[RELEASE] bravo.exe 0.2.3 RC2 | Automatic Captions and Censors

[theredreaper]

But also wanted to mention a few things that make this different from, and still arguably not matching the features of, the hacky-beta.exe program:

- This does not seem to do video. The beta.exe program does video the best and fastest of any "easy" solution I've seen. IE this one is promising but I can't seem to get it to work: https://github.com/8TBHomework/AutoCensorMPV

- Captions outside of image body is imo less interesting. I like the sensory overload (of the dual images too tbh).

- Caption system is much less robust, needing every variation to be its own line. beta.exe spits out near gibberish if configured wrong, but I do like the dynamic nature of it.

- I haven't been able to see the censoring in action because crashes, but I'm guessing this does not do highlighting.

- It appears you can only load one directory at a time.

Video is a lot more difficult to implement with the censoring, as the only way to make it work properly is to run the neural network against every single frame of the video, and the processing takes an incredible amount of time, even with the less than perfect model I'm currently using. It's faster with the model distributed by the authors of nudenet, but misses quite a lot. In general I'd rather not implement a feature than have to implement it in such a way that it introduces performance issues and/or just doesn't work properly. But this is something I'm looking into.

To your point about the captions in the image body and the dual images, I was not a fan of either feature personally. I might add them as options if enough people want them to be available, but as it stands now, they'd be pretty low on my list of priorities.

I'm not sure I'm following what you mean by this system of doing captions being less robust. With the ability to have the user provide their own captions, it's infinitely configurable, where beta.exe's approach was kind of a take it or leave it. If the included captions and their subject matter weren't to your taste, you didn't have a way to do anything about it.

Hopefully the latest release will take care of the issues you've had with the censoring. Not sure what you mean by "highlighting".

[devotes.julchen]

My implementation will load captions from a user-provided text-file. I will probably distribute a "default" file along with it, but it's hard to imagine coming up with any collection of that sort of thing that works for everyone. When it's finished (I'm 75ish percent there by now), I will most likely just open source it and release it to the community, along with the source code, so you will be more than welcome to contribute translations, more captions, etc to the project. I don't intend to do much more after the initial release than whatever bug fixes might become necessary, as I have already been contacted by a couple of the other developers on this forum about getting involved with some of the bigger projects going on here, and am quite excited about that.

Since the captions are in a txt file, the whole thing can be adapted and translated by anyone there. I would be happy if language variables were inserted for the texts used in the software and if language files were added for translation.

The reason this took awhile to get done was that this update added quite a lot of new logic, and therefore quite a lot of code to the project. It seems stable on my machine in its current form, and I wanted to get a release out with these bugfixes and the requested features so far, but I will be moving to an RC2 version very soon, as the code is now in dire need of some refactoring. Let me know how you all get by with this version.

https://drive.google.com/file/d/1A7i1Jx ... sp=sharing

I get this error when starting the slide show.

[theredreaper]

I get this error when starting the slide show.

I'm going to need a little more info on this one. Can you send me a screenshot or something of what is in the directory you've pointed it to for images?

Since the captions are in a txt file, the whole thing can be adapted and translated by anyone there. I would be happy if language variables were inserted for the texts used in the software and if language files were added for translation.

Doing translations on the fly in the software would take quite a lot of work. But you could just cut/paste the entire contents of a captions file into Google translate for your favored language and then save the result back to the captions file.

[Augustulus]

I'm going to need a little more info on this one. Can you send me a screenshot or something of what is in the directory you've pointed it to for images?

Since I'm currently testing and receiving the same error, I'll give you a quick overview.
In fact there are two errors right there. If you don't censor the images, you get the following error once starting the slideshow:

Code: Select all

Traceback (most recent call last):
  File "main.py", line 344, in <module>
  File "main.py", line 246, in next_slide
  File "main.py", line 159, in scale
  File "PIL\Image.py", line 3077, in open
  File "PIL\Image.py", line 548, in __getattr__
AttributeError: read

An unhandled read exception.

Once you censor them, you'll get the error already mentioned by devotes.julchen.

For me the folder I tried it with, solely contained jpg images:

Spoiler: show

bravo_error.jpg (442.13 KiB) Viewed 1182 times

[theredreaper]

Ok, that gives me a clue. It's not able to display .jpg images natively, so they must be converted to .png on the fly. I must have introduced a regression in the function that handles the conversion. I'll get to work on that now and see if I can figure it out.

