Jump to content
adrian

Tracy Debugger

Recommended Posts

Hi @adrian

today Tracy cost me 1 hour of work 😞 

$this->addHookAfter('Session::loginSuccess', function(HookEvent $event) {
  bd('Session::loginSuccess');
});

No matter what hook I tried (Session::login, ProcessLogin::loginSuccess etc) tracy did not dump anything. It seemed that the hook did never fire. I even placed a bd('Session.php') in Session.php... nothing! 😮

I tried different Session Handlers (DB, default), tried hooking ::logout (which worked) etc etc.

Then I found:

$this->addHookAfter('Session::loginSuccess', function(HookEvent $event) {
  bd('Session::loginSuccess');
  $this->message('Session::loginSuccess');
});

H6XN4mh.png

So the hook fires, but the BD() call get's lost somewhere! The dumps recorder is also empty (not shown in the screenshot)!

Any ideas what could be going on?

 

PS: My new balance

  • Tracy time costs: 1h
  • Tracy time savings: 500h+ (thx once more!! 🙂 )
  • Like 1

Share this post


Link to post
Share on other sites

Hey @bernhard - I just tested here with all combinations of the following on my local dev setup (MacOS)

1) SessionHandlerDB on and off
2) Your hook in init.php and ready.php and in the Console panel injected at init and ready

In all cases, the bd() shows up in one of the Tracy redirect bars.

Then I tried on one of my linux servers and in all tests, the bd() call was never activated. My thought is that it must be something to do with when Tracy is loaded. Because Tracy is loaded by PW's module system, it isn't loaded until site modules get loaded. Ryan made it so that it is the first of these site modules to be loaded which means we can always reliably use it to test development of other modules, but it doesn't load before the PW core which is why it's sometimes impossible to use it to debug the core. 

All that said, I don't know if load order is the problem in this case, or if it's something else. I'd actually love to know if you see a difference on different servers / PW installs. And anyone else who'd like to test also - it would be good to figure out if there is something consistent with when it does and doesn't work.

  • Like 1

Share this post


Link to post
Share on other sites

Thx @adrian

the strange thing is. If the load order of the modules was the problem, shouldn't the bd() throw an error that the function does not exist? The hook seems to be fired, because the $this->message() shows up. 🤔

  • Like 1

Share this post


Link to post
Share on other sites
57 minutes ago, bernhard said:

the strange thing is. If the load order of the modules was the problem, shouldn't the bd() throw an error that the function does not exist? The hook seems to be fired, because the $this->message() shows up. 🤔

Yeah, that makes sense 🙂

Ok, so after spending way too much time, it's not specifically a Tracy issue. If you do a Tracy log call, eg:

$this->addHookAfter('Session::loginSuccess', function(HookEvent $event) {
    l('Session::loginSuccess');
});

then it works fine. The issue is that dumps are stored in a variable and added to as needed during the current request. But it seems like PW hasn't actually started its session at Session::loginSuccess. If you try the same bd() call at Session::init then it works as expected.

$this->addHookAfter('Session::init', function(HookEvent $event) {
    bd('Session::init'); // works
});

So, I looked into the possibility of making it possible to do bd() calls and have them carry over to the next PW session (after the login has actually started the session) and there are possibilities, but I think they all come with downsides. I think it might help the Dumps Recorder panel though in that I'll be able to make it keep data over sessions, so I might consider that. I'll mull it over for a bit.

As a reminder (almost just for myself), Session::loginSuccess does work on some setups (eg my local dev environment), so there must be some differences in how sessions are handled that is impacting this.

  • Thanks 1

Share this post


Link to post
Share on other sites

Just a quick update to you all know that there is a new keyboard shortcut for reloading a snippet from disk, clearing previous output, and running the code which should be a nice time saver and also prevent you from accidentally running old code in the console that hasn't been updated from the disk version when you are using an external editor.

image.png.6d6d7b0e97ad966ef651f65f2d096ef9.png

I have also rejigged the execution of code when "Run" is called vs injecting. Hopefully the visual cues of the button changing from "Run" to "Inject" and the status in the header showing "Inject @ Ready" makes it more obvious what is going to happen when.

Another update is that the Dumps Recorder panel now stores its data in a file so that it will survive across sessions which may be useful if you are hooking certain session methods and other scenarios where the normal Dumps panel is not showing your bd() calls.

  • Like 5

Share this post


Link to post
Share on other sites

New toggle arrow to collapse / show the snippets side pane with the state remembered via cookie.

image.thumb.png.42b1fdba3b216cbca94919b935553c1c.png

VS

image.thumb.png.458796aa658f7f957069808dfc7cb72d.png

  • Like 2

Share this post


Link to post
Share on other sites

If you work with deeply nested / large JSON data and you use Adminer, the latest update will allow you to speed things up significantly with a few new config options. I've also included a PR to the JSON preview plugin which this functionality relies on - my changes ensure the max text length setting is applied to json that is kept as a string due to the max level setting. 

image.thumb.png.21dd7e8af779db08b729e882c6bb5f5e.png

  • Like 3

Share this post


Link to post
Share on other sites

Hey @adrian unfortunately I have some very strange issues with the dumps feature 😞 

This is the code I have (using Nette Forms):

[...]

    l('renderMail');
    bd('renderMail');

    // submit
    if($form->isSuccess()) {
      l('success!');
      bd('success!');
      $session->redirect('thank-you-page');

[...]

I have several problems:

1) The bd() calls do only show up in the dumps recorder, not in the regular AJAX bar dump.

2) The dumps recorder shows the entries twice whereas the tracy logs show them only once.

