The card creation process can be started at any time by pressing a, which
brings up a menu for selecting subtitles to include on the card. Once the sub
selection menu is active, the selected subtitles are shown at the bottom right
of the screen. Information on timing is at the bottom left, and a help menu
listing all available key bindings is at the top left.
Pressing a again adds the current subtitle to the selection. By default all
subtitles are added automatically. This behavior can be toggled with A, or
disabled in the script config.
q and e can be used to set the timing of the exported audio clip manually
to the start or end of the currently active subtitle. Q and E do the same
but using the current playback time instead. s sets the screenshot time to
the current frame. Pressing the timing keys again while a time is set resets
the corresponding time if the keypress would have set it to the same value.
For example, if the screenshot time is set to 15:00.000 and the current
timestamp in the video is also 15:00.000, pressing s will unset the
screenshot time.
If no times are set, the clip will begin at the start of the first selected
subtitle and end at the end of the last one. The screenshot will be the frame
that's visible when the export happens. Screenshots can be toggled with s
in the global menu.
y resets the selection and times. p plays the audio that would be exported
with the current selection/times. d ends the subtitle selection and begins
target word selection. f exports the selection immediately without any
target words. F does the same but opens the export menu
first.
By pressing K, line selection is skipped and the target word selection is
started immediately with the current subtitle. Ctrl+k skips the entirety of
the card creation process and instantly exports a card of the active subtitle
including an audio clip and screenshot. Ctrl+K opens the export menu first.
In this part of the export process you can select one or more target words. By default you are limited to one (in accordance with AJATT/MIA/Refold practices), but it's possible to increase this limit (or remove it entirely) in the script config.
Unneeded parts of the text are deletable using the bindings described in keys.md.
The word can be selected with the usual key combinations for selecting text:
(Ctrl+)(⇧+)←/→. ↑ and ↓ switch between subtitles and Alt+DEL
removes the current one from the selection.
Once you have chosen some text, you can look up any dictionary entries that
match it exactly using ⏎ and any entries starting with it using ⇧+⏎. If
you have lookup transformations configured you can use those with Ctrl+⏎.
Any whitespace characters (according to Unicode properties, including zero
width spaces (U+200B)) before or after the word are automatically removed
before searching.
If the search has any results, a new menu will open that allows you to choose
one of them with ↑/↓ and ⏎, or Ctrl+⏎ to add multiple definitions
without returning to the previous menu. ←/→ can be used to switch between
dictionaries (in the current group, if more than one is configured).
After selecting a definition, the word will be shown at the bottom right of
the screen. If you want, you can now add
pronunciation audio from Forvo using a (see below).
Like during subtitle selection, you can preview the selection audio with p.
Alt+⌫ removes the newest selected word.
When you are ready to export the card, you can do so using f, or F if you
want to open the export menu first.
If you hit a after selecting a target word, Immersive will load a list of
pronunciations of that word from Forvo. You will then be
able to choose one using ↑/↓. To play the audio of a pronunciation press
SPACE. If it isn't loaded yet it might take a moment before it starts
playing. If you want to load all audio files automatically set
forvo_preload_audio to yes in the script config.
Once you have selected a pronunciation that you are happy with, confirm it
using ⏎.
If your target language is not Japanese you have to change forvo_language
in the script config. The language code can be found on the Forvo website, for
example on this page at the right of:
日本 pronunciation in Japanese [ja]
The export menu allows exporting a card using Anki's 'Add' GUI (g), and
adding the export to an existing note.
The candidates for adding are found using the following query: "deck:<target deck>" "note:<target note type>" is:new, so all new cards of the target note
type from the target deck.
s exports by adding to the most recently added candidate. a allows
selecting a note from a list of candidates first. You can pick a note with
↑/↓ and ⏎. If you want to change how notes are displayed, you can use
the target config entry note_template:
| Variable | Type | Description |
|---|---|---|
note_template |
||
type |
single |
The note type |
id |
single |
The note ID |
tags |
list |
The tags of the note |
field_<field name> |
single |
Contents of the field <field name> |
Once you initiate the export, Immersive will first validate your Anki target config and only proceed if that is successful. After that, the audio clip and screenshot are encoded, as is the Forvo audio if there is any and it is configured to be.
Then the following template is applied to each field of the target note type,
as configured by the field:... entries of the target config, as well as to
all tags in the target's tags entry (any spaces in variable values are
replaced by underscores in tags):
| Variable | Type | Description |
|---|---|---|
field:<field name>/tags |
||
words |
list |
all target words |
selections |
list |
all target words, but exactly as selected from subtitles |
sentences |
list |
subtitles selected for export |
definitions |
list |
target word definitions as exported by the dictionary |
audio |
single |
audio clip as an Anki sound tag |
image |
single |
screenshot as an Anki/HTML image tag |
word_audio |
single |
pronunciation file as an Anki sound tag |
audio_file |
single |
filename of the audio clip |
image_file |
single |
filename of the screenshot |
word_audio_file |
single |
filename of the pronunciation audio |
path |
single |
path to the file mpv is currently playing (excluding the filename) the full URL for URLs |
filename |
single |
filename of the file mpv is currently playing identical to path for URLs
|
series_id |
single |
series ID as explained here |
series_title |
single |
series title as explained here |
prev_content |
single |
previous content of the field when adding to an existing note |
start |
single |
start time of the audio formatted as [HH:][MM:]SS |
end |
single |
end time of the audio formatted as [HH:][MM:]SS |
start_ms |
single |
start time of the audio formatted as [HH:][MM:]SS.mmm |
end_ms |
single |
end time of the audio formatted as [HH:][MM:]SS.mmm |
start_seconds |
single |
start time of the audio in seconds |
end_seconds |
single |
end time of the audio in seconds |
start_seconds_ms |
single |
start time of the audio in seconds and milliseconds |
end_seconds_ms |
single |
end time of the audio in seconds and milliseconds |
The card is exported with the resulting fields and tags.
The variables definitions and sentences can be filtered through
substitutions before being exported. These are defined in the target config
entries definition_substitutions and sentence_substitutions.
sentence_substitutions are applied when selecting a subtitle line,
definition_substitutions as the definition is added to the note
immediately before it is sent to Anki.
Both entries expect a multiline value, with each line representing a
substitution, e.g. this is what the default value of sentence_substitutions
would look like in a config file:
sentence_substitutions=[[
<（.-）
<%(.-%)
]]
Each line is split at the first <. Everything to the left of it becomes the
replacement and everything to the right becomes the pattern. If you want to
use a < in the replacement string, you can escape it as \<. Other common
escapes like \n or \t work as well, in both the replacement and pattern.
The pattern is a Lua pattern,
which means that the following characters need to be escaped by placing a %
before them if you want to use them literally: ^$()%.[]*+-?.
All occurrences of the pattern are replaced by the replacement string, or
deleted if the string is empty (i.e. < is the first character of the
substitution definition). The patterns are applied one after the other in the
order they are defined in.
The default value of sentence_substitutions contains the patterns （.-） and
%(.-%) (both with an empty replacement), which delete character names and
kanji readings from Japanese subtitles. If you want to disable them, simply
put this line in the global section of immersive-targets.conf:
sentence_substitutions=