Commit: 7b7412c4f75944b1b95e621241f474eeafd3ba07
Parent: b24689a62a180fc1274460c7fda60452b3cdd4fb
Author: Ren Tatsumoto
Date: Sun, 13 Dec 2020 22:44:35 +0300
rewrite forvo support
Diffstat:
| M | subs2srs.lua | | | 139 | ++++++++++++++++++++++++++++++++++++++++++++++++++++++++----------------------- |
1 file changed, 99 insertions(+), 40 deletions(-)
diff --git a/subs2srs.lua b/subs2srs.lua
@@ -412,46 +412,6 @@ local function join_media_fields(note1, note2)
return note1
end
-local function get_forvo_pronunciation(word)
- word = platform.windows and url_encode(word) or word
- local forvo_format = config.audio_extension:sub(2)
- local forvo_page = subprocess { 'curl', '-s', string.format('https://forvo.com/search/%s/ja', word) }.stdout
- local play_params = string.match(forvo_page, "Play%((.-)%);")
- if not play_params then
- return nil
- end
- local iter = string.gmatch(play_params, "'(.-)'")
- local formats = { mp3 = iter(), ogg = iter() }
- local audio_url = string.format('https://audio00.forvo.com/%s/%s', forvo_format, base64d(formats[forvo_format]))
- local pronunciation_filename = string.format('forvo_%s.%s', platform.windows and os.time() or word, forvo_format)
- local pronunciation_path = utils.join_path(config.collection_path, pronunciation_filename)
- mp.commandv('run', 'curl', audio_url, '-s', '-L', '-o', pronunciation_path)
- return string.format('[sound:%s]', pronunciation_filename)
-end
-
-local function append_forvo_pronunciation(note1, note2)
- if config.use_forvo == 'no' then
- return note1
- end
- if type(note2[config.vocab_audio_field]) ~= 'string' then
- return note1
- end
- if is_empty(note2[config.vocab_field]) then
- return note1
- end
- if config.use_forvo == 'always' or is_empty(note2[config.vocab_audio_field]) then
- local forvo_pronunciation = get_forvo_pronunciation(note2[config.vocab_field])
- if not is_empty(forvo_pronunciation) then
- if config.vocab_audio_field == config.audio_field then
- note1[config.vocab_audio_field] = forvo_pronunciation .. note1[config.vocab_audio_field]
- else
- note1[config.vocab_audio_field] = forvo_pronunciation
- end
- end
- end
- return note1
-end
-
local validate_config
do
local function is_webp_supported()
@@ -598,6 +558,103 @@ end
platform = is_running_windows() and init_platform_windows() or init_platform_nix()
------------------------------------------------------------
+-- utils for downloading pronunciations from Forvo
+
+local append_forvo_pronunciation
+do
+ local function audio_reencode(source_path, dest_path)
+ local args = {
+ 'mpv',
+ source_path,
+ '--loop-file=no',
+ '--video=no',
+ '--no-ocopy-metadata',
+ '--no-sub',
+ '--audio-channels=mono',
+ '--oacopts-add=vbr=on',
+ '--oacopts-add=application=voip',
+ '--oacopts-add=compression_level=10',
+ '--af-append=silenceremove=1:0:-50dB',
+ table.concat { '--oac=', config.audio_codec },
+ table.concat { '--oacopts-add=b=', config.audio_bitrate },
+ table.concat { '-o=', dest_path }
+ }
+ return subprocess(args)
+ end
+
+ local function reencode_and_store(source_path, filename)
+ local reencoded_path = utils.join_path(platform.tmp_dir(), 'reencoded_' .. filename)
+ audio_reencode(source_path, reencoded_path)
+ local result = ankiconnect.store_file(filename, reencoded_path)
+ os.remove(reencoded_path)
+ return result
+ end
+
+ local function curl_save(source_url, save_location)
+ local curl_args = { 'curl', source_url, '-s', '-L', '-o', save_location }
+ return subprocess(curl_args).status == 0
+ end
+
+ local function get_pronunciation_url(word)
+ local file_format = config.audio_extension:sub(2)
+ local forvo_page = subprocess { 'curl', '-s', string.format('https://forvo.com/search/%s/ja', url_encode(word)) }.stdout
+ local play_params = string.match(forvo_page, "Play%((.-)%);")
+
+ if play_params then
+ local iter = string.gmatch(play_params, "'(.-)'")
+ local formats = { mp3 = iter(), ogg = iter() }
+ return string.format('https://audio00.forvo.com/%s/%s', file_format, base64d(formats[file_format]))
+ end
+ end
+
+ local function get_forvo_pronunciation(word)
+ local audio_url = get_pronunciation_url(word)
+
+ if is_empty(audio_url) then
+ msg.warn(string.format("Seems like Forvo doesn't have audio for word %s.", word))
+ return
+ end
+
+ local filename = string.format('forvo_%s%s', platform.windows and os.time() or word, config.audio_extension)
+ local tmp_filepath = utils.join_path(platform.tmp_dir(), filename)
+
+ local result
+ if curl_save(audio_url, tmp_filepath) and reencode_and_store(tmp_filepath, filename) then
+ result = string.format('[sound:%s]', filename)
+ else
+ msg.warn(string.format("Couldn't download audio for word %s from Forvo.", word))
+ end
+
+ os.remove(tmp_filepath)
+ return result
+ end
+
+ append_forvo_pronunciation = function(appended_data, stored_data)
+ if config.use_forvo == 'no' then
+ return appended_data
+ end
+ if type(stored_data[config.vocab_audio_field]) ~= 'string' then
+ return appended_data
+ end
+ if is_empty(stored_data[config.vocab_field]) then
+ return appended_data
+ end
+ if config.use_forvo == 'always' or is_empty(stored_data[config.vocab_audio_field]) then
+ local forvo_pronunciation = get_forvo_pronunciation(stored_data[config.vocab_field])
+ if not is_empty(forvo_pronunciation) then
+ if config.vocab_audio_field == config.audio_field then
+ -- improperly configured fields. don't lose sentence audio
+ appended_data[config.vocab_audio_field] = forvo_pronunciation .. appended_data[config.vocab_audio_field]
+ else
+ appended_data[config.vocab_audio_field] = forvo_pronunciation
+ end
+ end
+ end
+ return appended_data
+ end
+end
+
+------------------------------------------------------------
-- provides interface for creating audio clips and snapshots
encoder = {}
@@ -749,8 +806,10 @@ ankiconnect.store_file = function(filename, file_path)
local _, error = ankiconnect.parse_result(ret)
if not error then
msg.info(string.format("File stored: '%s'.", filename))
+ return true
else
msg.error(string.format("Couldn't store file '%s': %s", filename, error))
+ return false
end
end