +
'C 2.1934191,9.1629397 2.5321441,8.1071654 3.384,8.007 L 3.5,8 H 6.913 V 2 ' +
'C 6.9131264,1.4926858 7.2931105,1.0658484 7.797,1.007 L 7.913,1 Z ' +
'm -1,2 h -6 v 7 H 6.086 L 11.913,15.297 17.74,10 H 14.914 V 3 Z" ' +
'fill-rule="evenodd"></path></svg>';
// Download symbol, downwards arrow with two horizontal lines below arrow
// the two horizontal lines are slightly wider than original DA download symbol
const extraDownload2Svg = '<svg viewBox="0 0 24 24" xmlns="http://www.w3.org/2000/svg">' +
'<path d="m 19.913,21 c 1.333333,0 1.333333,2 0,2 h -16 c -1.3333328,0 -1.3333328,-2 0,-2 z ' +
'm 0,-3 c 1.333333,0 1.333333,2 0,2 h -16 c -1.3333328,0 -1.3333328,-2 0,-2 z ' +
'm -4,-17 c 0.507356,-4.4041e-4 0.934639,0.379128 0.994,0.883 L 16.913,2 v 6 h 3.414 ' +
'c 0.853894,-5.62e-5 1.315031,1.0010965 0.76,1.65 l -0.088,0.09 -8.413,7.648 ' +
'c -0.345621,0.314527 -0.862993,0.347745 -1.246,0.08 L 11.24,17.388 2.828,9.74 ' +
'C 2.1934191,9.1629397 2.5321441,8.1071654 3.384,8.007 L 3.5,8 H 6.913 V 2 ' +
'C 6.9131264,1.4926858 7.2931105,1.0658484 7.797,1.007 L 7.913,1 Z ' +
'm -1,2 h -6 v 7 H 6.086 L 11.913,15.297 17.74,10 H 14.914 V 3 Z" ' +
'fill-rule="evenodd"></path></svg>';
const logPrefix = "TamperMonkey userscript " + GM_info.script.name + ": ";
// This recreates "__INITIAL_STATE__", "__INITIAL_I18N__" and few other JSON objects that normally exist on the global window object.
// The "initials" variable acts as container, replacing the default global window object.
// The global window object is not accessed nor changed in any way.
var initials = getInitials(document);
const i18nKeyFreeDownload = "actionbar.freeDownload";
const i18nKeyPrivateColl = "actionbar.collectPrivately.label";
const i18nKeyMoreActions = "deviationview.overlay.moreActions";
const ariaFreeDownload = initials.__INITIAL_I18N__.resources[i18nKeyFreeDownload]; // for "en": "Free download"
const ariaPrivateColl = initials.__INITIAL_I18N__.resources[i18nKeyPrivateColl]; // for "en": "Add to Private Collection"
const ariaMoreActions = initials.__INITIAL_I18N__.resources[i18nKeyMoreActions]; // for "en": "More Actions"
const idExtraOriginalDownload = GM_info.script.namespace + "_extraOriginalDownload";
const ariaExtraOriginalDownload = "download original media";
const idExtraResampledDownload = GM_info.script.namespace + "_extraResampledDownload";
const ariaExtraResampledDownload = "download resampled media";
var deviationId;
var deviationJson;
var deviationExtendedJson;
var deviationAuthorJson;
var deviationOriginalInfos; // array of structs created by parseInitials and sub-functions
var deviationResampledInfos; // array of structs created by parseInitials and sub-functions
var numDeviations; // size of array
var deviationsShown; // array of indices of deviations shown (length = 1 except for MultiImage deviations in "Scroll view" mode)
var linkOriginalDownload; // anchor ("a") element with the sub-hierarchy created by insertExtraDownloadButton
var linkResampledDownload; // anchor ("a") element with the sub-hierarchy created by insertExtraDownloadButton
var observerMain;
var observerMultiImage;
var multinode;
// Mamespace resolver for use with document.evaluate to search for "svg" tags
// Must use "svg:svg" in XPath to locate "svg" tags in document
function nsResolver(prefix) {
if (prefix === 'svg') {
return 'http://www.w3.org/2000/svg';
} else {
return null;
}
}
// Creates a new html element using an HTML source code snippet.
// This approach via a "template" element requires a new-ish browser.
function createElementFromHtml(aDocument, html) {
const template = aDocument.createElement('template');
template.innerHTML = html;
return template.content.firstElementChild;
}
// This ASYNC method returns a promise to retrieve the supplied URL via HTTP GET.