[theredreaper]

https://drive.google.com/file/d/1YiDhjk ... sp=sharing

Code: Select all

0.1.3 RC1 - 
    - addressed a regression that prevented .jpg images from being loaded correctly.

[devotes.julchen]

https://drive.google.com/file/d/1YiDhjk ... sp=sharing

Code: Select all

0.1.3 RC1 - 
    - addressed a regression that prevented .jpg images from being loaded correctly.

Code: Select all

Traceback (most recent call last):
  File "main.py", line 345, in <module>
  File "main.py", line 239, in next_slide
  File "PIL\Image.py", line 2264, in save
  File "PIL\Image.py", line 626, in _ensure_mutable
  File "PIL\Image.py", line 619, in _copy
  File "PIL\ImageFile.py", line 226, in load
ValueError: I/O operation on closed file.

Unfortunately I still get an error. I tried it with only files with the .jpg extension. Under C:\Users\****\.Nudenet are the two files classes and detector_v2_default_checkpoint.onnx but they are not generated automatically I have to copy them there myself. I haven't received any notification either.

[theredreaper]

Unfortunately I still get an error. I tried it with only files with the .jpg extension. Under C:\Users\****\.Nudenet are the two files classes and detector_v2_default_checkpoint.onnx but they are not generated automatically I have to copy them there myself. I haven't received any notification either.

Yeah, I screwed up the fix with the .jpg files. Sorry about that. It's fixed in this version.

I'm perplexed that you didn't get an error message from the program if it failed to install the detector and classes files. There is a three step process in place for that at this point. The logic basically translates to this:

1) Does directory C:\Users\****\.Nudenet exist?
Yes: Go to step 2
No: Try to create it, if it can't be created, raise an error message.
2) Does file C:\Users\****\.Nudenet\detector_v2_default_checkpoint.onnx exist?
Yes: Go to step 3
No: Try to copy it, if it can't be copied, raise an error message.
3) Does file C:\Users\****\.Nudenet\classes exist?
Yes: Continue execution
No: Try to copy it, if it can't be copied, raise an error message.

So you should have gotten an error message when the program failed to copy the files into the directory. This all works on both of my test systems. Out of curiosity, any of you that are troubleshooting this problem, would you try deleting your .Nudenet folder entirely and then run the program as an administrator? Right click bravo.exe > Run as Administrator

This wouldn't be my permanent solution to the issue, but if that works, (or doesn't) it will give me a clue as to what could be causing this headache for people.

Now on to the update

Moving to RC2 version after a major refactoring.

Patch notes:

Code: Select all

0.2.0 RC2 - 
    - changed the handling of program options and the generation and loading of config files.   program options are now stored in a dictionary and dumped to json data at program exit.   config filename has changed to 'config.json'
    - addressed an issue where all censor options would be loaded regardless of which ones the user had enabled.
    - addressed a small issue with the enabling/disabling of censor options upon enabling/disabling censoring.
    - created a new class for a slideshow object.   A slideshow object is instantiated when the user presses 'Start Slideshow' and garbage collected when a user presses 'Stop Slideshow' or changes any program options.  Changing the image directory, captions file, or any program options will now require the slideshow to be restarted.  This ensures that a user's changes get applied as one would expect, whereas previously, changing certain options would not take effect immediately, leading to an inconsistent and potentially confusing user experience.  overall performance should be improved and program memory usage reduced somewhat.
    - the regression introduced in 0.1.2 preventing .jpg images from being displayed should really be fixed this time.   really.
    - a new entry point to the package was created and logic had to be modified for correctly directing the working directory.

https://drive.google.com/file/d/1QybjiP ... sp=sharing

[Augustulus]

Now on to the update

Moving to RC2 version after a major refactoring.

