Merge pull request #43 from ROBERT-MCDOWELL/main

pull attempt into v2.0

app.py

@@ -6,9 +6,10 @@ import subprocess
 import sys
 
 from lib.conf import *
-from lib.lang import language_options, default_language_code
+from lib.lang import language_mapping, default_language_code
 
 script_mode = NATIVE
+share = False
 
 def check_python_version():
     current_version = sys.version_info[:2]  # (major, minor)
@@ -99,10 +100,10 @@ def is_port_in_use(port):
         return s.connect_ex(('0.0.0.0', port)) == 0
 
 def main():
-    global script_mode, ebooks_dir
+    global script_mode, share, ebooks_dir
     
     # Convert the list of languages to a string to display in the help text
-    language_options_str = ", ".join(language_options)
+    lang_list_str = ", ".join(list(language_mapping.keys()))
 
     # Argument parser to handle optional parameters with descriptions
     parser = argparse.ArgumentParser(
@@ -111,12 +112,12 @@ def main():
 Example usage:    
 Windows:
     headless:
-    ebook2audiobook.cmd --headless --ebook 'path_to_ebook' --voice 'path_to_voice' --language en --use_custom_model --custom_model 'model.zip' --custom_config config.json --custom_vocab vocab.json
+    ebook2audiobook.cmd --headless --ebook 'path_to_ebook' --voice 'path_to_voice' --language en --custom_model 'model.zip'
     Graphic Interface:
     ebook2audiobook.cmd
 Linux/Mac:
     headless:
-    ./ebook2audiobook.sh --headless --ebook 'path_to_ebook' --voice 'path_to_voice' --language en --use_custom_model --custom_model 'model.zip' --custom_config config.json --custom_vocab vocab.json
+    ./ebook2audiobook.sh --headless --ebook 'path_to_ebook' --voice 'path_to_voice' --language en --custom_model 'model.zip'
     Graphic Interface:
     ./ebook2audiobook.sh
 """,
@@ -124,8 +125,8 @@ Linux/Mac:
     )
     options = [
         "--script_mode", "--share", "--headless", "--ebook", "--ebooks_dir",
-        "--voice", "--language", "--device", "--use_custom_model", "--custom_model", 
-        "--custom_config", "--custom_vocab", "--custom_model_url", "--temperature",
+        "--voice", "--language", "--device", "--custom_model", 
+        "--custom_model_url", "--temperature",
         "--length_penalty", "--repetition_penalty", "--top_k", "--top_p", "--speed",
         "--enable_text_splitting", "--version"
     ]
@@ -142,37 +143,31 @@ Linux/Mac:
     parser.add_argument(options[5], type=str,
                         help="Path to the target voice file for TTS. Optional, uses a default voice if not provided.")
     parser.add_argument(options[6], type=str, default="en",
-                        help=f"Language for the audiobook conversion. Options: {language_options_str}. Defaults to English (en).")
+                        help=f"Language for the audiobook conversion. Options: {lang_list_str}. Defaults to English (en).")
     parser.add_argument(options[7], type=str, default="cpu", choices=["cpu", "gpu"],
                         help=f"Type of processor unit for the audiobook conversion. If not specified: check first if gpu available, if not cpu is selected.")
-    parser.add_argument(options[8], action="store_true",
-                        help="Use a custom TTS model. Defaults to False. Must be True to use custom models.")
-    parser.add_argument(options[9], type=str,
+    parser.add_argument(options[8], type=str,
                         help="Path to the custom model file (.pth). Required if using a custom model.")
-    parser.add_argument(options[10], type=str,
-                        help="Path to the custom config file (config.json). Required if using a custom model.")
-    parser.add_argument(options[11], type=str,
-                        help="Path to the custom vocab file (vocab.json). Required if using a custom model.")
-    parser.add_argument(options[12], type=str,
+    parser.add_argument(options[9], type=str,
                         help=("URL to download the custom model as a zip file. Optional, but will be used if provided. "
                               "Examples include David Attenborough's model: "
                               "'https://huggingface.co/drewThomasson/xtts_David_Attenborough_fine_tune/resolve/main/Finished_model_files.zip?download=true'. "
                               "More XTTS fine-tunes can be found on my Hugging Face at 'https://huggingface.co/drewThomasson'."))
-    parser.add_argument(options[13], type=float, default=0.65,
+    parser.add_argument(options[10], type=float, default=0.65,
                         help="Temperature for the model. Defaults to 0.65. Higher temperatures lead to more creative outputs.")
-    parser.add_argument(options[14], type=float, default=1.0,
+    parser.add_argument(options[11], type=float, default=1.0,
                         help="A length penalty applied to the autoregressive decoder. Defaults to 1.0. Not applied to custom models.")
-    parser.add_argument(options[15], type=float, default=2.0,
+    parser.add_argument(options[12], type=float, default=2.0,
                         help="A penalty that prevents the autoregressive decoder from repeating itself. Defaults to 2.0.")
-    parser.add_argument(options[16], type=int, default=50,
+    parser.add_argument(options[13], type=int, default=50,
                         help="Top-k sampling. Lower values mean more likely outputs and increased audio generation speed. Defaults to 50.")
-    parser.add_argument(options[17], type=float, default=0.8,
+    parser.add_argument(options[14], type=float, default=0.8,
                         help="Top-p sampling. Lower values mean more likely outputs and increased audio generation speed. Defaults to 0.8.")
-    parser.add_argument(options[18], type=float, default=1.0,
+    parser.add_argument(options[15], type=float, default=1.0,
                         help="Speed factor for the speech generation. Defaults to 1.0.")
-    parser.add_argument(options[19], action="store_true",
+    parser.add_argument(options[16], action="store_true",
                         help="Enable splitting text into sentences. Defaults to False.")
-    parser.add_argument(options[20], action="version",version=f"ebook2audiobook version {version}",
+    parser.add_argument(options[17], action="version",version=f"ebook2audiobook version {version}",
                         help="Show the version of the script and exit")
 