//
async function urlGetPromise(aUrl, aResponseType) {
return new Promise((resolve, reject) => {
GM_xmlhttpRequest({
method: 'GET',
url: aUrl,
responseType: (aResponseType) ? aResponseType : undefined,
onload: function(response) {
if ((response.readyState >= 2) && (response.status == 200)) {
resolve(response);
} else {
reject(response);
}
},
ontimeout: function(response) {
reject(response);
},
onerror: function(response) {
reject(response);
}
});
});
}
// Downloads a file from aUrl and triggers browser download save with filename aFilename
// Performs the worse the larger the file (downloads entire file before triggering browser download/SaveAs dialog)
function blobDownload(aDocument, aUrl, aFilename) {
// internally download the media file as a blob (worse/slower the larger the file is -> !DELAY!)
urlGetPromise(aUrl, "blob").then(
// SUCCESS -> DONE, trigger browser download
(result) => {
const tmpAnchor = aDocument.createElement('a');
const blob = new Blob([result.response], { type : 'application/octet-stream' });
const blobUrl = URL.createObjectURL(blob);
tmpAnchor.href = blobUrl;
tmpAnchor.download = (aFilename?.length > 0) ? aFilename : "";
tmpAnchor.click();
URL.revokeObjectURL(blobUrl);
},
// FAILURE -> log error
(errorresult) => {
console.error(logPrefix + "blobDownload - blob download of media file failed.", { cause: errorresult });
}
);
}
//
//
function eventHandlerClickDownload(aEvent) {
const el = aEvent.currentTarget;
const doc = (el.ownerDocument) ? el.ownerDocument : document;
// simple sanity-check if this is triggered on the correct element
if (el?.tagName == "A" && el?.href == "javascript:;" && el?.getAttribute("num") >= 1) {
aEvent.preventDefault(); // never let the link click be handled by the browser itself
const num = el.getAttribute("num");
for (var i=0; (i<num); i++) {
const download = el.getAttribute("download"+i);
const href = el.getAttribute("href"+i);
const timeout = el.getAttribute("timeout"+i);
// if time-out (plus 30sec buffer time) has elapsed -> refresh href url
// Such a timeout href URL is currently only offered by DA for the original download of the first/main deviation
// -> refresh mechanism is hard-coded here to extract exactly only this one URL from reloaded deviation page
// -> If there's ever more than one line to be refreshed like this -> then this most be re-designed ENTIRELY (use of multiple promises/async)
if (timeout && ( (Number.parseInt(timeout) - 30) < (Date.now() / 1000)) ) {
const refreshinddex = i; // use const for async, using i directly is unsafe
urlGetPromise(doc.URL).then(
// SUCCESS -> DONE, now extract new url with new timeout from document HTML source
(result) => {
const rt = result.responseText;
// the whole deviation page contains only one direct download link -> find it
var start = rt.indexOf('href="https://www.deviantart.com/download/');
if (start > 0) {
start = start + 6; // skip href="
const end = rt.indexOf('"', start); // find closing "
// rt is HTML-stringified - here, this only means "&" -> "&amp;"
const newhref = rt.substring(start, end).replaceAll("&amp;", "&");
el.setAttribute("href"+refreshinddex, newhref);
el.setAttribute("timeout"+refreshinddex, Number.parseInt(newhref.substring(newhref.indexOf("&ts=")+4)));
el.click(); // click and trigger this event handler again (NOTE: this is within an async method from urlGetPromise)
// -> This is one part that needs to be REDESIGNED if ever more than one link to refresh
} else {
console.error(logPrefix + "triggerDownload - old URL has timed out, could not find new download url in document HTML");
}
},
// FAILURE -> log error
(errorresult) => {
console.error(logPrefix + "triggerDownload - old URL has timed out, download of document with new URL failed.", { cause: errorresult });
}
);
return; // finish event, do nothing else, ABORT FOR LOOP. After refresh promise is done, a new link click will be executed (see el.click() above)
}
//  -> if not same-origin, the suggested filename is ignored
if (GM_info?.downloadMode == "native") {
// TamperMonkey with download mode "native", or ViolentMonkey (which has only native mode)
console.info(logPrefix + "eventHandlerClickDownload - using (hopefully) more efficient GM_download method");
GM_download(href, (download?.length > 0) ? download : undefined);
}
else {
// TamperMonkey with download mode "Browser API" or "disabled"
// GM_download doesn't help in this case -> ugly workaround via Blob download (can cause delays)
console.info(logPrefix + "eventHandlerClickDownload - using inefficient blob download method");
// internally download the media file as a blob (worse the larger the file is -> !DELAY!)
blobDownload(doc, href, download);
}
} // for
} // if (sanity check)
}
// Creates a new sub-hierarchy with a download link ("a" element) and an inline SVG as download symbol.
// It creates this new hierarchy based on an existing template structure.
// The existing hierarchy must start with a "span" and must contain a "button" element with a unique "aria-label" and an SVG symbol
// The "button" element is replaced by an anchor "a" element (but keeping all children of the "button", including the SVG.