Patch notes:

Code: Select all

0.2.0 RC2 - 
    - changed the handling of program options and the generation and loading of config files.   program options are now stored in a dictionary and dumped to json data at program exit.   config filename has changed to 'config.json'
    - addressed an issue where all censor options would be loaded regardless of which ones the user had enabled.
    - addressed a small issue with the enabling/disabling of censor options upon enabling/disabling censoring.
    - created a new class for a slideshow object.   A slideshow object is instantiated when the user presses 'Start Slideshow' and garbage collected when a user presses 'Stop Slideshow' or changes any program options.  Changing the image directory, captions file, or any program options will now require the slideshow to be restarted.  This ensures that a user's changes get applied as one would expect, whereas previously, changing certain options would not take effect immediately, leading to an inconsistent and potentially confusing user experience.  overall performance should be improved and program memory usage reduced somewhat.
    - the regression introduced in 0.1.2 preventing .jpg images from being displayed should really be fixed this time.   really.
    - a new entry point to the package was created and logic had to be modified for correctly directing the working directory.

https://drive.google.com/file/d/1QybjiP ... sp=sharing

Displaying uncensored jpgs now definitely works, but I get a conversion error between string and float now once I want to censor the images.
The big problem is that I can't reproduce the error for you, because of another bug: the program doesn't start at all .
For the first time I get a config.json generated inside the folder, but on start up, the program tries to load the config.json from my Users temporary folder and thus returns a FileNotFoundError:

Code: Select all

Traceback (most recent call last):
  File "cli.py", line 1, in <module>
  File "<frozen importlib._bootstrap>", line 991, in _find_and_load
  File "<frozen importlib._bootstrap>", line 975, in _find_and_load_unlocked
  File "<frozen importlib._bootstrap>", line 671, in _load_unlocked
  File "PyInstaller\loader\pyimod03_importers.py", line 495, in exec_module
  File "bravo\__init__.py", line 1, in <module>
  File "<frozen importlib._bootstrap>", line 991, in _find_and_load
  File "<frozen importlib._bootstrap>", line 975, in _find_and_load_unlocked
  File "<frozen importlib._bootstrap>", line 671, in _load_unlocked
  File "PyInstaller\loader\pyimod03_importers.py", line 495, in exec_module
  File "bravo\main.py", line 43, in <module>
  File "bravo\options.py", line 26, in load_options
FileNotFoundError: [Errno 2] No such file or directory: '<User path>\\AppData\\Local\\Temp\\_MEI72802\\config.json'

Once this is fixed I can provide you more insight into the censoring error.

[devotes.julchen]

I'm perplexed that you didn't get an error message from the program if it failed to install the detector and classes files. There is a three step process in place for that at this point.

1) Does directory C:\Users\****\.Nudenet exist?
Yes: Go to step 2
No: Try to create it, if it can't be created, raise an error message.
2) Does file C:\Users\****\.Nudenet\detector_v2_default_checkpoint.onnx exist?
Yes: Go to step 3
No: Try to copy it, if it can't be copied, raise an error message.
3) Does file C:\Users\****\.Nudenet\classes exist?
Yes: Continue execution
No: Try to copy it, if it can't be copied, raise an error message.

So you should have gotten an error message when the program failed to copy the files into the directory. This all works on both of my test systems. Out of curiosity, any of you that are troubleshooting this problem, would you try deleting your .Nudenet folder entirely and then run the program as an administrator? Right click bravo.exe > Run as Administrator

So I tested the whole thing again in the 0.2.0 RC2 version, I deleted the .Nudenet folder beforehand. He now creates the .Nudenet folder and the two corresponding files (classes and detector_v2_default_checkpoint.onnx). Namely as an administrator, but also when I run the file normally. I then tested when I delete the individual files, whether they are then created again or whether I then get a corresponding note. However, this is not the case.

I start a slideshow without settings. This is now possible without any problems. However, as soon as I change the settings in the options, I get the following error:

Code: Select all

Traceback (most recent call last):
  File "cli.py", line 4, in <module>
  File "bravo\main.py", line 141, in main
  File "bravo\slideshow.py", line 162, in start
ValueError: could not convert string to float: ''

After you have exited the program for the first time, the config.json is created in the main directory. However, if you run the program again after this, you get the following error:

Code: Select all

Traceback (most recent call last):
  File "cli.py", line 1, in <module>
  File "<frozen importlib._bootstrap>", line 991, in _find_and_load
  File "<frozen importlib._bootstrap>", line 975, in _find_and_load_unlocked
  File "<frozen importlib._bootstrap>", line 671, in _load_unlocked
  File "PyInstaller\loader\pyimod03_importers.py", line 495, in exec_module
  File "bravo\__init__.py", line 1, in <module>
  File "<frozen importlib._bootstrap>", line 991, in _find_and_load
  File "<frozen importlib._bootstrap>", line 975, in _find_and_load_unlocked
  File "<frozen importlib._bootstrap>", line 671, in _load_unlocked
  File "PyInstaller\loader\pyimod03_importers.py", line 495, in exec_module
  File "bravo\main.py", line 43, in <module>
  File "bravo\options.py", line 26, in load_options
FileNotFoundError: [Errno 2] No such file or directory: 'C:\\Users\\****\\AppData\\Local\\Temp\\_MEI104202\\config.json'

