The one small but relevant news I’ve seen is a recent mention that Illustrator can now/will soon support multi-threading:   https://blog.adobe.com/en/publish/2024/10/14/even-more-speed-precision-power-get-started-latest-illustrator-indesign-features-creative-professionals   Threading doesn’t really matter to Illustrator on desktop, where all interactions (not counting us automators) are a single human user manipulating one document at a time via its GUI, but it will be important to the headless Ai engine Adobe is putting into their Adobe Cloud.   The big problem is that Illustrator’ guts are 40 years old(!), hard-welded to a 1980s one user = one [G]UI = one active document model, so there’s a lot of ancient nasty internal plumbing that needs to be reworked or replaced to bring it into this century. (TBH they should’ve started this modernization 20 years ago but, eh, monopolies…) This isn’t to say UXP couldn’t be built atop the old core, but it’d be a hella nasty lash-up and Ai is already cranky and unstable as it is. So I don’t blame them for doing the whole job in a sensible order.   Once that blocker is out of the way, I’d expect UXP to progress quickly atop the new architecture. UXP (for its faults) powering Ai+PS+ID engines in the Adobe Cloud will be a Killer Product, and now Adobe’s woken up by fresh competition it’s pulling all the stops.   Alas, I’ve just checked out the Ai 29 SDK (what ExtendScript’s built on) and it’s the exact same gnarly 40yo API, so no new movements there. Here’s hoping for MAX ’25… ... View more

You’ve already got the short and long answers:   1. Illustrator as-is cannot represent CMYKOGV values, as Illustrator documents only support pre-digital RGB/CMYK [+optional spot] color models.   If you want to dig really deep, there is a C++ API (IAIColorSpace) that can represent additional color spaces and the `AIColor` struct that represents individual RGB/CMYK/spot/etc color values has a `kAdvanceColor` option which suggests it may be able to represent color values in other colorspaces. But Illustrator’s SDK documentation is hopelessly lacking and contains no information or examples on their usage, so I can’t even tell you if those APIs could be capable of doing what you want, never mind how to actually do it. So unless you’re an organization like Esko that pays a ton of money for a direct private line to Adobe’s developer teams, those APIs are closed to you.   2. Your whole request smells of this.   As a professional automation developer, one of my jobs is to advise my customer when the thing they’re asking for would be a very bad idea/already has a far much way of doing it which they’re maybe not aware of/is literally impossible. Of course, if someone insists on digging that hole for themselves anyway then that’s their perogative; but I certainly won’t be helping/billing them to bury themselves as that would be unethical on my part.   TL;DR: Unless you can tell us what your end goal is—e.g. getting your soft/hard proofs to match your Indigo output more accurately—we can’t help you further. And if you can tell us, the chances are the correct way to do it is not the way you’re wanting to do it. Either way, the people best placed to help you are your RIP vendors, so you should talk to them first. ... View more