ZawHTd4.png

The tracy logs should be the correct results, because after the form success I do a session redirect to the thank you page. No idea why things show up twice in the dumps recorder. Of course I have cleared the dumps before submitting the form 😉 And it's the latest version of tracy from today (4.21.10)

Explanation of the dumps recorder:

1) renderMail is called on the initial page view (correct, logs at 18:10:12)

2) renderMail is called when the form was submitted (correct, logs at 18:10:21)

3) success is fired because the form was successfully submitted (correct)

4) send mail is fired correctly

5) 2, 3 and 4 are dumped again (not correct)

 

Here another dump showing timestamps:

CSaF8jm.png

 

Any ideas? 😞 

Edit:

I checked by sending an email additionally to the "send mail" dump:

  public function send($message, $contact) {
    bd("send $message to $contact");
    $mail = new WireMail();
    $mail->to('foo@bar.com');
    $mail->from('foo@bar.com');
    $mail->subject("Test @ ".date('d.m.Y H:i:s'));
    $mail->send();
  }

The mail was actually sent twice, so it seems that the tracy dump is correct. So what about the tracy log? Shouldn't that also be listed twice then? No idea why that send() call fires twice at all... there is a session redirect that should prevent this 😞 

Share this post


Link to post
Share on other sites
3 minutes ago, bernhard said:

1) The bd() calls do only show up in the dumps recorder, not in the regular AJAX bar dump.

Is a new AJAX bar showing at all (even without a Dumps panel entry). If it's not, then it's probably an issue with Nette forms not sending the 'X-Requested-With': 'XMLHttpRequest'

8 minutes ago, bernhard said:

after the form success I do a session redirect to the thank you page

Or maybe it's related to that - if it's an AJAX form submission and then you do a redirect afterwards, that a) seems to defeat the reason for an ajax submission, and b) is probably why the AJAX bar is lost. Does it show in the ajax bar if you don't do the redirect?

11 minutes ago, bernhard said:

2) The dumps recorder shows the entries twice whereas the tracy logs show them only once.

Not sure about that - I'll look a little later and after I hear back about those possible reasons for #1

Share this post


Link to post
Share on other sites
2 hours ago, adrian said:

Is a new AJAX bar showing at all (even without a Dumps panel entry). If it's not, then it's probably an issue with Nette forms not sending the 'X-Requested-With': 'XMLHttpRequest'

There's only one single bar. But it's not an AJAX request. I'm submitting the form as regular POST and then doing a session->redirect to a new page. But usually tracy also log's redirected requests (like after page save).

Sorry about the ajax confusion! I meant those redirect bars:

bIxCZSn.png

Share this post


Link to post
Share on other sites
On 5/4/2020 at 6:24 PM, bernhard said:

The mail was actually sent twice, so it seems that the tracy dump is correct. So what about the tracy log? Shouldn't that also be listed twice then? No idea why that send() call fires twice at all... there is a session redirect that should prevent this 😞 

I found the reason for this: I'm rendering a repeater with several blocks. There are two subscription blocks with nette forms. A form submit triggers the redirect but the redirect does not happen immediately because it does finish() the active process, hence renders the other blocks as well. That's why I got 2 dumps (one for each block). I added a custom instandRedirect method to my repeater blocks:

public function instantRedirect($url) {
  header("Location: $url");
  exit(0);
}

 

Share this post


Link to post
Share on other sites

Hi @adrian, I'm getting the following error in the RequestInfo panel with  many of my custom modules, do you know what's causing this?

ProcessWire\WireException: Method TrelloWire::getArray does not exist or is not callable in this context in /var/www/vhosts/klforexpats.preview-server.dev/httpdocs/public/wire/core/Wire.php:529
Stack trace:
#0 /var/www/vhosts/klforexpats.preview-server.dev/httpdocs/public/wire/core/Wire.php(386): ProcessWire\Wire->___callUnknown()
#1 /var/www/vhosts/klforexpats.preview-server.dev/httpdocs/public/wire/core/WireHooks.php(823): ProcessWire\Wire->_callMethod()
#2 /var/www/vhosts/klforexpats.preview-server.dev/httpdocs/public/wire/core/Wire.php(450): ProcessWire\WireHooks->runHooks()
#3 /var/www/vhosts/klforexpats.preview-server.dev/httpdocs/public/wire/core/Wire.php(453): ProcessWire\Wire->__call()
#4 /var/www/vhosts/klforexpats.preview-server.dev/httpdocs/public/site/assets/cache/FileCompiler/site/modules/TracyDebugger/panels/RequestInfoPanel.php(295): ProcessWire\Wire->__call()
#5 /var/www/vhosts/klforexpats.preview-server.dev/httpdocs/public/site/assets/cache/FileCompiler/site/modules/TracyDebugger/tracy-2.7.x/src/Tracy/Bar/Bar.php(150): RequestInfoPanel->getPanel()
#6 /var/www/vhosts/klforexpats.preview-server.dev/httpdocs/public/site/assets/cache/FileCompiler/site/modules/TracyDebugger/tracy-2.7.x/src/Tracy/Bar/Bar.php(122): Tracy\Bar->renderPanels()
#7 /var/www/vhosts/klforexpats.preview-server.dev/httpdocs/public/site/assets/cache/FileCompiler/site/modules/TracyDebugger/tracy-2.7.x/src/Tracy/Bar/Bar.php(98): Tracy\Bar->renderHtml()
#8 /var/www/vhosts/klforexpats.preview-server.dev/httpdocs/public/site/assets/cache/FileCompiler/site/modules/TracyDebugger/tracy-2.7.x/src/Tracy/Debugger/Debugger.php(293): Tracy\Bar->render()
#9 [internal function]: Tracy\Debugger::shutdownHandler()
#10 {main}

This happens in the module configuration (in this case the module is TrelloWire, but I get this with most custom modules). Is this indicative of some problem with my module? Or just some strange interaction between the module config and Tracy?

Share this post


Link to post
Share on other sites

@MoritzLost - from my initial investigation, it looks like a PW core bug. You are using this (https://processwire.com/blog/posts/new-module-configuration-options/#enter-the-new-moduleconfig-class) approach to defining module settings which I have never used, but it appears that when that approach is used, the getArray() method doesn't exist on the module, hence the error.

Do you know of another module that uses that approach that I could test to confirm?

 

  • Like 1

Share this post


Link to post
Share on other sites

@adrian I have tested my other modules; I think the error only occurs when the module does NOT extend WireData, since the getArray method comes from there.

  • TextformatterPageTitleLinks extends Textformatter which extends WireData -> RequestInfoPanel is working.
  • ProcessCacheControl extends Process which extends WireData -> RequestInfoPanel is working.
  • TrelloWire extends Wire (instead of WireData) -> RequestInfoPanel displays the error.
    • When I change the class definition to extend WireData instead of Wire, the error disappears.