     for arg in sys.argv:
@@ -188,6 +183,7 @@ Linux/Mac:
         sys.exit(1)
     
     script_mode = args.script_mode if args.script_mode else script_mode
+    share =  args.share if args.share else share
     
     if script_mode == NATIVE:
         check_pkg = check_and_install_requirements(requirements_file)
@@ -236,11 +232,13 @@ Linux/Mac:
             else:
                 print(f"Error: The directory {ebooks_dir} does not exist.")
                 sys.exit(1)
+
         elif args.ebook:
             progress_status, audiobook_file = convert_ebook(args)
             if audiobook_file is None:
                 print(f"Conversion failed: {progress_status}")
                 sys.exit(1)
+
         else:
             print("Error: In headless mode, you must specify either an ebook file using --ebook or an ebook directory using --ebooks_dir.")
             sys.exit(1)       
@@ -249,7 +247,7 @@ Linux/Mac:
         allowed_arguments = {'--share', '--script_mode'}
         passed_args_set = {arg for arg in passed_arguments if arg.startswith('--')}
         if passed_args_set.issubset(allowed_arguments):
-             web_interface(args.script_mode, args.share)
+             web_interface(script_mode, share)
         else:
             print("Error: In non-headless mode, no option or only '--share' can be passed")
             sys.exit(1)