[deleted as off-topic] ... View more

Hi Monica,   The key phrase there is “proper color management”. Raw CMYK numbers mean nothing except in context of an accurate ICC profile.   That being said, 8-bit CMYK is still a rotten working space. A CMYK gamut with ICC press profile is small and low-fidelity; it doesn’t tolerate adjustments and doesn’t accomodate the far larger gamut of modern CMYKOGV presses (unless you stretch it to something silly, which would be its own world of pain). Hence the creation of Pantone inks, which bypass CMYK and ICC profiles entirely. Hence Adobe RGB/Large Gamut RGB as a working space in more modern workflows (although still not ideal).   In Illustrator’s 80s-era color model, 1 Pantone color = 1 separation. This model made sense when output was traditional litho presses:   1 Pantone color = 1 separation = 1 film = 1 plate + 1 bucket of pre-mixed ink => 1 color on paper.   But Pantones are just as meaningless as CMYK values in modern 6-8 color digital printing. They are supported purely for backwards-compatibility, i.e. serving a highly conservative upstream industry that is tightly wedded to working practices which are now 30 years out of date and terrified to change.   OP’s fundamental problem is, I think, that they don’t realize just how unconnected the “CMYK” and “Pantone” numbers they read in their Illustrator files are from what finally ends up on each “plate”. (Yes, I know Indigo uses rollers, eliminating plates—my whole point is that classical litho mindset should not be applied.)   The first thing OP’s RIP software will do is take apart their artwork file and translate all of its colors into something modern and sane—presumably 16-bit Lab colorspace. From there it will generate its actual separations, relying on math and exact knowledge of substrate and inks.   All of these conversions are mediated by the device-specific ICC profiles installed. Thus, any resemblance to OP’s original raw numbers is entirely coincidental. And any resemblance to OP’s soft- and/or hard-proofs is critically dependent on whether or not their ICC device profiles are any good in the first place.   The only thing artworking in an 8-bit CMYK colorspace gives you is poorer-quality data to send to the above workflow. Unfortunately, while Photoshop supports Lab/16 documents, Illustrator still does not. (Ironic for an industry-standard platform on which over a trillion dollars of print work gets artworked every year!)   The correct working colorspace for a modern production department should be 16-bit Lab. This represents all human-visible colors as absolute values completely independent of any and all input and output devices. Install high-quality screen and press ICC profiles for on-screen representation and out-of-gamut detection, and a color editing GUI that isn’t from AI 8, and Illustrator operators could be rocking the bleeding edge of print.   We can live in hope.   But hey, I’m old enough to remember when “color management” meant bending your CRT’s RGB knobs to vaguely resemble your most recent Cromalin. 🙂   ... View more

> Anyway, you are talking about manually making this conversion, using a scheme of your own as shown in your Excel document, so I will infer that the RIP doesn't convert PANTONE colors in the way you want.   This.   That OP presents an Excel with CMYK[OV] numbers written in it rather suggests it is OP’s production process which is dysfunctional. Asking automators here how to script a hack onto a kludge onto a bodge of a process won’t make their artworks better; it will make their process worse. The only right answer is: “Don’t do that.”   I would suggest OP has an honest and forthright conversation with their RIP vendor to understand what OP actually needs for a modern proofing process. (As opposed to what OP thinks they need… based on out-of-date working practices and deficient upstream tools, which their modern RIP software already expends a fair bit of effort compensating and correcting for). ... View more

Yes, we have INDIGO cmykov press. I want to check files before RIP.   Why? To what end?   Illustrator’s color model is rooted firmly in the 1980s and has hardly changed since. Raw CMYK numbers are pretty meaningless when your modern RIP software and digital press run on ICC profiles; you should be hanging your proofing process off that.   … p.s. Serif’s Affinity supports 16-bit Lab documents out of the box, so Adobe’s failure to keep pace with modern print tech should be especially embarrassing considering how much they originally invented. Honestly, the print world sacks off CMYK sliders and Pantone books for Lab, the better. ... View more

The previous script moved the leftmost selected object to the top of the stacking order.   This keeps the leftmost selected object at its current position in the stacking order:   var sel = app.activeDocument.selection; if (sel.length < 2) { alert("Please select 2 or more items."); } else { var items = []; for (var i = 0; i < sel.length; i++) { var item = sel[i]; items.push({item: item, x: item.position[0]}); } items.sort(function(a, b) { return a.x < b.x; }); var topItem = items.pop().item; while (items.length > 0) { items.shift().item.move(topItem, ElementPlacement.PLACEAFTER); } }   ... View more

var sel = app.activeDocument.selection; var items = []; for (var i = 0; i < sel.length; i++) { var item = sel[i]; items.push({item: item, x: item.position[0]}); } items.sort(function(a, b) { return a.x < b.x; }); // or use `>` to reverse order for (var i = 0; i < items.length; i++) { items[i].item.zOrder(ZOrderMethod.BRINGTOFRONT); } ... View more