// RETURNS the "a" element, i.e. *NOT* the root of the newly created hierarchy.
function insertExtraDownloadButton(aDocument, aAriaLabelLocator, aId, aAria, aSvg, aInfos) {
var template = aDocument.evaluate("./main//section//span[./div//button[@aria-label='" + aAriaLabelLocator + "']//svg:svg]",
aDocument.body, nsResolver, XPathResult.FIRST_ORDERED_NODE_TYPE, null).singleNodeValue;
if (template === null) {
// given aria label for existing hierarchy not found - fall back to secondary aria label
template = aDocument.evaluate("./main//section//span[./div//button[@aria-label='" + ariaMoreActions + "']//svg:svg]",
aDocument.body, nsResolver, XPathResult.FIRST_ORDERED_NODE_TYPE, null).singleNodeValue;
if (template === null) {
// secondary aria label also not found -> error
throw new Error(logPrefix + "insertExtraDownloadButton - template structure not found by aria-label: " + aAriaLabelLocator);
}
}
// use existing sub-hierarchy containing button with specified aria-label as template for our own, i.e.: clone entire structure
const clone = template.cloneNode(true);
clone.id = aId + "_top";
// check if this already exists -> delete old structure, if yes
const old = aDocument.getElementById(clone.id);
if (old) old.remove();
// old link should be deleted as child of overall old - but better safe than sorry...
const oldlink = aDocument.getElementById(aId);
if (oldlink) oldlink.remove();
// the "button" from the template needs to be replaced with an "a" anchor
const buttons = clone.getElementsByTagName("button");
// there must be only one single button in sub-hierarchy identified by aAriaLabelLocator
if (buttons.length !== 1) {
throw new Error(logPrefix + "insertExtraDownloadButton - template structure does not contain single button element", { cause: template });
}
const button = buttons[0];
// prepare to replace "button" element with "a" element (but keeping/bringing along its children)
const newLink = aDocument.createElement("a");
newLink.className = button.className;
newLink.ariaLabel = aAria;
newLink.title = (aInfos.length > 1) ? "Download multiple media files" : (aInfos[0].description) ? aInfos[0].description : aAria;
newLink.id = aId;
newLink.target = "_blank";
newLink.href = "javascript:;";
for (var i=0, j=0; (i<aInfos.length); i++) {
if (!(aInfos[i].url?.length > 0)) continue;
newLink.setAttribute("href"+j, aInfos[i].url);
newLink.setAttribute("download"+j, (aInfos[i].filename) ? aInfos[i].filename : "");
if (aInfos[i].timeout > 0) {
newLink.setAttribute("timeout"+j, aInfos[i].timeout);
}
j++;
}
newLink.setAttribute("num", j);
// event handler for downloading -> the anchor link itself ignores suggested download name (it's not same-origin)
newLink.addEventListener("click", eventHandlerClickDownload);
// there is always at least one child (the svg element)
while (button.hasChildNodes()) {
newLink.appendChild(button.firstChild);
}
// replace the "button" with new "a" element
button.parentNode.replaceChild(newLink, button);
// replace the "svg" element with our own
const svgs = clone.getElementsByTagName("svg");
if (svgs.length !== 1) {
throw new Error(logPrefix + "insertExtraDownloadButton - template structure does not contain single button element", { cause: template });
}
const newSvg = createElementFromHtml(aDocument, aSvg);
svgs[0].parentNode.replaceChild(newSvg, svgs[0]);
// insert the newly created sub-hierarchy into the DOM
template.before(clone);
return newLink;
}
// Returns the "span" element that forms the beginning of the DA original "Free Download" link button.
// Returns "null" if the button is not found (e.g.: if the deviation has download of original image file disabled).
function getFreeDownloadSpan(aDocument) {
return aDocument.evaluate("./main//section//span[./div//a[@aria-label='" + ariaFreeDownload +
"' and contains(@href,'&ts=') and contains(@href,'www.deviantart.com/download')]//svg:svg]",
aDocument.body, nsResolver, XPathResult.FIRST_ORDERED_NODE_TYPE, null).singleNodeValue;
}
// DA UI uses react, and common way to initialize React with data is to use a JSON "window.__INITIAL_STATE__" global object.
// It contains info this script needs - however, this object is deleted right after it is initially used by React.
// To get at the contained data, we thus need to recreate it, i.e. rerun the inline script that originally creates it.
// However, we don't want to contaminate the page's global scope, so we must rerun the script with a changed scope.
// Here, we use a local target "window" var, and we enable strict mode to prevent any other vars from leaking into global scope.
// This also return "window.__INITIAL_I18N__" - which does exist in the page's global scope window element.
// However, accessing the page's window from a userscript is unsafe, so better via this method, in same manner like "window.__INITIAL_STATE__".
function getInitials(aDocument) {
const scriptnode = aDocument.evaluate("//script[contains(.,'window.__INITIAL_STATE__ = JSON.parse(') and contains(.,'window.__URL_CONFIG__ = JSON.parse(')]",
aDocument, null, XPathResult.FIRST_ORDERED_NODE_TYPE, null).singleNodeValue;
if (scriptnode === null) return null;
// create and run the script code as a function body.