ebook2audiobook.cmd

@@ -37,7 +37,7 @@ if not exist "%CALIBRE_TEMP_DIR%" (
 
 icacls "%CALIBRE_TEMP_DIR%" /grant Users:(OI)(CI)F /T
 
-for %%A in (%*) do (
+for %%A in (%ARGS%) do (
 	if "%%A"=="%DOCKER_UTILS%" (
 		set "SCRIPT_MODE=%DOCKER_UTILS%"
 		break
@@ -228,7 +228,7 @@ if not "%DOCKER_BUILD_STATUS%"=="0" (
 net session >nul 2>&1
 if %errorlevel% equ 0 (
     echo Restarting in user mode...
-    start "" /b cmd /c "%~f0" %*
+    start "" /b cmd /c "%~f0" %ARGS%
     exit /b
 )
 goto dispatch
@@ -269,7 +269,7 @@ if "%SCRIPT_MODE%"=="%FULL_DOCKER%" (
 		call conda create --prefix %SCRIPT_DIR%\%PYTHON_ENV% python=%PYTHON_VERSION% -y
 		call conda activate %SCRIPT_DIR%\%PYTHON_ENV%
 		call python -m pip install --upgrade pip
-		call python -m pip install beautifulsoup4 coqui-tts ebooklib docker "gradio>=4.44.0" mecab mecab-python3 "nltk>=3.8.2" pydub translate tqdm unidic
+		call python -m pip install --upgrade -r requirements.txt
 		call python -m unidic download
 		call python -m spacy download en_core_web_sm
 		call python -m nltk.downloader punkt_tab

ebook2audiobook.sh

@@ -212,7 +212,7 @@ function conda_check {
 		source $CONDA_ENV
 		conda activate $SCRIPT_DIR/$PYTHON_ENV
 		python -m pip install --upgrade pip
-		python -m pip install beautifulsoup4 coqui-tts ebooklib docker "gradio>=4.44.0" mecab mecab-python3 "nltk>=3.8.2" pydub translate tqdm unidic
+		python -m pip install --upgrade -r requirements.txt
 		python -m unidic download
 		python -m spacy download en_core_web_sm
 		python -m nltk.downloader punkt_tab

ebooks/test4.txt

@@ -1 +0,0 @@
-this is the test four from the result of text file to audiobook conversion.

ebooks/test_ar.txt

@@ -0,0 +1 @@
+هذا هو الاختبار من نتيجة تحويل ملف نصي إلى كتاب صوتي.

ebooks/test_cs.txt

@@ -0,0 +1 @@
+Toto je test od výsledku převodu textového souboru na audioknihu.

ebooks/test_da.txt

@@ -0,0 +1 @@
+Dette er testen fra resultatet af konvertering af tekstfil til lydbog.

ebooks/test_de.txt

@@ -0,0 +1 @@
+Dies ist der Test des Ergebnisses der Konvertierung einer Textdatei in ein Hörbuch.

ebooks/test_el.txt

@@ -0,0 +1 @@
+Αυτή είναι η δοκιμή από το αποτέλεσμα της μετατροπής αρχείου κειμένου σε ηχητικό βιβλίο.

ebooks/test_en.txt

@@ -0,0 +1 @@
+This is the test from the result of text file to audiobook conversion.

ebooks/test_es.txt

@@ -0,0 +1 @@
+Esta es la prueba del resultado de la conversión de archivo de texto a audiolibro.

ebooks/test_fi.txt

@@ -0,0 +1 @@
+Tämä on testi tekstitiedoston muuntamisen tuloksesta äänikirjaksi.

ebooks/test_fr.txt

@@ -0,0 +1 @@
+Ceci est le test provenant d'un fichier text en livre audio.

ebooks/test_hr.txt

@@ -0,0 +1 @@
+Ovo je test rezultata pretvorbe tekstualne datoteke u audioknjigu.

ebooks/test_it.txt

@@ -0,0 +1 @@
+Questo è il test del risultato della conversione del file di testo in audiolibro.

ebooks/test_ja.txt

@@ -0,0 +1 @@
+これは、テキスト ファイルからオーディオブックへの変換結果のテストです。

ebooks/test_ko.txt

@@ -0,0 +1 @@
+이는 텍스트 파일을 오디오북으로 변환한 결과에 대한 테스트입니다.

ebooks/test_nb.txt

@@ -0,0 +1 @@
+Dette er testen fra resultatet av konvertering av tekstfil til lydbok.

ebooks/test_nl.txt

@@ -0,0 +1 @@
+Dit is de test op basis van het resultaat van de conversie van een tekstbestand naar een audioboek.

ebooks/test_pl.txt

@@ -0,0 +1 @@
+To jest test wyniku konwersji pliku tekstowego na audiobook.