At the time there was no config.json in the directory C:\Users\****\AppData\Local\Temp\_MEI170402

[theredreaper]

Discovered a couple of issues last night and got them patched this morning.

Code: Select all

0.2.1 RC2 -
    - addressed a bug that caused a program crash when using .jpg images when trying to do the conversion procedure on the last image in the list.
    - some temporary files created during execution are now cleaned up on application exit.
    - addressed a regression introduced in the 0.2.0 release that caused the application to look in the wrong directory for the config file.
    - addressed a regression introduced in the 0.2.0 release that could cause a program crash when starting a new slideshow after setting a user defined advance timer.

https://drive.google.com/file/d/1PN26kI ... sp=sharing

[theredreaper]

...

At the time there was no config.json in the directory C:\Users\****\AppData\Local\Temp\_MEI170402

Thank you again for you help with tracking down bugs. Both of the issues you mentioned should be corrected in the latest release.

[theredreaper]

Displaying uncensored jpgs now definitely works, but I get a conversion error between string and float now once I want to censor the images.
The big problem is that I can't reproduce the error for you, because of another bug: the program doesn't start at all .
For the first time I get a config.json generated inside the folder, but on start up, the program tries to load the config.json from my Users temporary folder and thus returns a FileNotFoundError:


Once this is fixed I can provide you more insight into the censoring error.

I believe both of these issues should be fixed in the latest release.

[wqslave]

I'm not sure I'm following what you mean by this system of doing captions being less robust. With the ability to have the user provide their own captions, it's infinitely configurable, where beta.exe's approach was kind of a take it or leave it. If the included captions and their subject matter weren't to your taste, you didn't have a way to do anything about it.

Beta.exe's approach is infinitely configurable, it just also works with a noun/adjective system, also with categories. So it looks at everything you've provided for the nouns/adjectives/names/ect, and each base sentence, and creates captions with regex replacements. So to make a caption there, you create a template sentence with spots for a bunch of parts that are put together at runtime. To do the same here you'd need to input each potential combination individually, which can quickly add-up to several thousand lines. It's not like a huge issue or anything that makes this unusable or something, it just isn't as robust.

Hopefully the latest release will take care of the issues you've had with the censoring. Not sure what you mean by "highlighting".

Highlighting is an alternative thing that beta.exe and some other censor-related softwares do where they put a box [] or circle () ect around a part. Kind of like a very basic identification box, but without a label. Beta.exe utilizes multiple censor styles (blur/black box/pixel) and then also highlight. It's additional sensory overload and/or to highlight "where you should be looking."

[devotes.julchen]

Thank you again for you help with tracking down bugs. Both of the issues you mentioned should be corrected in the latest release.

No problem. The first problem has been fixed, the second is still there.

Code: Select all

Traceback (most recent call last):
  File "cli.py", line 1, in <module>
  File "<frozen importlib._bootstrap>", line 991, in _find_and_load
  File "<frozen importlib._bootstrap>", line 975, in _find_and_load_unlocked
  File "<frozen importlib._bootstrap>", line 671, in _load_unlocked
  File "PyInstaller\loader\pyimod03_importers.py", line 495, in exec_module
  File "bravo\__init__.py", line 1, in <module>
  File "<frozen importlib._bootstrap>", line 991, in _find_and_load
  File "<frozen importlib._bootstrap>", line 975, in _find_and_load_unlocked
  File "<frozen importlib._bootstrap>", line 671, in _load_unlocked
  File "PyInstaller\loader\pyimod03_importers.py", line 495, in exec_module
  File "bravo\main.py", line 40, in <module>
  File "bravo\options.py", line 26, in load_options
FileNotFoundError: [Errno 2] No such file or directory: 'C:\\Users\\****\\AppData\\Local\\Temp\\_MEI147522\\config.json'