ripose-jp / memento

An mpv-based video player for studying Japanese

Memento

Memento is a FOSS, mpv-based video player for studying Japanese.

Features

• Grammar aware subtitle search
• Yomichan-style Kanji cards
• Support for Yomichan dictionaries
• Anki card creation through AnkiConnect
• Support for mpv upscalers, plugins, and configuration files

Dictionaries

• JMdict (Japanese vocabulary)
  • JMdict_dutch.zip
  • JMdict_english.zip
  • JMdict_french.zip
  • JMdict_german.zip
  • JMdict_hungarian.zip
  • JMdict_russian.zip
  • JMdict_slovenian.zip
  • JMdict_spanish.zip
  • JMdict_swedish.zip
• JMnedict (Japanese names)
  • JMnedict.zip
• KireiCake (Japanese slang)
  • kireicake.zip
• KANJIDIC (Japanese kanji)
  • KANJIDIC_english.zip
  • KANJIDIC_french.zip
  • KANJIDIC_portuguese.zip
  • KANJIDIC_spanish.zip
• Innocent Corpus (Term and kanji frequencies across 5000+ novels)
  • innocent_corpus.zip
• Kanjium (Pitch dictionary, see related project page for details)
  • kanjium_pitch_accents.zip

Troubleshooting

Can't Add Cards: Only the "Show in Anki" Button is Available

This means there is a mistake in your card template. When this happens, AnkiConnect will report that all potential cards are not addable. Double check your card template to see if the front of the card is missing something.

Secondary Subtitles Don't Work

Prior to mpv v0.35.0, it was impossible to set the visibility of primary and secondary subtitles independent of one another. If you haven't compiled Memento yourself and are using the provided binaries for v1.0.0 or later, this section doesn't apply to you.

Method 1

Go to Settings → Options → Search, uncheck "Hide mpv subtitles when subtitle search is visible", and check "Hide subtitle search when playing media".

If the searchable subtitles become harder to read, consider adding a background. This can be done in Interface settings. Make sure you set the alpha/opacity channel to 255 in order to avoid your background being transparent.

Method 2

When the Memento is paused, manually toggle subtitle visibility to reveal the secondary subtitle. Subtitle visibility is bound to v by default.

Windows: MSVCR100.dll is Missing When Trying to Stream

This error can be fixed by installing the Microsoft Visual C++ 2010 Service Pack 1 Redistributable Package (x86).

Windows: Updating youtube-dl/yt-dlp

If you're version of youtube-dl is out of date, you may experience degraded streaming performance or websites not working entirely.

Memento's version of youtube-dl can be updated by doing the following:

1. Download yt-dlp
2. Rename the file to youtube-dl.exe
3. Put youtube-dl.exe in Memento's install directory. For the portable version of Memento, this is located in the same folder as the executable. For the installed version, this is located at C:\Program Files\Memento by default.

macOS: Streaming video doesn't work

This means that mpv cannot find your youtube-dl installation.

To install youtube-dl, paste these commands into Terminal:

curl -L https://github.com/yt-dlp/yt-dlp/releases/latest/download/yt-dlp -o ~/Library/Preferences/memento/youtube-dl
chmod a+rx ~/Library/Preferences/memento/youtube-dl

macOS: Could not initialize MeCab

Move the Memento application to a directory that doesn't have spaces in the path like /Applications. Unfortunately, this is an limitation of MeCab and there is no workaround that can be implemented.

Linux: Audio and images aren't added to Anki

This only applies to users of the Anki Flatpak.

By default the Anki Flatpak doesn't have access to the /tmp directory. Memento stores audio and image files here temporarily when making them available to Anki. You can give Anki access to the /tmp directory by installing Flatseal. To do so, open Flatseal, select Anki, and add the /tmp directory under Other files.

Dependencies

• Qt
  • Base
  • SVG
• mpv
• sqlite3
• MeCab
  • With either ipadic or NAIST-jdic installed as a system dictionary on Linux and macOS. This only applies to self-compiled versions, not appimages or app bundles.
• Json-C
• libzip
• youtube-dl or yt-dlp (optional)
• Python (optional)

For the best experience, install Noto Sans JP and the Kanji Stroke Order fonts.

Building

I do not guarantee that any branch will successfully build or be bug-free. If you want to build a stable version of Memento, compile a release from source.