I have tested & confirmed this behaviour with another module I'm working on (currently not publicly available, will probably publish this next week or so). So I think it comes down to modules that don't extend WireData at all, which isn't required (at least according to the docs). I guess it's mostly a cosmetic choice, but for the modules that don't really keep track of data I didn't want to extend WireData. Maybe the RequestInfoPage could check for that case and use getModuleConfig instead?

  • Like 1

Share this post


Link to post
Share on other sites

Hi @adrian,

I've been cleanin' up my latest project and noticed this warning:

Cookie “tracyClearDumpsRecorderItems” will be soon rejected because it has the “sameSite” attribute set to “none” or an invalid value, without the “secure” attribute. To know more about the “sameSite“ attribute, read https://developer.mozilla.org/docs/Web/HTTP/Cookies

I don't know if it means anything. If so, great. If not, please ignore this post. 🙂

  • Like 1

Share this post


Link to post
Share on other sites

Thanks @MoritzLost for figuring that out. I have added a check to prevent that error when a module doesn't extend WireData. Honestly, I think the getting of module settings in PW is a bit of a mess which is why I use 3 different approaches in Tracy because they all return different things for some modules. Let me know if you have any problems with the new version.

Thanks @rick for the heads up on that. I took a quick look, but I don't have time to fix things at the moment - I set a lot of cookies in various places in Tracy and I am not using a standard method so it's not easy to change them all and I sometimes set in JS and sometimes in PHP. Obviously I don't want to make them "secure" because I am sure many of us are running local dev setups without https, so I probably need to just set the sameSite rule to Lax, but maybe Strict is better? Not sure yet, but it's on my list to take care of soon.

  • Like 1

Share this post


Link to post
Share on other sites
32 minutes ago, adrian said:

Thanks @MoritzLost for figuring that out. I have added a check to prevent that error when a module doesn't extend WireData. Honestly, I think the getting of module settings in PW is a bit of a mess which is why I use 3 different approaches in Tracy because they all return different things for some modules. Let me know if you have any problems with the new version.

Thanks for the fix @adrian, I just tested it and it seems to be working! I'm not getting errors any more with both modules I tested before 🙂 

34 minutes ago, adrian said:

Thanks @rick for the heads up on that. I took a quick look, but I don't have time to fix things at the moment - I set a lot of cookies in various places in Tracy and I am not using a standard method so it's not easy to change them all and I sometimes set in JS and sometimes in PHP. Obviously I don't want to make them "secure" because I am sure many of us are running local dev setups without https, so I probably need to just set the sameSite rule to Lax, but maybe Strict is better? Not sure yet, but it's on my list to take care of soon.

Just butting in here: I don't believe you want to use Strict, this may break the debugger for if you visit the site by clicking on a link from somewhere else. Since you can't really use Secure cookies because of the local dev environments like you said, you'll probably be best of with SameSite=Lax to most cookies. This article on web.dev is a good resource for the SameSite changes.

  • Like 1

Share this post


Link to post
Share on other sites
1 minute ago, MoritzLost said:

Just butting in here: I don't believe you want to use Strict, this may break the debugger for if you visit the site by clicking on a link from somewhere else. Since you can't really use Secure cookies because of the local dev environments like you said, you'll probably be best of with SameSite=Lax to most cookies. This article on web.dev is a good resource for the SameSite changes.

Much appreciated - I'll admit to not knowing too much about these options yet, so great to get some more knowledgeable input. Thanks!

  • Like 1

Share this post


Link to post
Share on other sites

Hey @adrian,

would it somehow be possible to get the tracy console in a separate process module? I'm working on my RockFinder and there I have built a "RockFinderTester" process module that does basically the same thing like the tracy console, only a lot worse 😄 I have many modules where a console feature would be great for testing. At the moment I'm using the console for lots of things, but I wonder how complicated it would be to move the console to a separate page that we can modify to our needs.

So for example for testing a RockFinder it might be great to have the code input on the top and below that it would be great to have an area where I can not only dump variables but also show the result as paginated table (using tabulator.js). The tabulator part could totally be done by me - I'd only need to have the proper JS events I can listen for.

I know we have the console snippets at the moment, but I think the console is such a powerful tool, that it would be great if it could be used by other modules as well.

Do you understand my idea? Would that be very complicated to achieve? Thanks in advance for thinking about it 🙂 

Share this post


Link to post
Share on other sites
16 minutes ago, bernhard said:

I wonder how complicated it would be to move the console to a separate page that we can modify to our needs

I have thought about making it a separate Process module so it can be opened on it's own page in the PW admin, but I am not sure about making it modifiable - what would you need to modify for you needs? Based on your tabulator example, it sounds like you are talking about extending the output options? But how best to do that? Is it just a matter of being able to inject (in the background) some additional JS to process the code and format as a table?

I am happy to talk about this more, but I haven't enough time at the moment to do anything substantial on this unfortunately. If you could think through how it should work in some more detail, that might make make it easier for me to implement. Of course PRs are also welcome, but I am happy to help implement if you figure out a plan for how it should work.

 

Share this post


Link to post
Share on other sites

Thx adrian,

this is what I have currently in my RockFinder2 module:

lTFDooA.png

This is basically a sandbox where one can test finders, sees a dump and also sees the data listed as tabulator. When looking at the code of RockFinder2 today I was quite surprised that it has +1200 lines of code. I had in my mind that it is quite lightweight compared to RockFinder1 because it builds upon the pw core query classes. Then I realized that I actually have lots of stuff packed into RockFinder2 that imho does not belong there. The finder tester is one of those things! I want to remove that part from the module so that RockFinder3 will just do one thing: Finding PW data and returning it as array of objects.

Nevertheless the finder tester is useful! So I don't really want to drop it as a whole. I can use the tracy console and do this:

kuqEJZp.png

This is already great, but it would be even greater if I could see the data in a tabulator. That's what I have in my process module, BUT: What I do not have there is the dumping features of the tracy console. Look at the first screenshot: There is a dump, but I always dump the returned finder object. I cant to any other dumps, because it's a custom implementation of the console. I also don't have the same settings and shortcuts as in the console.

Vision: Ideally I'd like to build RockFinder3 with only the features that it needs and pack the tester in a separate module. This module would require tracy (but only the process module, not the finder module) and could use the tracy console process. This process would show an InputfieldForm that everybody could modify easily via hooks. So I could just add a new InputfieldMarkup to that form below the tracy console, add some lines of javascript that intercept the AJAX events sent by the console and then show the result as tabulator grid.