Here is @moluapple’s GitHub page:   https://github.com/moluapple   This is the last version of the AI_datasetMakeup.jsx file before it was deleted from the repository (there is no explanation as to why):   https://github.com/moluapple/ExtendScripts4AI-ID/blob/95bd64c73a9180de5abdda16cad6a56ef1e8d182/AI_datasetMakeup.jsx   This may also be useful to you:   https://community.adobe.com/t5/illustrator-discussions/import-csv-file/m-p/14427733#M397733   -- ... View more

On Mac, use AppleScript. Paste the following into a new Script Editor document (/Applications/Utilities/Script Editor.app) and save as `aitool.applescript`:   #!/usr/bin/env osascript property _name : "aitool" on run argv if argv is {} then set argv to {"-h"} set opt to item 1 of argv set argv to rest of argv if {opt} is in {"-h", "--help"} then log "Usage:" log _name & " -h" log _name & " -o /path/to/file.ai [ file ... ]" log _name & " -r /path/to/file.jsx [ argument ... ]" else if {opt} is in {"-o", "--open"} then repeat with pth in argv set f to POSIX file pth try tell application "Adobe Illustrator" to open f without dialogs on error msg number num log msg & " (" & num & ")" end try end repeat else if {opt} is in {"-r", "--run"} then set pth to item 1 of argv set argv to rest of argv set f to POSIX file pth try tell application "Adobe Illustrator" do javascript f with arguments argv end tell on error msg number num log msg & " (" & num & ")" end try else log "Unknown option: " & opt end if end run     In Terminal.app, make it executable and move it to a directory that on your shell $PATH:   chmod +x aitool.applescript sudo mv aitool.applescript /usr/local/bin/aitool     You can now run from Terminal, e.g.   aitool -r /Users/foo/myscript.jsx     Support for relative file paths, more robust error handling (e.g. file not found), etc is left as an exercise. (This may help.) ... View more

Two files in the Illustrator SDK look relevant to your interests, courtesy of a third-party repo that posted it (Adobe’s own developer page isn’t working in my Firefox):   https://github.com/WestonThayer/Bloks/blob/master/BloksAIPlugin/Vendor/illustratorapi/ate/ATESuites.h   The CharFeaturesSuite has functions for querying and setting OT features on a range of text in a text frame. (Look for the phrase “OpenType feature”.) I expect JSX’s CharacterAttributes.stylisticSets uses that under the hood.   https://github.com/WestonThayer/Bloks/blob/master/BloksAIPlugin/Vendor/illustratorapi/illustrator/AIFont.h   AIFontSuite’s functions provides a lot more font information  than JSX’s painfully limited TextFont class. (Look for “glyph set”.)   See if there’s anything there that whets your appetite. You’ll need a C/C++ developer to make use of it if you don’t know C yourself; there’s a few around these forums, typically under the SDK tag.   ... View more

It seems to be the number of files opened and closed that triggers it. Nothing you can do about that. Implement the workaround, file a bug report if you want, and move along. ... View more

AI crashes after opening and closing >100 docs? I remember dealing with that bug from more than a decade ago. (Using the AppleScript API, but the JSX API shares a lot of the plumbing.) Quitting and restarting AI after every 100 docs was the workaround; you can easily do that using AppleScript. (If your script is JSX, write an AppleScript helper app that performs the relaunch, then have your JSX run it via `File.execute`.) ... View more