Linux

To install Memento on Linux, type the following commands:

make
sudo make install

Windows

1. Install MSYS2
2. Open MSYS2 MinGW 64-bit
3. Make sure MSYS2 is up to date by running this command multiple times until it stops doing anything:

   pacman -Syuu
   

4. Install the necessary tools and dependencies:

   pacman -S git make mingw-w64-x86_64-gcc mingw-w64-x86_64-ninja mingw-w64-x86_64-cmake mingw-w64-x86_64-python mingw-w64-x86_64-sqlite3 mingw-w64-x86_64-qt6 mingw-w64-x86_64-mpv mingw-w64-x86_64-mecab mingw-w64-x86_64-json-c mingw-w64-x86_64-libzip
   

5. Clone the repository:

   git clone https://github.com/ripose-jp/Memento.git
   

6. Build Memento:

   cd Memento
   ./windows/build.sh x86_64
   

7. The resulting file will be in

   build/Memento_x86_64
   

macOS

Important: Clang is the only officially supported compiler for building Memento on macOS.

1. Install Homebrew with this command:

   /bin/bash -c "$(curl -fsSL https://raw.githubusercontent.com/Homebrew/install/HEAD/install.sh)"
   

2. Install the necessary tools and dependencies:

   brew install git cmake sqlite3 qt6 mpv mecab mecab-ipadic json-c libzip
   

3. Clone the repository:

   git clone https://github.com/ripose-jp/Memento.git
   

4. Build Memento:

   cd Memento
   make
   

5. The resulting executable will be:

   Memento/build/src/Memento
   

macOS App Bundle

1. Follow steps 1 - 3 of the macOS build instructions.

2. dylibbundler is also needed when creating an app bundle, so install it with:

brew install dylibbundler

1. Open the Keychain Access app.

2. Go to Keychain Access > Certificate Assistant > Create a Certificate... in the menubar.

3. Put the name of your certificate in the 'Name' field, set the 'Certificate Type' to 'Code Signing', and click 'Create'.

4. Return to your terminal and input:

   cd Memento
   make appbundle CERT_NAME='<name entered in the last step>'
   

5. The resulting app bundle will located at:

   Memento/build/src/Memento.app
   

Adding OCR Support

To build with OCR support, make sure that Python is installed and run:

pip install manga-ocr

Any problems you may have getting manga-ocr installed using pip is beyond the scope of this project. I wish you the best of luck.

Add -DOCR_SUPPORT=ON to the CMAKE_ARGS environment variable:

export CMAKE_ARGS='-DOCR_SUPPORT=ON'

From here follow normal build instructions for your platform.

Note: Theoretically OCR is supported on Windows. Assuming Memento was built against msys2's version of Python, you will have to set the environment variable PYTHONHOME to C:\msys64\mingw64.

Configuration

Most mpv shaders, plugins, and configuration files will work without modification.

It is important to note that mpv and Memento's configuration directories are separate. This means mpv configuration files, scripts, etc. intended to modify Memento's behavior should be placed in Memento's configuration directory.

The Memento configuration directory is located at:

Linux

~/.config/memento

Windows

Version 0.5.1 and later

%APPDATA%\Local\memento

Version 0.5.0-1 and earlier

[installation directory]\config

macOS

~/Library/Preferences/memento

If any mpv binds or plugins do not work, please create an issue in the issue tracker.

Contributing

Before making a pull request, please read CONTRIBUTING.md.

Animebook

If you hate Memento, try Animebook.

https://github.com/animebook/animebook.github.io

Acknowledgements

• MpvWidget code is based off of libmpv example code by w4m
  • https://github.com/mpv-player/mpv-examples/tree/master/libmpv/qt_opengl
• Some code based off of Baka-MPlayer
  • https://github.com/u8sand/Baka-MPlayer
• Hardware acceleration achieved on Linux using mpc-qt code
  • https://github.com/mpc-qt/mpc-qt
• UI inspired by and dictionaries provided by Yomichan.
  • https://foosoft.net/projects/yomichan/
• OCR backend written kha-white and contributors
  • https://github.com/kha-white/manga-ocr
• Fullscreen icons sourced from here
  • https://www.iconfinder.com/iconsets/material-core
• Various icons sourced from here
  • https://www.iconfinder.com/iconsets/ionicons
• Pause, play, skip, and seek icons sourced from here and used under the CC 3.0 License
  • https://www.iconfinder.com/iconsets/play-rounded
• mingw-bundledlls.py based on this script with a slightly expanded blacklist
  • https://github.com/mpreisler/mingw-bundledlls
• Flatpak manifest heavily based on the mpv Flatpak
  • https://github.com/flathub/io.mpv.Mpv

memento's Issues

make it so that you can change the position, size and font of the scannable sub-line

make it so that the sub line you can scan can be modified like the standard subtitle line can in mpv ie change size, position and font. would also be cool if those settings could be saved as default so you dont have to change it each time you open the program