Hm. I see that the ajax response at the moment does only send the html of the dumps. Maybe it could instead send the response as JSON? One property being the dump: "<foo>bar</foo>" and the other property being the returned value:

3Uob1Z6.png

In this case the code would return a RockFinder2 instance, so we would need something on the backend that transforms the returned data (here the rockfinder) in something readable by the client. Maybe something like this:

// pseudo code
$wire->addHookAfter("Console::getPayload", function($event) {
  $rockfinder = $event->arguments(0);
  $array = $rockfinder->toPlainArrayReadyForJavaScript(); // ['foo', 'bar']
  $event->return = $array;
});

This would result in this ajax response:

{
	dump: "<foo>bar</foo>",
	return: ['foo', 'bar']
}

So the console could just dump the markup and I could intercept that response with JS:

$(document).on('tracyConsoleResponse', function() {
  // get the return value of the xhr
  // render tabulator
});

Maybe that kind of refactoring could also make sense for the console? Maybe this approach could also make sense in an even bigger context: What if all panels where process modules? I for example have no idea how to build a panel for tracy (I'd have to study my hello world module again 😉 ), but it would be piece of cake to build a process module. Maybe I'm requesting too much here. But again: Maybe refactoring could make sense - I leave that assessment to you 🙂 

---

Side note: Also I have this bug when I open the console on my process module: The code of my ACE field gets quirky.

6S0Xr7M.png

That's not a request how to fix it - just to let you know why it would be great to have one central solution that can be modified instead of building our own solution and fix bugs 🙂 

 

  • Like 1

Share this post


Link to post
Share on other sites

Hey @bernhard - thanks for all the details. I'll have to get back to you on most of them, but I can explain the reason for the quirky stuff in your module. It is due to the really old version of ACE included with the InputfieldAceExtended module that you are making use of. The same thing happens with Ryan's HannaCode module - there is an issue posted about that here: https://github.com/ryancramerdesign/ProcessHannaCode/issues/21

  • Like 1

Share this post


Link to post
Share on other sites

Just had another idea that might be even more powerful and simpler to implement: What if we could define custom dumping functions? For example:

$rf = new RockFinder2();
[...]
tabulator($rf);

// modify tabulator and dump another one
tabulator($rf);

instead of a regular dump() this would render the markup of a tabulator, eg:

<div id="1905terf8dvuid"></div>
<script>
var table = new Tabulator("#1905terf8dvuid", {
 	height:205,
 	data:[
 	{id:1, name:"Oli Bob", age:"12", col:"red", dob:""},
 	{id:2, name:"Mary May", age:"1", col:"blue", dob:"14/05/1982"},
 	{id:3, name:"Christine Lobowski", age:"42", col:"green", dob:"22/05/1982"},
 	{id:4, name:"Brendon Philips", age:"125", col:"orange", dob:"01/08/1980"},
 	{id:5, name:"Margret Marmajuke", age:"16", col:"yellow", dob:"31/01/1999"},
 	],
 	layout:"fitColumns", //fit columns to width of table (optional)
 	columns:[ //Define Table Columns
	 	{title:"Name", field:"name", width:150},
	 	{title:"Age", field:"age", hozAlign:"left", formatter:"progress"},
	 	{title:"Favourite Color", field:"col"},
	 	{title:"Date Of Birth", field:"dob", sorter:"date", hozAlign:"center"},
 	],
 	rowClick:function(e, row){ //trigger an alert message when the row is clicked
 		alert("Row " + row.getData().id + " Clicked!!!!");
 	},
});
</script>

So basically we'd only need a way to define custom functions like "tabulator()" that define the returned markup and that's it 🙂 

  • Like 1

Share this post


Link to post
Share on other sites

@bernhard

Just a little rough experimentation - I've added a new echo() method to the TD class which outputs exactly what is sent to it.

With this addition, I managed to do this:

image.png.43a3b4c134549a7238c5c1c0e19b12b4.png

Obviously in this case I am not passing any data to the tabulator() method, but you can make this happen easily by modifying your new tabulator() function.

In a process module I have a couple of things to make this all work:

1) Load the tabulator js and css files in the init() method:

        $this->wire('config')->scripts->add("https://unpkg.com/tabulator-tables@4.6.3/dist/js/tabulator.min.js");
        $this->wire('config')->styles->add("https://unpkg.com/tabulator-tables@4.6.3/dist/css/tabulator.min.css");

2) Define a new tabulator function - note that in this case this is a global function outside the process module class, but you could also make it a class function and call $rf->tabulator()

function tabulator($str, $title = NULL) {
    \TD::echo('<div id="tabletest"></div>
    <script>
    var table = new Tabulator("#tabletest", {
         height:205,
         data:[
         {id:1, name:"Oli Bob", age:"12", col:"red", dob:""},
         {id:2, name:"Mary May", age:"1", col:"blue", dob:"14/05/1982"},
         {id:3, name:"Christine Lobowski", age:"42", col:"green", dob:"22/05/1982"},
         {id:4, name:"Brendon Philips", age:"125", col:"orange", dob:"01/08/1980"},
         {id:5, name:"Margret Marmajuke", age:"16", col:"yellow", dob:"31/01/1999"},
         ],
         layout:"fitColumns", //fit columns to width of table (optional)
         columns:[ //Define Table Columns
             {title:"Name", field:"name", width:150},
             {title:"Age", field:"age", hozAlign:"left", formatter:"progress"},
             {title:"Favourite Color", field:"col"},
             {title:"Date Of Birth", field:"dob", sorter:"date", hozAlign:"center"},
         ],
         rowClick:function(e, row){ //trigger an alert message when the row is clicked
             alert("Row " + row.getData().id + " Clicked!!!!");
         },
    });
    </script>', $title);
}


You can try it out by replacing the attached TD.php for your Tracy install (in the includes subfolder).

Note that you can supply two arguments to the \TD::echo method: $str and $title, just like with d() and bd()

Does that work for you ok?

TD.php

  • Like 1
  • Thanks 1

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

  • Recently Browsing   0 members

    No registered users viewing this page.