// pass in a fake "window" object as a parameter to redirect from the originally intended global scope window object to our own Object.
const windowHolder = {};
Function("window", "'use strict'; " + scriptnode.textContent)(windowHolder);
if (Object.hasOwn(windowHolder, "__INITIAL_STATE__") && Object.hasOwn(windowHolder, "__INITIAL_I18N__")) {
return windowHolder;
} else {
throw new Error (logPrefix + "getInitials - Unable to extract __INITIAL_STATE__ or __INITIAL_I18N__ JSON objects", { cause: scriptnode });
}
}
function getSizeText(sizeInBytes) {
if (sizeInBytes == null) {
return "";
}
if (sizeInBytes >= 1048576) {
return (sizeInBytes / 1048576).toFixed(1) + " MB";
}
else if (sizeInBytes >= 1024) {
return (sizeInBytes / 1024).toFixed(0) + " KB";
}
else {
return sizeInBytes.toFixed(0) + " B";
}
}
// Determines the best resampled image (mediatype with "c" field).
// Non-resampled media (original, video, etc.) are ignored (items without "c" field")
// IMPORTANT: The WIX image resampling converts animated images (GIF, etc) into static JPEGs :-(
// NOTE: this function "pre-selects" the media type "preview" over other types with equal height/resolution
//       The reason is to derive a nicer proposed filename: "...-preview.jpg" instead of "...-300w-2x.jpg"
//   aMediaTokens can be undefined (e.g.: video deviations use no tokens at all, not even for WIX transformations)
// returns an object struct with members:
//   keyString   - a string that is unique to this media, allowing to identify it when analyzing HTML snippet (used for MultiImage deviations)
//   url         - the download url
//   filename    - proposed filename (or empty string) - to be used with "download" attrib of anchor ("a" ) element
//   description - proposed description for display as hover pop with "title" attrib
//   original    - always FALSE for this function - boolean indicating if this is the original (non-resampled) media file
function findBestResampledImageInfo(aMediaTypes, aMediaBaseUri, aPrettyName, aMediaTokens) {
if (!aMediaBaseUri || !aPrettyName || !aMediaTypes) {
throw new Error(logPrefix + "findBestResampledImageInfo - some required parameters are not set");
}
if (!(aMediaTypes.length > 0)) {
throw new Error(logPrefix + "findBestResampledImageInfo - aMediaTypes is empty or wrong type");
}
var bestHeight = -1;
var bestWidth = -1;
var bestC = null; // WIX image transformation/resampling URI path (like "/v1/fill/w_1063,h_752,q_70,strp/<prettyName>-pre.jpg")
var bestT = null; // name/type of media ("fullview", "preview", etc.)
var bestR = -1; // URL protection token
// Media type with "c" field uses a WIX image transformation URL -> and the proposed filename pattern is part of the URL.
// This pattern is extracted - as browser-based download won't honor non-same-origin "download" attribute
// Instead, we need to process the pattern (replace "<prettyName>" placeholder with actual name).
// File size information in "f" attrib is *INCORRECT* (it's the size of the original file, not the transformed one).
// ADDITIONALLY:
//   There can be a larger transformation image nested inside this type (typically "-2x", within "ss" array).
//   However, this is not always the case: (a) nested "ss" array might not be there, or (b) nested inside might be exact same size.
//   these "sub-types" use the same access token as the "outer" media type (token index in attrib "r")
for (var i = 0; (i < aMediaTypes.length); i++) {
const mediatype = aMediaTypes[i];
var r = mediatype.r; // URL protection token
if (!(r >=0)) continue; // skip if not protected by a token (like static placeholder or error image)
const t = mediatype.t; // name/type of media ("fullview", "preview", etc.)
var c = mediatype.c; // WIX image transformation/resampling URI path (like "/v1/fill/w_1063,h_752,q_70,strp/<prettyName>-pre.jpg")
if (c === undefined) continue; // skip if this is not a resampled media
if (c.includes("/v1/crop/") || c.includes("/v1/fit/")) continue; // skip types that crop image or reduce in size
var h = mediatype.h; // height
if ((h > bestHeight) || ((h === bestHeight) && ((t == "preview") || (t == "fullview")))) {
bestHeight = h;
bestWidth = mediatype.w;
bestC = c;
bestT = t;
bestR = r;
}
var ss = mediatype.ss; //check for additional media variants (like "/v1/fill/w_1063,h_752,q_70,strp/<prettyName>-pre-2x.jpg")
for (var j = 0; (j < ss?.length); j++) {
const mediatypess = ss[j];
h = mediatypess.h;
if (h > bestHeight) {
bestHeight = h;
bestWidth = mediatypess.w;
bestC = mediatypess.c;
bestT = t;
bestR = r;
}
}
}
// Note: image size info for resampled media is either wrong (it's the size of the original file) or missing entirely
if (bestHeight > 0) {
return {
keyString: aMediaBaseUri,
filename: bestC.substring(bestC.indexOf("<prettyName>")).replaceAll("<prettyName>", aPrettyName),
description: "" + bestWidth + "x" + bestHeight + // width and height
" - " + bestC.substring(bestC.lastIndexOf(".")+1).toUpperCase() + // file type extension
" - " + bestT + // sample t: "fullview"
" - download resampled media",
url: aMediaBaseUri + bestC.replaceAll("<prettyName>", aPrettyName) + ((bestR >= 0) ? "?token=" + aMediaTokens[bestR] : ""),
original: false
};
} else {
return { keyString: null, filename: "", description: "no resampled media found", url: "", original: false };
}
}
// Determines the best media file that is NOT a resampled image (no WIX image transformation)
// This should only be one of the following: Either an original image, or a video/document.