ebooks/test_pt.txt

@@ -0,0 +1 @@
+Este é o teste do resultado da conversão de ficheiro de texto em audiolivro.

ebooks/test_ro.txt

@@ -0,0 +1 @@
+Acesta este testul de la rezultatul conversiei fișierului text la cartea audio.

ebooks/test_ru.txt

@@ -0,0 +1 @@
+Это тест результата конвертации текстового файла в аудиокнигу.

ebooks/test_sl.txt

@@ -0,0 +1 @@
+To je preizkus rezultata pretvorbe besedilne datoteke v zvočno knjigo.

ebooks/test_sv.txt

@@ -0,0 +1 @@
+Detta är testet från resultatet av konvertering av textfil till ljudbok.

ebooks/test_zh.txt

@@ -0,0 +1 @@
+这是从文本文件到有声读物的转换结果进行的测试。

lib/functions.py

@@ -206,21 +206,18 @@ def download_and_extract(path_or_url, extract_to=models_dir):
     except Exception as e:
         raise DependencyError(e)
 
-def load_spacy_model(language):
-    model_name = f"{language}_core_web_sm" 
-    try:
-        nltk.data.find('tokenizers/punkt_tab')
-    except LookupError:
-        print("Downloading NLTK punkt tokenizer...")
-        nltk.download('punkt_tab')
-    if not is_package(model_name):
-        try:
-            print(f"Downloading model: {model_name}")
-            download_package(model_name)  # Download the model if not installed
-        except Exception as e:
-            print(f"Error downloading model {model_name}: {e}")
-            return None
-    return spacy.load(model_name)
+#def load_spacy_model(language):
+#    lang_pack = language_mapping[language]["model"]
+#    try:
+#        nlp = spacy.load(lang_pack)
+#    except OSError:
+#        print("Spacy model not found. Tyring to download it...")
+#        try:
+#            subprocess.run(["python", "-m", "spacy", "download", lang_pack])
+#            nlp = spacy.load(lang_pack)
+#        except OSError:
+#             raise ValueError(f"Spacy model does not exist for {language_mapping[language]['name']}...")
+#    return nlp
 
 def translate_pronouns(language):
     global ebook_pronouns  
@@ -385,7 +382,7 @@ def concat_audio_chapters(metadatas, cover_file):
                         
                 languages = metadatas.get('Languages', None)
                 if languages:
-                    ffmpeg_metadata += f"language={languages}\n\n"  # Language
+                    ffmpeg_metadata += f"language={languages}\n\n"
                 
             else:
                 print("Warning: metadatas is None. Skipping metadata generation.")
@@ -676,10 +673,11 @@ def split_long_sentence(sentence, language='en', max_pauses=10):
     :return: A list of sentence parts that meet the criteria.
     """
     # Get the Max character length for the selected language -2 : with a default of 248 if no language is found
-    max_length = (char_limits.get(language, 250)-2)
+    char_limits = language_mapping[language]["char_limit"]
+    max_length = (char_limits - 2)
 
     # Adjust the pause punctuation symbols based on language
-    if language == 'zh-cn':
+    if language == 'zh':
         punctuation = ['，', '。', '；', '？', '！']  # Chinese-specific pause punctuation including sentence-ending marks
     elif language == 'ja':
         punctuation = ['、', '。', '；', '？', '！']  # Japanese-specific pause punctuation
@@ -724,7 +722,7 @@ def convert_chapters_to_audio(device, temperature, length_penalty, repetition_pe
             target_voice_file = default_target_voice_file
         