  • Similar Content

    • By FireWire
      Hello community!

      I want to share a new module I've been working on that I think could be a big boost for multi-language ProcessWire sites.

      Some background, I was looking for a way for our company website to be efficiently translated as working with human translators was pretty laborious and a lack of updating content created a divergence between languages. I, and several other devs here, have talked about translation integrations and have recognized the power that DeepL has. DeepL is an AI deep learning powered service that delivers translation quality beyond any automated service available. After access to the API was opened up to the US, I built Fluency, a DeepL translation integration for ProcessWire.
      Fluency brings automated translation to every multi-language field in the admin, and also provides a translation tool allowing the user to translate their text to any language without it being inside a template's field. With Fluency you can:
      Translate any plain textarea or text input Translate any CKEditor content (yes, with markup) Translate page names for fully localized URLs on every page Translate your in-template translation function wrapped strings Translate modules DeepL offers translations to the following languages: English (US), English (UK), German, French, Spanish, Portuguese (EU), Portuguese (Brazil, Italian, Dutch, Polish, Russian, Japanese, Chinese (Simplified)
      Installation and usage is completely plug and play. Whether you're building a new multi-language site, need to update a site to multi-language, or simply want to stop manually translating a site and make any language a one-click deal, it could not be easier to do it. Fluency works by having you match the languages configured in ProcessWIre to DeepL's. You can have your site translating to any or all of the languages DeepL translates to in minutes (quite literally).
      Let's break out the screenshots...
      When the default language tab is shown, a message is displayed to let users know that translation is available. Clicking on each tab shows a link that says "Translate from English". Clicking it shows an animated overlay with the word "Translating..." cycling through each language and a light gradient shift. Have a CKEditor field? All good. Fluency will translated it and use DeepL's ability to translate text within HTML tags. CKEditor fields can be translated as easily and accurately as text/textarea fields.

      Repeaters and AJAX created fields also have translation enabled thanks to a JavaScript MutationObserver that searches for multi-language fields and adds translation as they're inserted into the DOM. If there's a multi-language field on the page, it will have translation added.

      Same goes for image description fields. Multi-language SEO friendly images are good to go.

      Creating a new page from one of your templates? Translate your title, and also translate your page name for native language URLs. (Not available for Russian, Chinese, or Japanese languages due to URL limitations). These can be changed in the "Settings" tab for any page as well so whether you're translating new pages or existing pages, you control the URLs everywhere.

      Language configuration pages are no different. Translate the names of your languages and search for both Site Translation Files (including all of your modules)

      Translate all of the static text in your templates as well. Notice that the placeholders are retained. DeepL is pretty good at recognizing and keeping non-translatable strings like that. If it is changed, it's easy to fix manually.

      Fluency adds a "Translate" item to the CMS header. When clicked this opens up a modal with a full translation tool that lets the user translate any language to any language. No need to leave the admin if you need to translate content from a secondary language back to the default ProcessWire language. There is also a button to get the current API usage statistics. DeepL account owners can set billing limitations via character count to control costs. This may help larger sites or sites being retrofitted keep an eye on their usage. This tool is available for all users with the page-edit permission.

      It couldn't be easier to add Fluency to your new or existing website. Simply add your API key and you're shown what languages are currently available for translation from/to as provided by DeepL. This list and all configuration options are taken live from the API so when DeepL releases new languages you can add them to your site without any work. No module updates, just an easy configuration. Just match the language you configured in ProcessWire to the DeepL language you want it to be associated with and you're done. Fluency also allows you to create a list of words/phrases that will not be translated which can prevent items such as brands and company names from being translated when they shouldn't

       
      Limitations:
      No "translate page" - Translating multiple fields can be done by clicking multiple translation links on multiple fields at once but engineering a "one click page translate" is not feasible from a user experience standpoint. The time it takes to translate one field can be a second or two, but cumulatively that may take much longer (CKEditor fields are slower than plain text fields). There may be a workaround in the future but it isn't currently on the roadmap. No "translate site" - Same thing goes for translating an entire website at once. It would be great, but it would be a very intense process and take a very (very) long time. There may be a workaround in the future but it isn't on the roadmap. No current support for Inline CKEditor fields - Handling for CKEditor on-demand hasn't been implemented yet, this is planned for a future release though and can be done. I just forgot about it because I've never really used that feature personally.. Alpha release - This module is in alpha. Releases should be stable and usable, but there may be edge case issues. Test the module thoroughly and please report any bugs via a Gitlab issue on the repository or respond here. Please note that the browser plugin for Grammarly conflicts with Fluency (as it does with many web applications). To address this issue it is recommended that you disable Grammarly when using Fluency, or open the admin to edit pages in a private window where Grammarly may not be loaded. This is an issue that may not have a resolution as creating a workaround may not be possible. If you have insight as to how this may be solved please visit the Gitlab page and file a bugfix ticket.
      Requirements:
      ProcessWire  3.0+ UIKit Admin Theme That's Fluency in a nutshell. A core effort in this module is to create it so that there is nothing DeepL related hard-coded in that would require updating it when DeepL offers new languages. I would like this to be a future-friendly module that doesn't require developer work to keep it up-to-date.
      It's Free
      This is my first real module and I want to give it back to the community as thanks. This is the best CMS I've worked with (thank you Ryan & contributors) and a great community (thank you dear reader). The only cost to use this is a subscription fee for the DeepL Pro API. Find out more and sign up here.
      Download & Feedback
      Download the latest version here
      https://gitlab.com/SkyLundy/fluency-processwire/-/archive/master/fluency-processwire-master.zip
      Gitlab repository:
      https://gitlab.com/SkyLundy/fluency-processwire
      File issues and feature requests here (your feedback and testing is greatly appreciated):
      https://gitlab.com/SkyLundy/fluency-processwire/-/issues
       
      Thank you! ¡Gracias! Ich danke Ihnen! Merci! Obrigado! Grazie! Dank u wel! Dziękuję! Спасибо! ありがとうございます! 谢谢你!

    • By horst
      ---------------------------------------------------------------------------------------------------------------------------------
        when working with PW version 2.6+, please use Pim2, not Pim!
        read more here  on how to change from the older to the newer version in existing sites
      ---------------------------------------------------------------------------------------------------------------------------------
      PageImage Manipulator, API for version 1 & 2

      The Page Image Manipulator is a module that let you in a first place do ImageManipulations with your PageImages. - And in a second place there is the possibility to let it work on any imagefile that exists in your servers filesystem, regardless if it is a 'known PW-image'.