//   aMediaTokens can be undefined (e.g.: video deviations use no tokens at all, not even for WIX transformations)
//   aOrigHeight can be undefined if it is not known
// returns an object struct with members:
//   keyString   - a string that is unique to this media, allowing to identify it when analyzing HTML snippet (used for MultiImage deviations)
//   url         - the download url
//   filename    - proposed filename (or empty string) - to be used with "download" attrib of anchor ("a" ) element
//   description - proposed description for display as hover pop with "title" attrib
//   original    - boolean indicating if this is the original (non-resampled) media file (videos need aOrigHeight to determine)
function findBestOtherInfo(aMediaTypes, aMediaBaseUri, aPrettyName, aIsVideo, aIsDoc, aFileType, aMediaTokens, aOrigHeight) {
if (!aMediaBaseUri || !aPrettyName || !aMediaTypes) {
throw new Error(logPrefix + "findBestOtherInfo - some required parameters are not set");
}
if (!(aMediaTypes.length > 0)) {
throw new Error(logPrefix + "findBestOtherInfo - aMediaTypes is empty or wrong type");
}
var bestHeight = -1;
var bestWidth = -1;
var bestB = null; // video URL (if exist)
var bestS = null; // document URL (if exist)
var bestT = null; // name/type of media ("fullview", "preview", etc.)
var bestR = -1; // URL protection token
var bestF = -1; // file size
for (var i = 0; (i < aMediaTypes.length); i++) {
const mediatype = aMediaTypes[i];
if (mediatype.c) continue; // ignore WIX image transformations
if (!aIsDoc && mediatype.s) continue; // ignore static images (like error images)
if (aIsVideo && (mediatype.t != "video")) continue; // only use "video" types
if (aIsDoc && (mediatype.t != aFileType)) continue; // only use document types
const h = mediatype.h; // height
if (h > bestHeight) {
bestHeight = h;
bestWidth = mediatype.w;
bestB = mediatype.b;
bestS = mediatype.s;
bestT = mediatype.t;
bestR = mediatype.r;
bestF = mediatype.f;
}
}
if (bestHeight >= 0) {
if (bestB) {
// this media type uses a URL independent of the baseUri but specific to the deviation - this is used for videos
const filetype = bestB.substring(bestB.lastIndexOf(".")+1);
// !! ASSUMPTION !!
//     => If height here is same a height of orig file, then we have original video here, as well.
// !! ASSUMPTION !!
const orig = (aOrigHeight > 0) ? ((bestHeight == aOrigHeight) ? "original " : "resampled ") : "";
return {
keyString: aMediaBaseUri, // not using bestB - there will always be a video thumbnail using aMediaBaseUri - but likely NOT always the URI of the largest video variant
filename: aPrettyName + "_" + bestHeight + "p." + filetype,
description: "" + bestWidth + "x" + bestHeight +
((bestF > 0) ? " (" + getSizeText(Number.parseInt(bestF)) + ") - " : " - ") +
filetype.toUpperCase() + " - " +
bestT + " - download " + orig + "media file",
url: bestB + ((bestR >= 0) ? "?token=" + aMediaTokens[bestR] : ""),
original: (bestHeight == aOrigHeight)
};
}
else if (aIsDoc && bestS) {
// this media type uses a static URL independent of the baseUri - this is used for documents
return {
keyString: aMediaBaseUri, // not using bestS - there will always be a document thumbnail using aMediaBaseUri
filename: aPrettyName + "." + aFileType,
description: "" + ((bestF > 0) ? " (" + getSizeText(Number.parseInt(bestF)) + ") - " : " - ") +
aFileType.toUpperCase() + " - " +
bestT + " - download original file",
url: bestS,
original: true
};
}
else
{
// this media type does not use WIX image transformations, and it is not a static URL (like used for "social_preview" media type)
// this means it is the original full size media, without any transformation - and with a URL that does NOT time-out quickly :-)
// but the URL will be with a cryptic GUID-like filename - so we must suggest better filename
const filetype = aMediaBaseUri.substring(aMediaBaseUri.lastIndexOf(".")+1);
return {
keyString: aMediaBaseUri,
filename: aPrettyName + "." + filetype,
description: "" + bestWidth + "x" + bestHeight +
((bestF > 0) ? " (" + getSizeText(Number.parseInt(bestF)) + ") - " : " - ") +
filetype.toUpperCase() + " - " +
bestT + " - download original media file",
url: aMediaBaseUri + ((bestR >= 0) ? "?token=" + aMediaTokens[bestR] : ""),
original: true
};
}
}
else
{
return { keyString: null, filename: "", description: "no non-resampled media found", url: "", original: false };
}
}
// Parses the JSON object "initials.__INITIAL_STATE__" from the "initials" parameter as created by "getInitials()".