The script error looks like Illustrator being a buggy jerk: e.g. `paragraphs.length` reporting 3 paragraphs when there are only 2. You can work around it like this:     var foundParagraphs = []; for (var i = 0; i < doc.stories.length; i++) { var paragraphs = doc.stories[i].paragraphs; for (var j = 0; j < paragraphs.length; j++) { try { foundParagraphs.push(paragraphs[j].contents.replace(/^\s+|\s+$/g)); } catch (e) { // bug workaround where paragraphs.length reports e.g. 3 items but there are only 2 if (e.number !== 1302) { throw e; } } } } foundParagraphs.sort();     I have to say, testing with the supplied files (which you should really anonymize before sharing, btw) the results are disappointing. That’s not really surprising to me: I was prototyping this sort of QA/revision autonomation over a decade ago and it is never as straightforward as it sounds. Pack copy documents rarely match exactly what appears on the artwork, e.g. “en_GB • Jute roller blind” ends up as “Jute roller blind”, and some fields don’t appear as text at all (e.g. picto names like “Wash care symbol - professional cleaning_do not dry clean.pdf”).   BTW, the SourceIllustrator file is not an improvement over the plain text file (which itself is not great), so I wouldn’t bother with that. If you can get the original pack copy document in XML or other structured format, that’s a better starting point as that tells you what each piece of text actually means. From there, you can selectively clean each one according to the same rules the artist follows, e.g. removing the “en_GB/fr_FR/etc •” crud, before trying to match it in the artwork.   I must emphasize: your entire ability to automate anything hangs on the quality of your inputs. If the customer’s pack copy is a sloppy mess that only a human artworker/QAer can reliably interpret, don’t even try. Maybe when AGI gets good enough, you’ll be able to throw unpredictable data at it and it’ll do as good a job as a human at making sense of it, but classical automation demands strictly defined, absolutely consistent inputs or you only dig yourself an even bigger hole.   The part of my script which extracts the artwork text would be improved by having it ignore the non-pack copy text on “cutter” and “legend” layers. My implementation (which was really just proof of concept) gets the document’s stories, ’cos that’s simplest, but it should really iterate through the individual text frames on the “artwork” layer to extract their text. Oh, and it should also check for any text box overflows while it’s doing that, as an obvious QA issue. (This still won’t guarantee the text is visible on pack, e.g. a text frame could be hidden under an image or solid fill instead of on top of it, but it’d be a very badly built artwork that did that. For that level of verification, you’ll need an OCR approach—which has its own challenges.)   Beyond that, for the text comparison step, you really want to use fuzzy text matching, e.g. fuse.js. This will identify text which is mostly, but not quite, identical. Touch-wood that should match up most of the pack copy fields to the text that appears in the artwork, at which point the fine differences (which might just be extra whitespace or fixed punctuation) can be highlighted for manual review. Find yourself a Node.js developer if you can’t write that script yourself. Interop between ExtendScript and Node scripts is a chore (for sharing data between them it’s easiest to read and write JSON files).   …   TBH, it’d be better to autonomate first artwork production, using scripts to place all the pack copy into a tear-off sheet of pre-built text frames/panels which the artworker can drop into place, or take a lead artwork and replace its original text to assist in producing the adaptations. Eliminating manual copy-paste will reduce QA as the text is already correctly formatted within Illustrator and most text frames be dropped straight into the artwork and sized and positioned to fit. Prebuilt text frames/panels can also be tagged so that QA scripts can easily locate them and compare them for changes. It is much easier to automate checking/amending an artwork that was partly/fully created by automation than an artwork built completely manually. Some human interaction will be unavoidable with artworks of this complexity and range, but you can reduce the variability that human operators always introduce.   Sorry that’s not an easy answer, but it’s not an easy problem, both technological and logistical. (I was 90% automating those artworks over a decade ago and the industry still hasn’t progressed any. And while I had first artworks solved I still hadn’t figured a satisfactory solution to QA and amends by the time I left, so I can’t give you all the answers as I don’t know them myself.) ... View more

Ma’am, this is a Wendy’s.   Try: https://community.adobe.com/t5/indesign/ct-p/ct-indesign ... View more