      The Page Image Manipulator is a Toolbox for Users and Moduledevelopers. It is written to be as close to the Core ImageSizer as possible. Besides the GD-filterfunctions it contains resize, crop, canvas, rotate, flip, sharpen, unsharpMask and 3 watermark methods.



      How does it work?

      You can enter the ImageManipulator by calling the method pim2Load(). After that you can chain together how many actions in what ever order you like. If your manipulation is finished, you call pimSave() to write the memory Image into a diskfile. pimSave() returns the PageImage-Object of the new written file so we are able to further use any known PW-image property or method. This way it integrates best into the ProcessWire flow.

      The three examples above put out the same visual result: a grayscale image with a width of 240px. Only the filenames will slightly differ.

      You have to define a name-prefix that you pass with the pimLoad() method. If the file with that prefix already exists, all operations are skipped and only the desired PageImage-Object gets returned by pimSave(). If you want to force recreation of the file, you can pass as second param a boolean true: pim2Load('myPrefix', true).

      You may also want to get rid of all variations at once? Than you can call $pageimage->pim2Load('myPrefix')->removePimVariations()!

      A complete list of all methods and actions are at the end of this post.
       
      You may also visit the post with tips & examples for users and module developers.


      How to Install
      Download the module Place the module files in /site/modules/PageImageManipulator/ In your admin, click Modules > Check for new modules Click "install" for PageImageManipulator Done! There are no configuration settings needed, just install and use it. Download    (version 0.2.0)
      get it from the Modules Directory History of origins

      http://processwire.com/talk/topic/3278-core-imagemanipulation/


      ----------------------------------------------------------------------------------------------------------


      Page Image Manipulator - Methods

      * pimLoad  or  pim2Load, depends on the version you use!

      pimLoad($prefix, $param2=optional, $param3=optional)
      param 1: $prefix - (string) = mandatory! param 2: mixed, $forceRecreation or $options param 3: mixed, $forceRecreation or $options return: pim - (class handle) $options - (array) default is empty, see the next method for a list of valid options! $forceRecreation - (bool) default is false It check if the desired image variation exists, if not or if forceRecreation is set to true, it prepares all settings to get ready for image manipulation
      -------------------------------------------------------------------

      * setOptions

      setOptions(array $options)
      param: $options - (array) default is empty return: pim - (class handle) Takes an array with any number valid options / properties and set them by replacing the class-defaults and / or the global config defaults optionally set in the site/config.php under imageSizerOptions or imageManipulatorOptions.

      valid options are:
      quality = 1 - 100 (integer) upscaling = true | false (boolean) cropping = true | false (boolean) autoRotation =true | false (boolean) sharpening = 'none' | 'soft' | 'medium' | 'strong' (string) bgcolor = (array) css rgb or css rgba, first three values are integer 0-255 and optional 4 value is float 0-1, - default is array(255,255,255,0) thumbnailColorizeCustom = (array) rgb with values for colorize, integer -255 - 255 (this can be used to set a custom color when working together with Thumbnails-Module)
        outputFormat = 'gif' | 'jpg' | 'png' (Attention: outputFormat cannot be specified as global option in $config->imageManipulatorOptions!) set {singleOption} ($value)
      For every valid option there is also a single method that you can call, like setQuality(90), setUpscaling(false), etc.
      -------------------------------------------------------------------

      * pimSave

      pimSave()
      return: PageImage-Object If a new image is hold in memory, it saves the current content into a diskfile, according to the settings of filename, imagetype, targetFilename and outputFormat. Returns a PageImage-Object!
      -------------------------------------------------------------------

      * release

      release()
      return: void (nothing) if you, for what ever reason, first load image into memory but than do not save it, you should call release() to do the dishes! 😉 If you use pimSave() to leave the ImageManipulator, release() is called automatically.
      -------------------------------------------------------------------

      * getOptions

      getOptions()
      return: associative array with all final option values example:
      ["autoRotation"] bool(true) ["upscaling"] bool(false) ["cropping"] bool(true) ["quality"] int(90) ["sharpening"] string(6) "medium" ["targetFilename"] string(96) "/htdocs/site/assets/files/1124/pim_prefix_filename.jpg" ["outputFormat"] string(3) "jpg" get {singleOption} ()
      For every valid option there is also a single method that you can call, like getQuality(), getUpscaling(), etc. See method setOptions for a list of valid options!
      -------------------------------------------------------------------

      * getImageInfo

      getImageInfo()
      return: associative array with useful informations of source imagefile example:
      ["type"] string(3) "jpg" ["imageType"] int(2) ["mimetype"] string(10) "image/jpeg" ["width"] int(500) ["height"] int(331) ["landscape"] bool(true) ["ratio"] float(1.5105740181269) ["bits"] int(8) ["channels"] int(3) ["colspace"] string(9) "DeviceRGB" -------------------------------------------------------------------

      * getPimVariations

      getPimVariations()
      return: array of Pageimages Collect all pimVariations of this Pageimage as a Pageimages array of Pageimage objects. All variations created by the core ImageSizer are not included in the collection.
      -------------------------------------------------------------------

      * removePimVariations

      removePimVariations()
      return: pim - (class handle) Removes all image variations that was created using the PIM, all variations that are created by the core ImageSizer are left untouched!
      -------------------------------------------------------------------
      * width

      width($dst_width, $sharpen_mode=null)
      param: $dst_width - (integer) param: $auto_sharpen - (boolean) default is true was deleted with version 0.0.8, - sorry for breaking compatibility param: $sharpen_mode - (string) possible: 'none' | 'soft' | 'medium' | 'strong', default is 'soft' return: pim - (class handle) Is a call to resize where you prioritize the width, like with pageimage. Additionally, after resizing, an automatic sharpening can be done with one of the three modes.
      -------------------------------------------------------------------