Typo in uninstaller

When you are going to uninstall Memento in windows and it asks for confirmation the message says "permanantly uninstall Memento?" and it should be "Permanently uninstall Memento?".

Although I think another question would be better, like "Are you sure you want to permanently remove Memento?" or something like that.

Screenshot:

Add the ability to have the audio of multiple subtitle lines.

Sometimes just the current subtitle line is not enough to give context to the word you want to mine, so it would be good to have a way of selecting multiple adjacent subtitle lines, maybe in the subtitle browser and then when you click the button to mine it records the audio of the selected subtitle lines and add them to the sentence field.

Possible selection of more characters?

Currently there is a problem that I can't select say all string as sentence. And sometimes even Kanji+Kana won't work, for example 楽しい, I can select only Kanji, or only Kana which is pointless.

Accept mpv-style subtitle argument from command line

I usually start my video player from the command line. With mpv, i can specify the right subtitle file with --sub-file=<sub-file>. I know i can still add subtitle files in memento, but it takes a few clicks in the gui with every new video file.

It would be nice if memento could also accept --sub-file(https://mpv.io/manual/master/#options-sub-file) where the given subtitle file is appended to the sub list.

This is all i really want, but as a side idea, it could be interesting to use ALL given command line options and pass them along to mpv. This would allow for some pretty powerful usage ideas.

Using arrow keys for skipping forward and backwards is slightly broken

Upon pressing either of the left/right arrow keys the video goes forwards or backwards twice, aka 10 seconds rather than 5.

Rather than fixing it so it goes forward once I think it would be better to skip to the previous or next subtitle, to add to this, also have the enter key replay the current subtitle.

Create chocolatey installer package

Hey,
what a useful video player you have here! Exactly what I was looking for 👏🏻.
I maintain some packages for Chocolatey, so I created one for memento.

https://community.chocolatey.org/packages/memento.portable
(The package has to undergo a moderation process now.)

If you could provide an installer binary, I could also create and maintain packages memento (metapackage) and memento.install.

why dont you publish this in AUR?

for easy access to arch linux users. please publish this on AUR.

yomichan audio?

can't i play audio via yomichan inside memento?

Shift select behavior bug

Selecting continuous blocks of subtitles from the subtitle list with shift click has unexpected behavior in some scenarios.

Subtitles in question:

892
00:59:21,015 --> 00:59:24,477 
あとは どうやって忘れるかだけだ

893
00:59:24,561 --> 00:59:24,727 
（せきばらい）

894
00:59:24,727 --> 00:59:26,104 
（せきばらい）

895
00:59:24,727 --> 00:59:26,104 
役に立たなかったでは
終えられんしな

896
00:59:26,104 --> 00:59:27,146 
役に立たなかったでは
終えられんしな

897
00:59:27,272 --> 00:59:31,818 
し… しかし 今回の計画は
各国の注目を集めており―

898
00:59:31,943 --> 00:59:35,154 
共和国側も
我々の情報を欲しています

899
00:59:35,280 --> 00:59:38,825 
それを 敵の属国リマダとの
国境で打ち上げる…

900
00:59:38,950 --> 00:59:42,370 
だから そうするんだ
打ち上げなぞは どうでもいい

901
00:59:42,495 --> 00:59:45,540 
目当ては 共和国のリマダ駐屯軍だ

Other instances I have observed also have consecutive subtitle entries that start and end at the same time while having the same text content. Perhaps this is causing problems with the hash map from subtitle text to timing information?

Unexpected behaviour while in full screen mode

I've been having an issue since I installed Memento on a new machine a few days ago. Whenever I watch I watch a video and enter full screen mode, the screen will flash black for a second and jump around for a bit and starts behaving erratically. The full screen playback acts as if it is on top of everything else (e.g. the windows popup volume indication when changing system volume is invisible) and if I manually take a screenshot, it will always be from the moment I first entered full screen rather than the frame I took the screenshot of. Similarly, when I start recording my screen when playing a video in windowed mode and switching into full screen, the audio will continue playing normally but the video recording is stuck in the frame when I entered full screen.