         # Handle custom model or use standard TTS model
-        print("Loading model...")
+        print("Loading TTS ...")
         if custom_model:
             config_path = custom_model['config']
             model_path = custom_model['model']
@@ -732,22 +730,22 @@ def convert_chapters_to_audio(device, temperature, length_penalty, repetition_pe
             config = XttsConfig()
             config.models_dir = models_dir
             config.load_json(config_path)
-            model = Xtts.init_from_config(config)
-            model.load_checkpoint(config, checkpoint_dir=model_path, vocab_path=vocab_path)
+            tts = Xtts.init_from_config(config)
+            tts.load_checkpoint(config, checkpoint_dir=model_path, vocab_path=vocab_path)
         else:
             #selected_tts_model = "tts_models/multilingual/multi-dataset/xtts_v2"
-            #model = TTS(selected_tts_model, progress_bar=False).to(device)
-            base_dir = os.path.join(models_dir,"tts_models--multilingual--multi-dataset--xtts_v2")
+            #tts = TTS(selected_tts_model, progress_bar=False).to(device)
+            base_dir = os.path.join(models_dir,"XTTS-v2")
             config_path = os.path.join(base_dir,"config.json")
             config = XttsConfig()
             config.models_dir = models_dir
             config.load_json(config_path)
-            model = Xtts.init_from_config(config)
-            model.load_checkpoint(config, checkpoint_dir=base_dir)
+            tts = Xtts.init_from_config(config)
+            tts.load_checkpoint(config, checkpoint_dir=base_dir)
           
-        model.to(device)
+        tts.to(device)
         print("Computing speaker latents...")
-        gpt_cond_latent, speaker_embedding = model.get_conditioning_latents(audio_path=[target_voice_file])
+        gpt_cond_latent, speaker_embedding = tts.get_conditioning_latents(audio_path=[target_voice_file])
         
         chapters_dir_audio_fragments = os.path.join(ebook_chapters_audio_dir, "fragments")
         os.makedirs(chapters_dir_audio_fragments, exist_ok=True)
@@ -755,35 +753,37 @@ def convert_chapters_to_audio(device, temperature, length_penalty, repetition_pe
         # Calculate the total number of chapters and segments (fragments) to set progress bar correctly
         total_segments = 0
         total_chapters = len([f for f in os.listdir(ebook_chapters_dir) if f.endswith('.txt')])
+        
+        # Check if the language is nltk compatible
+        nltk_language = language_mapping[language]["name"].lower() if language_mapping[language] else None
 
         # Pre-calculate total segments (sentences + fragments per chapter)
         for chapter_file in sorted(os.listdir(ebook_chapters_dir)):
             if cancellation_requested.is_set():
-                stop_and_detach_tts(model)
+                stop_and_detach_tts(tts)
                 msg = "Cancel requested"
                 raise ValueError(msg)
             if chapter_file.endswith('.txt'):
                 with open(os.path.join(ebook_chapters_dir, chapter_file), 'r', encoding='utf-8') as file:
                     chapter_text = file.read()
-                    nltk_language = language_mapping.get(language)
-                    if nltk_language:
-                        sentences = nltk.tokenize.sent_tokenize(chapter_text, language=nltk_language)
-                    else:
-                        sentences = [chapter_text]
-                    
+                    #if nltk_language is not None:
+                        #sentences = nltk.tokenize.sent_tokenize(chapter_text, language=nltk_language)
+                        #sentences = nltk.tokenize.word_tokenize(chapter_text, language=nltk_language, preserve_line=False)
+                    #else:
+                    sentences = [chapter_text]
+
                     # Calculate total fragments for this chapter
                     for sentence in sentences:
                         fragments = split_long_sentence(sentence, language=language)
                         total_segments += len(fragments)
 
-        # Initialize progress tracking
         current_progress = 0
         total_progress = total_segments + total_chapters  # Total is chapters + segments/fragments
 
         with tqdm(total=total_progress, desc="Processing 0.00%", bar_format='{desc}: {n_fmt}/{total_fmt} ', unit="step") as t:
             for chapter_file in sorted(os.listdir(ebook_chapters_dir)):
                 if cancellation_requested.is_set():
-                    stop_and_detach_tts(model)
+                    stop_and_detach_tts(tts)
                     msg = "Cancel requested"
                     raise ValueError(msg)
                 if chapter_file.endswith('.txt'):
@@ -800,30 +800,29 @@ def convert_chapters_to_audio(device, temperature, length_penalty, repetition_pe
 
                     with open(chapter_file_path, 'r', encoding='utf-8') as file:
                         chapter_text = file.read()
-                        nltk_language = language_mapping.get(language)
-                        
-                        if nltk_language:
-                            sentences = nltk.tokenize.sent_tokenize(chapter_text, language=nltk_language)
-                        else:
-                            sentences = [chapter_text]
+                        #if nltk_language is not None:
+                        #    sentences = nltk.tokenize.sent_tokenize(chapter_text, language=nltk_language)
+                            #sentences = nltk.tokenize.word_tokenize(chapter_text, language=nltk_language, preserve_line=False)
+                        #else:
+                        sentences = [chapter_text]
                         