      * height

      height($dst_height, $sharpen_mode=null)
      param: $dst_height - (integer) param: $auto_sharpen - (boolean) default is true was deleted with version 0.0.8, - sorry for breaking compatibility param: $sharpen_mode - (string) possible: 'none' | 'soft' | 'medium' | 'strong', default is 'soft' return: pim - (class handle) Is a call to resize where you prioritize the height, like with pageimage. Additionally, after resizing, an automatic sharpening can be done with one of the three modes.
      -------------------------------------------------------------------

      * resize

      resize($dst_width=0, $dst_height=0, $sharpen_mode=null)
      param: $dst_width - (integer) default is 0 param: $dst_height - (integer) default is 0 param: $auto_sharpen - (boolean) default is true was deleted with version 0.0.8, - sorry for breaking compatibility param: $sharpen_mode - (string) possible: 'none' | 'soft' | 'medium' | 'strong', default is 'soft' return: pim - (class handle) Is a call to resize where you have to set width and / or height, like with pageimage size(). Additionally, after resizing, an automatic sharpening can be done with one of the three modes.
      -------------------------------------------------------------------

      * stepResize

      stepResize($dst_width=0, $dst_height=0)
      param: $dst_width - (integer) default is 0 param: $dst_height - (integer) default is 0 return: pim - (class handle) this performs a resizing but with multiple little steps, each step followed by a soft sharpening. That way you can get better result of sharpened images.
      -------------------------------------------------------------------

      * sharpen

      sharpen($mode='soft')
      param: $mode - (string) possible values 'none' | 'soft'| 'medium'| 'strong' return: pim - (class handle) Applys sharpening to the current memory image. You can call it with one of the three predefined pattern, or you can pass an array with your own pattern.
      -------------------------------------------------------------------

      * unsharpMask

      unsharpMask($amount, $radius, $threshold)
      param: $amount - (integer) 0 - 500, default is 100 param: $radius - (float) 0.1 - 50, default is 0.5 param: $threshold - (integer) 0 - 255, default is 3 return: pim - (class handle) Applys sharpening to the current memory image like the equal named filter in photoshop.
      Credit for the used unsharp mask algorithm goes to Torstein Hønsi who has created the function back in 2003.
      -------------------------------------------------------------------

      * smooth

      smooth($level=127)
      param: $level - (integer) 1 - 255, default is 127 return: pim - (class handle) Smooth is the opposite of sharpen. You can define how strong it should be applied, 1 is low and 255 is strong.
      -------------------------------------------------------------------

      * blur

      blur()
      return: pim - (class handle) Blur is like smooth, but cannot called with a value. It seems to be similar like a result of smooth with a value greater than 200.
      -------------------------------------------------------------------

      * crop

      crop($pos_x, $pos_y, $width, $height)
      param: $pos_x - (integer) start position left param: $pos_y - (integer) start position top param: $width - (integer) horizontal length of desired image part param: $height - (integer) vertical length of desired image part return: pim - (class handle) This method cut out a part of the memory image.
      -------------------------------------------------------------------

      * canvas

      canvas($width, $height, $bgcolor, $position, $padding)
      param: $width = mixed, associative array with options or integer, - mandatory! param: $height = integer, - mandatory if $width is integer! param: $bgcolor = array with rgb or rgba, - default is array(255, 255, 255, 0) param: $position = one out of north, northwest, center, etc, - default is center param: $padding = integer as percent of canvas length, - default is 0 return: pim - (class handle) This method creates a canvas according to the given width and height and position the memory image onto it.
      You can pass an associative options array as the first and only param. With it you have to set width and height and optionally any other valid param. Or you have to set at least width and height as integers.
      Hint: If you want use transparency with rgba and your sourceImage isn't of type PNG, you have to define 'png' as outputFormat with your initially options array or, for example, like this: $image->pimLoad('prefix')->setOutputFormat('png')->canvas(300, 300, array(210,233,238,0.5), 'c', 5)->pimSave()
      -------------------------------------------------------------------

      * flip

      flip($vertical=false)
      param: $vertical - (boolean) default is false return: pim - (class handle) This flips the image horizontal by default. (mirroring)
      If the boolean param is set to true, it flips the image vertical instead.
      -------------------------------------------------------------------

      * rotate

      rotate($degree, $backgroundColor=127)
      param: $degree - (integer) valid is -360 0 360 param: $backgroundColor - (integer) valid is 0 - 255, default is 127 return: pim - (class handle) This rotates the image. Positive values for degree rotates clockwise, negative values counter clockwise. If you use other values than 90, 180, 270, the additional space gets filled with the defined background color.
      -------------------------------------------------------------------

      * brightness

      brightness($level)
      param: $level - (integer) -255 0 255 return: pim - (class handle) You can adjust brightness by defining a value between -255 and +255. Zero lets it unchanged, negative values results in darker images and positive values in lighter images.
      -------------------------------------------------------------------

      * contrast

      contrast($level)
      param: $level - (integer) -255 0 255 return: pim - (class handle) You can adjust contrast by defining a value between -255 and +255. Zero lets it unchanged, negative values results in lesser contrast and positive values in higher contrast.
      -------------------------------------------------------------------

      * grayscale

      grayscale()
      return: pim - (class handle) Turns an image into grayscale. Remove all colors.
      -------------------------------------------------------------------

      * sepia

      sepia()
      return: pim - (class handle) Turns the memory image into a colorized grayscale image with a predefined rgb-color that is known as "sepia".
      -------------------------------------------------------------------

      * colorize

      colorize($anyColor)
      param: $anyColor - (array) like css rgb or css rgba - but with values for rgb -255 - +255,  - value for alpha is float 0 - 1, 0 = transparent  1 = opaque return: pim - (class handle) Here you can adjust each of the RGB colors and optionally the alpha channel. Zero lets the channel unchanged whereas negative values results in lesser / darker parts of that channel and higher values in stronger saturisation of that channel.
      -------------------------------------------------------------------

      * negate

      negate()
      return: pim - (class handle) Turns an image into a "negative".
      -------------------------------------------------------------------

      * pixelate

      pixelate($blockSize=3)
      param: $blockSize - (integer) 1 - ??, default is 3 return: pim - (class handle) This apply the well known PixelLook to the memory image. It is stronger with higher values for blockSize.
      -------------------------------------------------------------------

      * emboss

      emboss()
      return: pim - (class handle) This apply the emboss effect to the memory image.
      -------------------------------------------------------------------