I've recorded the following video to demonstrate: First I enter and leave full screen twice to demonstrate the flashes and jumping around, then I watch the same scene in windowed mode and then in full screen mode. For some reason, the full screen playback is a still frame in the recording, even though it displayed fine on my end: https://files.catbox.moe/tzh2xl.mp4

This only occurs in Memento, the latest version of regular mpv work perfectly fine.

Add Yomichan-style search for creating cards

I hate to be the one asking for feature creep, but here is something to consider. Currently, when looking up a word such as追いはぎ, 持ちこたえる, 叩きだす, etc, where the first part is in the conjunctive form and the second part is in kana, Memento is not able to look up the whole word but separates it into two (at least with my set of dictionaries). After editing the subtitle to rewrite the second part of the word in kanji also (追い剥ぎ、持ち堪える、叩き出す), Memento is able to look up the whole word.

I think this is a difficult problem and Yomichan does not handle this case well either. But there have been other cases where the word form is too weird for Memento to recognize (one I can think of is 混とん vs 混沌). I wonder if any of the following are possible:

• allow the user to select part of the subtitle to look up as the headword as a whole (I doubt this is what causes the issue in the example; I ran these words through Mecab and each is recognized as an individual token)

• have an interface that the user could open & manually input a word (in the dictionary form that Memento should be able to recognize) to look up with Memento’s dictionary, and be able to add the result to Anki with the relevant screenshot and context. This seems like a catch-all solution for these scenarios that don’t happen very often.

Audio padding

Sometimes the audio gets slightly cut off when exporting a card to anki would be nice if the plugin had some padding on either side to minimize this issue

Transparent subtitle bug introduced in 4bb801f

In versions since 4bb801f, the search (subtitle) widget is transparent (but still searchable on hover). I tested the three previous versions which all display correctly. I don't understand why it is caused in that commit though.

Duplicate cards even though duplicate policy is marked as "none"

The program lets you create a new card of a word you already have a card for (tho you havent reviewed it yet) and theres no warning shown on your yomichan, unlike the browser version. On options > anki integration > duplicate policy I marked "none" btw.
Other than that everything looks pretty good.

OBS: idk if I missed anything, if I did something dumb I'll delete this issue

25-30% CPU usage when idle with subs on the screen

When video is stopped and no subtitles on the screen CPU usage is 0-0.2%, and when subtitles are on the screen CPU usage jumps to constant 25-30%. Other UI elements(menu bar and video controls) don't affect CPU usage much(1.5-2% maximum and only for a moment). I don't know what happens, maybe subtitles are being redrawn constantly(but why if nothing changes?).

100% reproducibility.
OS: Windows 10
Memento version: 0.5.1

Can't build with CMake versions older than 3.13

install(TARGETS) only supports installing targets created in the same directory on CMake releases older than 3.13. Memento's minimum required cmake version is 3.1.

see: https://stackoverflow.com/questions/34443128/cmake-install-targets-in-subdirectories

Ubuntu 18.04 LTS still ships with CMake 3.10.

Add option to resize screenshots before adding to anki

The user should be able to resize mpv screenshots, because otherwise images might be too big on anki cards especially when watching something in fullscreen.

Locally this was pretty easy to implement, starting on line 717 in mpvadapter.cpp, I added

    QImage img(filename);
    img = img.scaled(1000,300,Qt::KeepAspectRatio)
    img.save(filename)

Unforunately I'm not very good at qt or c++, so I don't know what the best way to add options to the settings ui is, but I think it would be a useful feature to be able to customize the screenshot size; and maybe an option to choose to preserve aspect ratio to height or width.

Display all subtitle entries in the subtitle list?

Currently it seems to only add entries to the list as the video is being played. Would it be possible to be able to browse the subtitles for the entire video?

Much thanks!

Unnecessary quotation marks in Memento's version

If you install Memento through the installer when you go to the Programs and Features on Windows 10 to uninstall the program the version number has quotation marks unlike the other programs.

Screenshot, Git included for comparison:

Showing timestamp

add the option to show timestamp on the left subtitles list visibility.