                         for sentence in sentences:
                             if cancellation_requested.is_set():
-                                stop_and_detach_tts(model)
+                                stop_and_detach_tts(tts)
                                 msg = "Cancel requested"
                                 raise ValueError(msg)
                             fragments = split_long_sentence(sentence, language=language)
                             for fragment in fragments:
                                 if cancellation_requested.is_set():
-                                    stop_and_detach_tts(model)
+                                    stop_and_detach_tts(tts)
                                     msg = "Cancel requested"
                                     raise ValueError(msg)
                                 if fragment != "":
                                     print(f"Generating fragment: {fragment}...")
                                     fragment_file_path = os.path.join(chapters_dir_audio_fragments, f"{count_fragments}.wav")
                                     
-                                    #if custom_model:
-                                    out = model.inference(
+                                    #if custom_tts:
+                                    out = tts.inference(
                                         fragment, language, gpt_cond_latent, speaker_embedding, 
                                         temperature=temperature, repetition_penalty=repetition_penalty, 
                                         top_k=top_k, top_p=top_p, speed=speed, enable_text_splitting=enable_text_splitting
@@ -876,11 +875,11 @@ def convert_chapters_to_audio(device, temperature, length_penalty, repetition_pe
     except Exception as e:
         raise DependencyError(e)
     
-def stop_and_detach_tts(model):
-    # Move the model to CPU if on GPU
-    if next(model.parameters()).is_cuda:
-        model.to('cpu')
-    del model
+def stop_and_detach_tts(tts):
+    # Move the tts to CPU if on GPU
+    if next(tts.parameters()).is_cuda:
+        tts.to('cpu')
+    del tts
     if torch.cuda.is_available():
         torch.cuda.empty_cache()
 
@@ -913,10 +912,6 @@ def convert_ebook(args):
             device = args.device.lower()
             target_voice_file = args.voice
             language = args.language
-            use_custom_model = args.use_custom_model
-            custom_model_file = args.custom_model
-            custom_config_file = args.custom_config
-            custom_vocab_file = args.custom_vocab
             temperature = args.temperature
             length_penalty = args.length_penalty
             repetition_penalty = args.repetition_penalty
@@ -924,6 +919,7 @@ def convert_ebook(args):
             top_p = args.top_p
             speed = args.speed
             enable_text_splitting = args.enable_text_splitting
+            custom_model_file = args.custom_model
             custom_model_url = args.custom_model_url
 
             if not os.path.splitext(ebook_src)[1]:
@@ -951,8 +947,7 @@ def convert_ebook(args):
                 ebook_pronouns = translate_pronouns(language)
                 
             # Load spaCy model for language analysis (you can switch models based on language)
-            nlp = load_spacy_model(language)
-
+            #if load_spacy_model(language):
             # Prepare tmp dir and properties
             if prepare_dirs(args.ebook) : 
                 
@@ -961,7 +956,7 @@ def convert_ebook(args):
 
                 # Handle custom model if the user chose to use one
                 custom_model = None
-                if use_custom_model and custom_model_file and custom_config_file and custom_vocab_file:
+                if custom_model_file and custom_config_file and custom_vocab_file:
                     custom_model = {
                         'model': custom_model_file,
                         'config': custom_config_file,
@@ -969,7 +964,7 @@ def convert_ebook(args):
                     }
 