      * edgedetect

      edgedetect()
      return: pim - (class handle) This apply the edge-detect effect to the memory image.
      -------------------------------------------------------------------

      * getMemoryImage

      getMemoryImage()
      return: memoryimage - (GD-Resource) If you want apply something that isn't available with that class, you simply can check out the current memory image and apply your image - voodoo - stuff
      -------------------------------------------------------------------

      * setMemoryImage

      setMemoryImage($memoryImage)
      param: $memoryImage - (GD-Resource) return: pim - (class handle) If you are ready with your own image stuff, you can check in the memory image for further use with the class.
      -------------------------------------------------------------------

      * watermarkLogo

      watermarkLogo($pngAlphaImage, $position='center', $padding=2)
      param: $pngAlphaImage - mixed [systemfilepath or PageImageObject] to/from a PNG with transparency param: $position - (string) is one out of: N, E, S, W, C, NE, SE, SW, NW,
      - or: north, east, south, west, center, northeast, southeast, southwest, northwest
      default is 'center' param: $padding - (integer) 0 - 25, default is 5, padding to the borders in percent of the images length! return: pim - (class handle) You can pass a transparent image with its filename or as a PageImage to the method. If the watermark is bigger than the destination-image, it gets shrinked to fit into the targetimage. If it is a small watermark image you can define the position of it:
      NW - N - NE | | | W - C - E | | | SW - S - SE  
      The easiest and best way I have discovered to apply a big transparency watermark to an image is as follows:
      create a square transparent png image of e.g. 2000 x 2000 px, place your mark into the center with enough (percent) of space to the borders. You can see an example here! The $pngAlphaImage get centered and shrinked to fit into the memory image. No hassle with what width and / or height should I use?, how many space for the borders?, etc.
      -------------------------------------------------------------------

      * watermarkLogoTiled

      watermarkLogoTiled($pngAlphaImage)
      param: $pngAlphaImage - mixed [systemfilepath or PageImageObject] to/from a PNG with transparency return: pim - (class handle) Here you have to pass a tile png with transparency (e.g. something between 150-300 px?) to your bigger images. It got repeated all over the memory image starting at the top left corner.
      -------------------------------------------------------------------

      * watermarkText

      watermarkText($text, $size=10, $position='center', $padding=2, $opacity=50, $trueTypeFont=null)
      param: $text - (string) the text that you want to display on the image param: $size - (integer) 1 - 100, unit = points, good value seems to be around 10 to 15 param: $position - (string) is one out of: N, E, S, W, C, NE, SE, SW, NW,
      - or: north, east, south, west, center, northeast, southeast, southwest, northwest
      default is 'center' param: $padding - (integer) 0 - 25, default is 2, padding to the borders in percent of the images length! param: $opacity- (integer) 1 - 100, default is 50 param: $trueTypeFont - (string) systemfilepath to a TrueTypeFont, default is freesansbold.ttf (is GPL & comes with the module) return: pim - (class handle) Here you can display (dynamic) text with transparency over the memory image. You have to define your text, and optionally size, position, padding, opacity for it. And if you don't like the default font, freesansbold, you have to point to a TrueTypeFont-File of your choice.
      Please have a look to example output: http://processwire.com/talk/topic/4264-release-page-image-manipulator/page-2#entry41989
      -------------------------------------------------------------------





      PageImage Manipulator - Example Output


    • By d'Hinnisdaël
      Format Datetime fields as Carbon instances.
      You can find the latest release and the complete readme on Github.
      Installation
      composer require daun/datetime-carbon-format Usage
      All Datetime fields will now be formatted as Carbon instances instead of strings. Some examples:
      // $page->date is a Datetime field // Output format: j/n/Y echo $page->date; // 20/10/2020 echo $page->date->add('7 days'); // 27/10/2020 echo $page->date->format('l, F j'); // Monday, October 20 echo $page->date->year; // 2020 echo $page->date->diffForHumans(); // 28 minutes ago Frontend only
      The ProcessWire admin seems to expect datetime fields to be strings. This module will only return Carbon instances on frontend page views.
      Date output format
      When casting a Carbon instance to a string (usually when outputting the field in a template), the field's date output format will be respected.
      Links
      GitHub • Readme • Carbon docs
       
       
      PS. I remember reading about a Carbon module in a recent newsletter, but couldn't find it anywhere. Was that you, @bernhard?
    • By MoritzLost
      TrelloWire
      This is a module that allows you to automatically create Trello cards for ProcessWire pages and update them when the pages are updated. This allows you to setup connected workflows. Card properties and change handling behaviour can be customized through the extensive module configuration. Every action the module performs is hookable, so you can modify when and how cards are created as much as you need to. The module also contains an API-component that makes it easy to make requests to the Trello API and build your own connected ProcessWire-Trello workflows.
      Warning: This module requires ProcessWire 3.0.167 which is above the current stable master release at the moment.
      Features
      All the things the module can do for you without any custom code: Create a new card on Trello whenever a page is added or published (you can select applicable templates). Configure the target board, target list, name and description for new cards. Add default labels and checklists to the card. Update the card whenever the page is updated (optional). When the status of the card changes (published / unpublished, hidden / unhidden, trashed / restored or deleted), move the card to a different list or archive or delete it (configurable). You can extend this through hooks in many ways: Modifiy when and how cards are created. Modify the card properties (Target board & list, title, description, et c.) before they are sent to Trello. Create your own workflows by utilizing an API helper class with many convenient utility methods to access the Trello API directly. Feedback & Future Plans
      Let me know what you think! In particular:
      If you find any bugs report them here or on Github, I'll try to fix them. This module was born out of a use-case for a client project where we manage new form submissions through Trello. I'm not sure how many use-cases there are for this module. If you do use it, tell me about it! The Trello API is pretty extensive, I'll try to add some more helper methods to the TrelloWireApi class (let me know if you need anything in particular). I'll think about how the module can support different workflows that include Twig – talk to me if you have a use-case! Next steps could be a dashboard to manage pages that are connected to a Trello card, or a new section in the settings tab to manage the Trello connection. But it depends on whether there is any interest in this 🙂 Links
      Repository on Github Complete module documentation (getting started, configuration & API documentation) TrelloWire in the modules directory Module configuration

×
×
  • Create New...