// Nearly all page information that this script needs, comes from __INITIAL_STATE__ -> thus this function is rather large...
function parseInitials(initials) {
'use strict'; // make double sure that this function uses strict mode (strict mode on should be inherited, anyway).
try {
const entities = initials.__INITIAL_STATE__["@@entities"];
console.info("TEST-entities: ", entities);
var tmpVal;
// extra info about deviation author
tmpVal = Object.keys(entities.user).at(0); // has only one child, with node name being the authors ID
deviationAuthorJson = entities.user[tmpVal];
const authorId = deviationAuthorJson.userId;
console.assert(authorId == tmpVal, logPrefix + "entities.user JSON key ("+tmpVal+") and userId ("+authorId+") should be identical", entities);
const authorName = deviationAuthorJson.username;
// extract info about deviation basics
tmpVal = Object.keys(entities.deviation).at(0); // has only one child, with node name being the authors ID
deviationJson = entities.deviation[tmpVal];
deviationId = deviationJson.deviationId;
console.assert(deviationId == tmpVal, logPrefix + "entities.deviation JSON key ("+tmpVal+") and deviationId ("+deviationId+") should be identical", entities);
tmpVal = deviationJson.author;
console.assert(authorId == tmpVal, logPrefix + "user.userId ("+authorId+") and deviation.author ("+tmpVal+") should be identical", entities);
// extract extended info (info about original file)
tmpVal = Object.keys(entities.deviationExtended).at(0); // has only one child, with node name being the authors ID
deviationExtendedJson = entities.deviationExtended[tmpVal];
console.assert(deviationId == tmpVal, logPrefix + "entities.deviationExtended JSON key ("+tmpVal+") and deviationId ("+deviationId+") should be identical", entities);
const isDownloadable = deviationJson.isDownloadable;
const isAiGenerated = deviationJson.isAiGenerated;
// isLocked indicates that the deviation is inaccessible - all available images will be blurred
const isLocked = (deviationJson.tierAccess == "locked") ||
(deviationJson.premiumFolderData?.hasAccess == false);
const isVideo = deviationJson.isVideo;
const fileType = deviationJson.type;
const isDoc = !isVideo && ((fileType == "pdf") || (deviationJson.canBeMultiImage == false));
const isMultiImage = deviationJson.isMultiImage;
const deviationAdditionalMedia = (isMultiImage) ? deviationExtendedJson.additionalMedia : null;
console.assert((!isMultiImage) || (deviationAdditionalMedia?.length > 0), logPrefix + "MultiImage deviations should have additionalMedia array within deviationExtended JSON", entities);
// prep global variables
numDeviations = (deviationAdditionalMedia?.length > 0) ? deviationAdditionalMedia.length + 1 : 1;
deviationOriginalInfos = new Array(numDeviations);
deviationResampledInfos = new Array(numDeviations);
// *****
// ***** parse the FIRST/MAIN deviation media - if this is multi-image, additional media must be parsed, as well
// *****
// extract info about file name and media types
const media = deviationJson.media;
const mediaPrettyName = media.prettyName; // deviation basic filename without extension
// modify filename to my preference:
//  -> prefix author name and additional "AI" prefix if "AI generated" flag is set
const myPrettyName = authorName.toLowerCase() + ((isAiGenerated) ? "_AI_" : "_") + mediaPrettyName;
// extract info about full size and/or preview media file
const mediaBaseUri = media.baseUri; // sample: "https://images-wixmp-ed30a86b8c4ca887773594c2.wixmp.com/i/<guid>/<otherguid>.jpg"
const mediaTokens = media.token; // optional tokens to access media types - if present, baseUri alone will result in access denied error
// extract info about original file
// results in deviationOriginalInfos[0] either set or null, never undefined
const download = deviationExtendedJson.download;
if (download) {
console.assert(isDownloadable, logPrefix + "download info is available but isDownloadable is false. Ignoring.", entities);
const downloadType = download.url.match(/\.([^.?]+)\?token/i)[1]; // extract file extension from url (not the same as download.type)
deviationOriginalInfos[0] = {
keyString: mediaBaseUri,
filename: myPrettyName + "." + downloadType,
description: "" + download.width + "x" + download.height + " (" + getSizeText(Number.parseInt(download.filesize)) + ") - " +
downloadType.toUpperCase() + " - download original media file",
url: download.url,
original: true,
timeout: Number.parseInt(download.url.substring(download.url.indexOf("&ts=")+4))
};
} else {
console.assert(!isDownloadable, logPrefix + "isDownloadable is true but no download info is available. Ignoring.", entities);
deviationOriginalInfos[0] = null;
}