Cannot build on Windows 10

Anon@Navi MSYS ~/Memento
$ ./windows-build.sh x86_64
mkdir: cannot create directory ‘build’: File exists
CMake Error: Could not create named generator MSYS Makefiles

Generators
* Unix Makefiles               = Generates standard UNIX makefiles.
  Ninja                        = Generates build.ninja files.
  Ninja Multi-Config           = Generates build-<Config>.ninja files.
  CodeBlocks - Ninja           = Generates CodeBlocks project files.
  CodeBlocks - Unix Makefiles  = Generates CodeBlocks project files.
  CodeLite - Ninja             = Generates CodeLite project files.
  CodeLite - Unix Makefiles    = Generates CodeLite project files.
  Eclipse CDT4 - Ninja         = Generates Eclipse CDT 4.0 project files.
  Eclipse CDT4 - Unix Makefiles= Generates Eclipse CDT 4.0 project files.
  Kate - Ninja                 = Generates Kate project files.
  Kate - Unix Makefiles        = Generates Kate project files.
  Sublime Text 2 - Ninja       = Generates Sublime Text 2 project files.
  Sublime Text 2 - Unix Makefiles
                               = Generates Sublime Text 2 project files.

Error: could not load cache
mkdir: cannot create directory ‘Memento_x86_64’: File exists
cp: cannot stat 'src/memento.exe': No such file or directory
./windows-build.sh: line 147: windeployqt: command not found

Followed the tutorial and it's not working... any suggestion?

Unexpected mpv config directory on MacOS

On macOS, the input.conf at ~/.config/mpv/input.conf apparently is not read. This can be solved by copying it to Memento's own config directory at ~/Library/Preferences/memento. I have not yet tested the case with mpv.conf.

The behavior is certainly unexpected, though perhaps it can be beneficial in allowing bespoke mpv configurations for Memento. In any case, we probably should note it somewhere?

Thanks for this really cool work!

Discovered while checking the following issue:

Nice catch with the double seek, I'll fix it in the next release.

As for changing the default keybinds, I'd rather not so Memento remains functionally as close to mpv as possible.

You can implement the behavior you want pretty easily by adding these lines to your input.conf.

ENTER   sub-seek        0
RIGHT   sub-seek        1
LEFT    sub-seek        -1

Check the README to find the location of the config directory for your operating system.

These links will be helpful for finding out what the default binds are and how to configure you own custom binds:
https://mpv.io/manual/stable/
https://github.com/mpv-player/mpv/blob/master/etc/input.conf

Originally posted by @ripose-jp in #3 (comment)

Option to normalize {audio-media} or {audio-context} volume against {audio}

Might be a nice feature to have, but I can understand if it is difficult to implement without introducing more dependencies or cause performance issues.

Video not showing, just a blank screen.

I initially used the program in another computer since the one I normally use was broken, I had it fixed some days ago and I installed the program in it and installed dictionaries but when I went to play the video it didn't show anything just a blank screen, the audio plays, the subs show but the video doesn't show up:

And it happens with every video. When I play the video in mpv it doesn't happen though:

I tried reinstalling the program but the same happens.

How do I uninstall Memento on Linux?

I found about Memento today and tried it out. But I can't uninstall it. I installed it via Github following your instructions.

audio padding settings arent saved on exit

when using the app and setting the audio padding on reopen the settings are reset to their default paddings

Double subtitles now displaying

When selecting second row of subtitles they don't display. Tried different files but issue persists.
Win10 21H1, latest memento.

A way to stop showing UI on hover over?

It's rather jarring to see the seeking menu pull up every time I'm trying to hover over a word.

OS X Port Bugs

This is an issue intended to list all the problems with Memento on OS X so that they can easily be fixed at a later date or by someone who wants to fix them.

I was able to get Memento to compile in VM, but because the VM doesn't support video acceleration or audio, so there is really no feasible way for me to develop or test Memento on OS X.

Bugs

• Fullscreen is broken

  • The way Memento handles fullscreen is that it waits for the playerFullscreenChanged signal to be omitted and adjusts the window size of Memento accordingly. The window resizing is all handled in MainWindow.

• The name of Memento in the OS X menu bar is lowercase

  • This can probably be fixed by changing the name of Memento in main.cpp. If this is changed, it should be in a preprocessor directive specific to OS X or it will break configurations on other OSes. Didn't work. The name in the menu bar is likely tied to the name of the executable file.