Code looks like generative AI garbage. A complete waste of time trying to understand it. Honestly, you should just say if you don’t know how to write a script yourself and ask if someone here is willing to write it for you (free or paid).   As it happens, I just had a someone asking a very similar question so figured I’d see how long it takes to write a script in practice. (Answer: this is as much as I’m prepared to do as a freebie, but at least it points you in the right direction.)   function readTextFile(f) { f.encoding = 'UTF-8'; // presumably f.open('r'); var result = f.read(); f.close(); return result; } function appendTextFile(f, s) { f.encoding = 'UTF-8'; // presumably f.open('a'); f.writeln(s); f.close(); } var doc = app.activeDocument; var f = doc.fullName; if (!f.exists) { throw new Error("Please save document and run again."); } var originalTextFile = File.openDialog("Choose the original text file."); var originalText = readTextFile(originalTextFile); var originalParagraphs = originalText.replace(/\s*[\r\n]\s*/g, '\n').split(/\n/g); originalParagraphs.sort(); var foundParagraphs = []; for (var i = 0; i < doc.stories.length; i++) { var paragraphs = doc.stories[i].paragraphs; for (var j = 0; j < paragraphs.length; j++) { foundParagraphs.push(paragraphs[j].contents.replace(/^\s+|\s+$/g)); } } foundParagraphs.sort(); // _matched.txt = text that is found in both documents // _missing.txt = text that is in original text file but not in artwork // _unknown.txt = text that is in artwork but not in original text file var matchedFile = new File(originalTextFile.fullName.replace(/\.[^.]+$/, '_matched.txt')); var missingFile = new File(originalTextFile.fullName.replace(/\.[^.]+$/, '_missing.txt')); var unknownFile = new File(originalTextFile.fullName.replace(/\.[^.]+$/, '_unknown.txt')); matchedFile.remove(); missingFile.remove(); unknownFile.remove(); for (var i = 0; i < originalParagraphs.length; i++) { var originalParagraph = originalParagraphs[i]; while (foundParagraphs.length > 0 && foundParagraphs[0] < originalParagraph) { appendTextFile(unknownFile, foundParagraphs.shift()); } if (foundParagraphs[0] === originalParagraph) { appendTextFile(matchedFile, foundParagraphs.shift()); } else { appendTextFile(missingFile, originalParagraph); } } while (foundParagraphs.length > 0) { appendTextFile(unknownFile, foundParagraphs.shift()); }   ... View more

From the ExtendScript documentation:   18.5.5 execute() fileObj.execute() Opens this file using the appropriate application, as if it had been double-clicked in a file browser. You can use this method to run scripts, launch applications, and so on. Returns true immediately if the application launch was successful.   Using JSX’s `execute` to run external scripts is limited and fragile. What you really want is an `app.doScript()` method as in InDesign, but AI doesn’t provide that. You could use `execute` to run an AppleScript app that writes the information to a temporary file at a known location; the JSX script can then wait until the file exists to read the information from it, then delete it.   Alternatively, if you can settle for using the current user’s login name, you can read the OS’s environment variables via JSX’s global “dollar” object:     $.getenv('USER')     ... View more

Look into using Illustrator’s COM APIs. e.g. For Python:   https://www.google.com/search?q=python+win32com+adobe+illustrator   https://pypi.org/project/pywin32/   For Node.js:   https://www.npmjs.com/package/winax     ... View more

https://community.adobe.com/t5/illustrator-discussions/import-csv-file/m-p/14427733#M397733 ... View more

I can confirm this issue (28.5), and would agree it’s a bug.   Have you tried the same script in JSX? AI’s AS and JS APIs share a lot of common plumbing so chances are it’s affected too, and this should increase priority for a fix. ... View more