const origFile = deviationExtendedJson.originalFile;
const origFileSize = origFile?.filesize;
const origWidth = origFile?.width;
const origHeight = origFile?.height;
// decide which of the available downloads / media types to use for Original (if any) and for Resampled
if (!isLocked) {
if (!isVideo && !isDoc) {
deviationResampledInfos[0] = findBestResampledImageInfo(media.types, mediaBaseUri, myPrettyName, mediaTokens);
if (!deviationOriginalInfos[0]) { // no download url (the one that times out) -> try to find original in mediaTypes
const bestOther = findBestOtherInfo(media.types, mediaBaseUri, myPrettyName, isVideo, isDoc, fileType, mediaTokens, origHeight);
deviationOriginalInfos[0] = (bestOther?.original) ? bestOther : null;
}
}
else // isVideo == true
{
const bestOther = findBestOtherInfo(media.types, mediaBaseUri, myPrettyName, isVideo, isDoc, fileType, mediaTokens, origHeight);
if (bestOther?.original) {
deviationOriginalInfos[0] = bestOther;
deviationResampledInfos[0] = null;
} else {
deviationOriginalInfos[0] = null;
deviationResampledInfos[0] = bestOther;
}
}
}
else {
// nothing we can do if the deviation is locked
deviationResampledInfos[0] = null;
}
// *****
// ***** If MultiImage -> parse additional media
// *****
for (var mediaindex = 0; (mediaindex < (numDeviations-1)); mediaindex++) {
const additionalMedia = deviationAdditionalMedia[mediaindex];
const origFileSize = additionalMedia.filesize;
const origWidth = additionalMedia.width;
const origHeight = additionalMedia.height;
const media = additionalMedia.media;
const mediaPrettyName = media.prettyName; // deviation basic filename without extension
// modify filename of MultiImage items to my preference:
//  -> MultiImage deviations often seem to include the file extension as part of the pretty name - try to detect and remove
//  -> prefix the ENTIRE pretty name of the first/main deviation (this will also "inherit" the author and potential AI prefixes)
const myMultiPrettyName = myPrettyName + "_" + mediaPrettyName.replace(/\.[a-zA-Z0-9]+$/, "");
// extract info about full size and/or preview media file
const mediaBaseUri = media.baseUri; // sample: "https://images-wixmp-ed30a86b8c4ca887773594c2.wixmp.com/i/<guid>/<otherguid>.jpg"
const mediaTokens = media.token; // optional tokens to access media types - if present, baseUri alone will result in access denied error
if (!isLocked) {
if (!isVideo && !isDoc) {
deviationResampledInfos[mediaindex+1] = findBestResampledImageInfo(media.types, mediaBaseUri, myMultiPrettyName, mediaTokens);
const bestOther = findBestOtherInfo(media.types, mediaBaseUri, myMultiPrettyName, isVideo, isDoc, fileType, mediaTokens, origHeight);
deviationOriginalInfos[mediaindex+1] = (bestOther?.original) ? bestOther : null;
}
else // there should be no videos in MuiltiImage deviations - but who knows... better safe than sorry...
{
const bestOther = findBestOtherInfo(media.types, mediaBaseUri, myMultiPrettyName, isVideo, isDoc, fileType, mediaTokens, origHeight);
if (bestOther?.original) {
deviationOriginalInfos[mediaindex+1] = bestOther;
deviationResampledInfos[mediaindex+1] = null;
} else {
deviationOriginalInfos[mediaindex+1] = null;
deviationResampledInfos[mediaindex+1] = bestOther;
}
}
}
else {
// nothing we can do if the deviation is locked
deviationResampledInfos[mediaindex+1] = null;
}
}
}
catch (e) {
throw new Error(logPrefix + "parseInitials - Failure parsing __INITIAL_STATE__", { cause: e });
}
}
function determineDeviationIndex(aDocument) {
var newDeviationsShown = [];
if (!multinode || !(numDeviations > 0)) {
return newDeviationsShown;
}
if (numDeviations == 1) {
newDeviationsShown = [0];
return newDeviationsShown;
}
for (var i=0; (i < numDeviations); i++) {
// if the below "keys" occur in the target DOM, then this identifies which deviation is currently being shown
const keyResampled = deviationResampledInfos[i]?.keyString;
const keyOriginal = deviationOriginalInfos[i]?.keyString;
var matchR = null;
if (keyResampled?.length > 0) {
matchR = aDocument.evaluate(".//figure//img[not(ancestor::button) and (contains(@src,'" + keyResampled + "') or contains(@srcset,'" + keyResampled + "'))]" +
" | " +
".//div[contains(@style, '" + keyResampled + "')]", // for video thumbnail overlay: style='... background-image: url("https:/ ...'
multinode, null, XPathResult.FIRST_ORDERED_NODE_TYPE, null).singleNodeValue;
}
var matchO = null;
if (keyOriginal?.length > 0) {
matchO = aDocument.evaluate(".//figure//img[not(ancestor::button) and (contains(@src,'" + keyOriginal + "') or contains(@srcset,'" + keyOriginal + "'))]" +
" | " +
".//div[contains(@style, '" + keyOriginal + "')]", // for video thumbnail overlay: style='... background-image: url("https:/ ...'