• The startup message box points OS X users to the wrong location for adding dictionaries

  • OS X manages to move things in the menu bar around automatically to make their OS consistent. This message can be changed in MainWindow's showEvent method.

• OS X will sleep on inactivity (Implemented not tested)

  • If the player is unpaused, it will not prevent the screen from dimming or going to sleep. This behavior is implemented on other OSes in MpvWidget. The import on line 37 was intended to be used to implement this, but I'm not entirely certain how to do this on OS X, so I didn't implement it.

If anyone has any more bugs to contribute, post them below and I'll see if I can give any insight on how to fix them.

mpvacious .lua script along side memento?

is it possible to use mpvacious with memento?.

Add dark mode to the subtitle list visibitly.

Where can I find the settings?

No dictionaries are installed. For subtitle searching to work, please install a dictionary.
Dictionaries can be found here.
To install a dictionary, go to Settings -> Options -> Dictionaries.

But where can I find settings? I assume it is the hamburger menu in the bottom-right corner, but nothing happens when I click it.
I'm on Linux, Manjaro with KDE.

More info: I installed the latest release 0.4.4 with

make
sudo make install

Allow dictionaries whose entries are not added to Anki by default

Seeing that recent commits refactor dictionary related code and add the ability to disable dictionaries, I wonder if it might be an good time to bring this up. Currently, Memento allows individual dictionary lookup results to be optionally toggled off so that they are not added to Anki. It might be nice if in dictionary settings, certain dictionaries can be configured with the checkbox off by default. This applies when:

• the user has many dictionaries loaded and would like to reference all of them, but only add the top result(s) to Anki
• the user uses special dictionaries whose results should not be added to Anki. One I know of is [補足] 漢字遣い参考, which only contains different forms of words (押し殺す is listed as おしころす【圧し殺す・押し殺す・押殺す・圧しころす・押しころす】) to address the scenario originally described in #47.

Error adding dictionaries

Tried several dictionaries zip files I already used for Anki
But still receiving same error

Error adding dictionaries
Could not open database

Thanks

Can't take screenshot

Manual selection in subtitle

Sometimes MeCab will not split up the words correctly, but there is no easy way to lookup the correct word (unless you open the Term search window and manually type it).
It would be handy if one could select any part of the sentence and look it up.

Search through subtitles

Search for words in the subtitles with a search bar.

For example, if I want to know how many "僕" there are in an episode, I will look it up via the search bar and jump straight to it.

error while building the soruce code of Memento on Arch linux.

terminal log:
mkdir -p build
cd build; cmake -DCMAKE_BUILD_TYPE=Release ..
-- Configuring done
CMake Error: The following variables are used in this project, but they are set to NOTFOUND.
Please set them or make sure they are set and tested correctly in the CMake files:
LIBZIP
linked by target "dictionary_db" in directory /home/me/Downloads/Memento-0.5.0-1-beta/src/dict
MECAB
linked by target "dictionary_db" in directory /home/me/Downloads/Memento-0.5.0-1-beta/src/dict

-- Generating done
CMake Generate step failed. Build files cannot be regenerated correctly.
make: *** [Makefile:4: release] Error 1

Unable to open media from URL's

When trying to open media from a URL (I tried with both Youtube and TVer.jp which works with MPV player) I was unable to open any media in the player. The screen is just black. Opening from local files is fine and has no problems.
It says It's unable to find MSVCR100.dll

Exact error message:
MSVCR100.dllが見つからないため、コードの実行を続行できません。プログラムを再インストールすると、この間題が解決する可能性があります。

I downloaded using the stable version on the release page, I did not build from source.

Character isn't highlighted if the cursor is on the right side of a character

Currently, if your cursor is in the right ~60% of a certain glyph, that glyph isn't highlighted and looked up. Here's a short video to explain what I mean.

I think this is an important thing to change because of how central the text selection is to the program. Currently the most reliable way to select a word is to move the cursor to the very left of the character, or the gap between the character and the next one to the left. You should be able to move your cursor directly to the character, this is how yomichan works on the browser: when the cursor is even slightly on the next glyph, it's selected.

I think this is different to latin-script highlighting, where the current behavior makes sense; characters are much more discrete in CJK.

I know this is really nitpicky, so if this is a can't-fix or is just too much work then please close the issue.