                 # If a custom model URL is provided, download and use it
-                if use_custom_model and custom_model_url:
+                if custom_model_url:
                     print(f"Received custom model URL: {custom_model_url}")
                     model_dir = get_model_dir_from_url(custom_model_url)
                     if download_and_extract(custom_model_url, model_dir):
@@ -1007,6 +1002,8 @@ def convert_ebook(args):
                         raise DependencyError("convert_chapters_to_audio() failed!")
                 else:
                     return None, None
+                #else:
+                #    return None, None
             else:
                 print(f"Temporary directory {tmp_dir} not removed due to failure.")
                 return None, None
@@ -1017,12 +1014,14 @@ def convert_ebook(args):
 
 def web_interface(mode, share):
     global ebook_src, is_converting, interface, cancellation_requested, is_gui_process, script_mode, is_gui_shared, audiobooks_ddn
-    
+
     script_mode = mode
     is_gui_process = True
     is_gui_shared = share
     audiobook_file = None
-    
+    language_options = [details["native_name"] for details in language_mapping.values()]
+    default_language_native = language_mapping[default_language_code]["native_name"]
+
     theme = gr.themes.Origin(
         primary_hue="amber",
         secondary_hue="green",
@@ -1042,8 +1041,11 @@ def web_interface(mode, share):
         gr.HTML(
             """
             <style>
+                input[type="checkbox"] {
+                    border-color: #fafafa !Important;
+                }
                 .svelte-1xyfx7i.center.boundedheight.flex{
-                    height: 110px !important;
+                    height: 120px !important;
                 }
                 .block.svelte-5y6bt2 {
                     padding: 10px !important;
@@ -1056,6 +1058,10 @@ def web_interface(mode, share):
                     margin: 0 !important;
                     font-size: 12px !important;
                 }
+                .block.svelte-5y6bt2.padded {
+                    height: auto !important;
+                    padding: 10px !important;
+                }
                 .block.svelte-5y6bt2.padded.hide-container {
                     height: auto !important;
                     padding: 0 !important;
@@ -1074,25 +1080,25 @@ def web_interface(mode, share):
                     padding: 0;
                     margin: 0;
                 }
+                #component-8, #component-9, #component-34 {
+                    height: 119px !important;
+                }
             </style>
             """
         )
-        with gr.Tabs():  # Create tabs for better UI organization
+        with gr.Tabs():
             with gr.TabItem("Input Options"):
                 with gr.Row():
                     with gr.Column(scale=3):
                         ebook_file = gr.File(label="eBook File")
-                        target_voice_file = gr.File(label="Target Voice File (Optional)")
-                        language = gr.Dropdown(label="Language", choices=language_options, value="en")
                         device = gr.Radio(label="Processor Unit", choices=["CPU", "GPU"], value="CPU")
-
+                        language = gr.Dropdown(label="Language", choices=language_options, value=default_language_native)  
                     with gr.Column(scale=3):
-                        use_custom_model = gr.Checkbox(label="Use Custom Model")
-                        custom_model_file = gr.File(label="Custom Model File (Optional)", visible=False)
-                        custom_config_file = gr.File(label="Custom Config File (Optional)", visible=False)
-                        custom_vocab_file = gr.File(label="Custom Vocab File (Optional)", visible=False)
-                        custom_model_url = gr.Textbox(label="Custom Model Zip URL (Optional)", visible=False)
-
+                        with gr.Group():
+                            target_voice_file = gr.File(label="Cloning Voice* (a .wav or .mp3 no more than 12sec)")
+                            custom_model_file = gr.File(label="Model* (a .zip containing config.json, vocab.json, model.pth)")
+                            custom_model_url = gr.Textbox(placeholder="https://www.example.com/model.zip", label="Model from URL*")
+                            gr.Markdown('<p>* Optional</p>')
             with gr.TabItem("Audio Generation Preferences"):
                 gr.Markdown(
                     """
@@ -1154,9 +1160,8 @@ def web_interface(mode, share):
                     info="Splits long texts into sentences to generate audio in chunks. Useful for very long inputs."
                 )
                 