I’m guessing this script is AI-generated. The `applyDropShadow` function is 100% nonsense. There is no `app.effects` property in AI’s JSX API (not that that line does anything useful with the result), nor is there a `DropShadowOptions` constructor.   Take out that function and the rest of the script appears to run to completion (I won’t vouch that it produces the result you want).   I can’t think offhand of a way to apply a custom drop shadow via AI’s own automation—neither ExtendScript nor Actions can control that feature. There might be something in the C++ SDK. Or you might be able to manipulate AI’s GUI menus and dialogs using an external automation tool, e.g. AutoIt (Windows) or GUI Scripting (Mac), although that is another brittle mess entirely.   Better to find some other way to add the drop shadow element to the artwork. You’d need to provide more information and multiple examples. ... View more

Your JavaScript is fine. The problem is your file path. If the file doesn’t exist, the `do javascript` command throws a “JavaScript code was missing” error. (The error message is misleading—it should say “file not found”.)   Here’s a useful tip: drag your `test.jsx` file from Finder into the Script Editor window. That will add the correct file path to your AppleScript code. ... View more

POSIX paths do not begin with the hard drive name. Try:   property scriptPath : "/用户/china144/桌面/TEST/test.jsx" as POSIX file   ... View more

// fit_text_to_frame.jsx // resizes text larger/smaller to fit the selected text frame(s) // (the text frame must be fully selected; point text and non-text items are ignored) // current limitations: this assumes all characters are same size; leading is not adjusted (function() { var MIN_SIZE = 1, MAX_SIZE = 1296 // practical limits, in pts var MIN_STEP = 0.1 // the smaller this is, the more precise (and slower) the fit function fitTextToFrame(textFrame) { // textFrame : TextFrame -- path or area text if (textFrame.typename !== 'TextFrame' || textFrame.kind === TextType.POINTTEXT) { return } var length = textFrame.contents.replace(/\s+$/, "").length // length of printable text (ignores trailing whitespace) if (length === 0) { return } // skip empty text frame // overflow checker; this checks length of printable text to index of last visible character in frame var hasOverflow = function() { var lastLine = textFrame.lines[textFrame.lines.length-1] if (!lastLine) { return true } // no lines visible (frame is too small) var visibleLength = (lastLine.characters[0].characterOffset + lastLine.length - 1) return visibleLength < length } // word count checker; assuming no overflow, this detects large words that are force-wrapped onto >1 line var hasSplitWords = function() { var wc = 0 for (var i = 0; i < textFrame.lines.length; i++) { wc += textFrame.lines[i].words.length } return wc > textFrame.words.length } // (caution: do not use JS’s +=, -=, *=, /= operators on AI properties as JSX doesn't handle them right) var style = textFrame.textRange.characterAttributes var size = style.size // 1. resize text so it is just overflowing while ((hasOverflow() || hasSplitWords()) && size >= MIN_SIZE * 2) { // decrease size until no overflow size /= 2.0 style.size = size } if ((hasOverflow() || hasSplitWords())) { return } // hit minimum size and it's still overflowing, so give up while (!(hasOverflow() || hasSplitWords()) && size <= MAX_SIZE / 2) { // increase size until text overflows size *= 2.0 style.size = size } if (!(hasOverflow() || hasSplitWords())) { return } // hit maximum size and it hasn't overflowed, so give up // 2. now do binary search for the largest size that will just fit the text frame var step = size / 2.0 while (step > MIN_STEP) { step /= 2.0 size += (hasOverflow() || hasSplitWords()) ? -step : step style.size = size } // 3. if text ends up overflowing after last step, step it back down while ((hasOverflow() || hasSplitWords())) { style.size = size - step step *= 1.1 } } var items = app.activeDocument.selection for (var i = 0; i < items.length; i++) { // iterate over currently selected items fitTextToFrame(items[i]) } })() ... View more

The linked script doesn’t use Illustrator’s variables/data merge feature, so can insert multiple CSV rows per artwork. Did you test the script with the supplied example files?   If that script is not suitable for your needs as-is (e.g. you need to replace portions of text within the text frame) then you will need to provide more details if folks here are to help you. I recommend adding Before and After examples of your artwork, plus a sample CSV file, to your original post. ... View more