Looking up words on this subtitle causes segmentation fault

Attempting to look up any part of the following subtitle immediately causes a segmentation fault on macOS 0.5.1-2.

1
00:00:00,000 --> 00:00:05,000
〝民間南極観測隊
３年ぶりに派遣決まる〞

The payload is attached here (.srt file before changing extension): test.txt. The timing has been changed from the original.

This happens with my custom dictionaries loaded, no dictionaries loaded, and just jmdict_english loaded.

Second subtitle track feature not working on latest release (v0.4.6-alpha) on Windows 10

What title says... I checked on the older version i had installed (v0.4.4-alpha) and the feature still works in that version. I haven't tested in linux.

Ability to use secondary sub as Anki export

The ability to export the secondary sub file from the player as an anki export option would be nice

Maintain subtitle style with .ass subtitles

I wish there was an option of maintaining the style of .ass subtitles instead of the player overriding it, there are color-coded subtitles that look so beautiful but that is lost when using the player's subtitles instead of mpv's ones.

Memento's subtitles:

MPV's subtitles:

allow creating cards through automated subs

It's not always 100% correct, but Japanese videos usually has harded subs and those subs usually come at the right timing. All you gotta do is finish the card yourself by editing the card with a screenshot of the hard sub or something.

Sometimes fails to populate {context}, {audio-media}, or {screenshot-video} for Anki card

Apologies that I have been unable to reproduce any error on a consistent basis. When playing back local media, deleting the card and adding it again in Memento usually works.

Empty {context}

The Anki card field populated with {context} is entirely blank when this happens. I think I noticed once when this happened that no subtitle seemed to be visually selected in the subtitle list when the card was being added. But as I have not been able to reproduce exactly this behavior I cannot say this certainly happened.

Error with {audio-media}

The Anki card field populated with {audio-media} can have values like this:

[Errno 2] No such file or directory: '/private/var/folders/31/xkn6lwrs7_382lqhy5f9wzbc0000gn/T/memento.apAyPt.aac'

[Errno 2] No such file or directory: '/private/var/folders/31/xkn6lwrs7_382lqhy5f9wzbc0000gn/T/memento.lmvQIY.aac'

or be entirely blank.

Such error values can be consistently reproduced by viewing this YouTube video and adding vocab in the Japanese subtitle to Anki. However I feel it may not be entirely the same problem experienced when playing local media, where it only fails intermittently.

Edit: include {screenshot-video} error

Error with {screenshot-video}

The Anki card field populated with {audio-media} can have a value like this:

[Errno 2] No such file or directory: '/private/var/folders/31/xkn6lwrs7_382lqhy5f9wzbc0000gn/T/memento.ddcjlh.jpg'

I have not encountered a case where the field is empty. This happens rarely but seemingly randomly when playing local media.

Platform: macOS 0.5.1-beta

Option to separately (de)select {audio-context}

Currently, selected lines of subtitles can have their text and audio exported to Anki with {context} and {audio-context} respectively. But there are cases where having less audio than text is preferable.

Consider a line at the end of an exchange, whose context is provided by the lines preceding it. If a lot of lines preceding it have to be selected, then {audio-context} exports a long segment of unrelated audio before getting to the vocab itself, which slows down the pacing of reviews. A more ideal option might be to keep the all the text while only exporting the audio of the relevant line (or fewer lines).

I think it could be nice if there is an option to enable a checkbox for each (selected) line in the subtitle list. When the user selects the line, they are checked by default, but the user has the option to uncheck lines to remove them from audio-context.

Clicking on subs lags a lot the application

First of all, very nice initiative of coming up with this. It will help a lot of people.

I didn't quite get if it's a hover thing or a click thing, but either way it lags the program A LOT.
It doesn't show up properly and when it does it's a blink of a second.
I had to keep my cursor on top of it for it to show and took a few seconds which is bothersome.

Suggestion:
Hold some key and hover on the word for it to appear.
Or simple click on it.

[Linux] Can't install dictionaries: "Could not open database"

I can't seem to open any dictionaries, even jmdict_english.zip. I get "Could not open database."
I am on arch linux, and i built the source from 0.4.7-alpha from the makefile.

Running the program from the command line gives nothing out of the ordinary other than [file] Cannot open file '/home/<name>/.config/memento//input.conf': No such file or directory which makes sense because i never made one, and also "Could not open dictionary database." I'm not sure what could be causing that.