multinode, null, XPathResult.FIRST_ORDERED_NODE_TYPE, null).singleNodeValue;
}
if (matchR || matchO) newDeviationsShown.push(i);
}
return newDeviationsShown;
}
// Inserts up to two extra download buttons, requires that __INITIAL_STATE__ has been parsed and processed
//  - button "original" will be present if media download in original, non-resampled quality is available
//  - button "resampled" will be present if a resampled media download is available
// Also checks if previously inserted buttons need to be updated (needed for MultiImage deviations)
function insertUpdateExtraButtons(aDocument) {
var newDeviationsShown = determineDeviationIndex(aDocument);
// only continue if no buttons inserted or if devinationsShown has changed
// CAREFUL: likely endless loop otherwise, as observer includes this DOM area
if (linkResampledDownload || linkOriginalDownload) {
if (deviationsShown &&
newDeviationsShown.length === deviationsShown?.length &&
newDeviationsShown.every((e, i) => e === deviationsShown[i])) {
return;
}
}
deviationsShown = newDeviationsShown;
const resampledInfos = [];
const originalInfos = [];
for (var i=0; (i<deviationsShown?.length); i++) {
const oi = deviationOriginalInfos[deviationsShown[i]];
const ri = deviationResampledInfos[deviationsShown[i]];
if (oi?.url?.length > 0) {
originalInfos.push(oi);
}
if (ri?.url?.length > 0) {
resampledInfos.push(ri);
}
}
if (linkResampledDownload) {
linkResampledDownload.remove();
linkResampledDownload = null;
}
if (resampledInfos.length > 0) {
linkResampledDownload = insertExtraDownloadButton(
aDocument, ariaPrivateColl, idExtraResampledDownload,
ariaExtraResampledDownload, extraDownload2Svg,
resampledInfos);
}
if (linkOriginalDownload) {
linkOriginalDownload.remove();
linkOriginalDownload = null;
}
if (originalInfos.length > 0) {
linkOriginalDownload = insertExtraDownloadButton(
aDocument, ariaPrivateColl, idExtraOriginalDownload,
ariaExtraOriginalDownload, extraDownload1Svg,
originalInfos);
}
}
function installMultiObserver(aDocument) {
// this observer is to deal multi image deviations -  it must be re-initialized after full-screen zoom (observer node gets removed on full-screen zoom)
observerMultiImage = new MutationObserver(function(mutations) {
// don't need to iterate over added or removed nodes!
// instead, the below function checks globally by ID (fast! simple!) if extra buttons already exist or not
insertUpdateExtraButtons(aDocument);
});
// Configuration of the observer
// look at everything below a certain div that encapsulates both the deviation and the button panel below (where the extra buttons are)
// It is removed from DOM when switching to full-screen.
// It's descendants structure is different for image deviations and video deviations
// It's descendants structure and/or descendant attributes change for MultiImage deviation when navigating images and/or when switching between single image and all image display modes
const multiconfig = { attributes: true, childList: true, characterData: false, subtree: true };
multinode = aDocument.evaluate("(./main/div[1])[preceding-sibling::header]/div[child::section]",
aDocument.body, null, XPathResult.FIRST_ORDERED_NODE_TYPE, null).singleNodeValue;
if (multinode) {
observerMultiImage.observe(multinode, multiconfig);
} else {
observerMultiImage = null;
}
}
function installObservers(aDocument) {
// create an observer instances to check if extra buttons need to be re-added or modified (override React-based re-rendering of DOM)
// this observer is to deal with zoom/full-screen display modes - it only looks at direct children of <main>
observerMain = new MutationObserver(function(mutations) {
if (!(multinode?.connected == true)) {
if (observerMultiImage) {
observerMultiImage.disconnect();
}
installMultiObserver(aDocument);
}
// don't need to iterate over added or removed nodes!
// instead, the below function checks globally by ID (fast! simple!) if extra buttons already exist or not
insertUpdateExtraButtons(aDocument);
});
// Configuration of the observer
// "main" node is the start of React managed hierarchy.
// It changes in major way when switching to full-screen.
// Fortunately, it is sufficient to just observe direct children - all sub-hierarchy elements get added/removed right along with them
const mainconfig = { attributes: false, childList: true, characterData: false, subtree: false };
const mainnode = aDocument.evaluate("./main", aDocument.body, null, XPathResult.FIRST_ORDERED_NODE_TYPE, null).singleNodeValue;
observerMain.observe(mainnode, mainconfig);
}
//
// ***** START OF USER SCRIPT *****
//
// extract all info needed for extra download button fROM __INITIAL_STATE__ object ("initials" itself is already initialized)
parseInitials(initials);
// process already loaded nodes (the initial posts before scrolling down for the first time)
insertUpdateExtraButtons(document);
// mutation observer to re-create extra download buttons if React re-renders DOM (e.g. expanding to and exiting from full-screen view).
installMultiObserver(document);
installObservers(document);
})();