-            session_status = gr.Textbox(label="Session")
-            session = gr.Textbox(label="Session", visible=False)
-
+        session_status = gr.Textbox(label="Session")
+        session = gr.Textbox(label="Session", visible=False)
         conversion_progress = gr.Textbox(label="Progress")
         convert_btn = gr.Button("Convert", variant="primary", interactive=False)
         audio_player = gr.Audio(label="Listen", type="filepath", visible=False)
@@ -1231,11 +1236,20 @@ def web_interface(mode, share):
             return gr.Button("Convert", variant="primary", interactive=False), None, audiobook_file, update_audiobooks_ddn()
 
         def refresh_audiobook_list():
-            if not os.path.isdir(audiobooks_dir):
-                os.makedirs(audiobooks_dir, exist_ok=True)
-            files = [f for f in os.listdir(audiobooks_dir)]
-            files.sort(key=lambda x: os.path.getmtime(os.path.join(audiobooks_dir, x)), reverse=True)
+            files = []
+            if audiobooks_dir is not None:
+                if not os.path.isdir(audiobooks_dir):
+                    os.makedirs(audiobooks_dir, exist_ok=True)
+                files = [f for f in os.listdir(audiobooks_dir)]
+                files.sort(key=lambda x: os.path.getmtime(os.path.join(audiobooks_dir, x)), reverse=True)
             return files
+
+        def update_audiobook_link(audiobook):
+            if audiobooks_dir is not None:
+                if audiobook:
+                    link = os.path.join(audiobooks_dir, audiobook)
+                    return link, link, gr.update(visible=True)
+            return None, None, gr.update(visible=False)
             
         def disable_convert_btn():
             return gr.Button("Convert", variant="primary", interactive=False)
@@ -1244,13 +1258,6 @@ def web_interface(mode, share):
             files = refresh_audiobook_list()
             return gr.Dropdown(choices=files, label="Audiobooks", value=files[0] if files else None)
 
-        def update_audiobook_link(audiobook):
-            if audiobook:
-                link = os.path.join(audiobooks_dir, audiobook)
-                return link, link, gr.update(visible=True)
-            else:
-                return None, None, gr.update(visible=False)
-
         def change_ebook_file(btn, f):
             global ebook_src, is_converting, cancellation_requested
             if f is None:
@@ -1269,13 +1276,13 @@ def web_interface(mode, share):
             data["event"] = 'change_data'
             return data
 
-        def process_conversion(session, device, ebook_file, target_voice_file, language, use_custom_model, custom_model_file, custom_config_file, custom_vocab_file, custom_model_url, temperature, length_penalty, repetition_penalty, top_k, top_p, speed, enable_text_splitting):                             
+        def process_conversion(session, device, ebook_file, target_voice_file, language, custom_model_file, custom_model_url, temperature, length_penalty, repetition_penalty, top_k, top_p, speed, enable_text_splitting):                             
             global ebook_src, is_converting, audiobook_file
+
             ebook_src = ebook_file.name if ebook_file else None
             target_voice_file = target_voice_file.name if target_voice_file else None
             custom_model_file = custom_model_file.name if custom_model_file else None
-            custom_config_file = custom_config_file.name if custom_config_file else None
-            custom_vocab_file = custom_vocab_file.name if custom_vocab_file else None
+            language = next((code for code, details in language_mapping.items() if details["native_name"] == language), None)
 
             if not ebook_src:
                 return "Error: eBook file is required."
@@ -1288,10 +1295,7 @@ def web_interface(mode, share):
                 ebook=ebook_src,
                 voice=target_voice_file,
                 language=language,
-                use_custom_model=use_custom_model,
                 custom_model=custom_model_file,
-                custom_config=custom_config_file,
-                custom_vocab=custom_vocab_file,
                 custom_model_url=custom_model_url,
                 temperature=float(temperature),
                 length_penalty=float(length_penalty),
@@ -1339,11 +1343,6 @@ def web_interface(mode, share):
                 audiobooks_dir = os.path.join(audiobooks_host_dir, f"web-{data['session_id']}")
             return [data, f"{warning_text}{warning_text_extra}", data["session_id"], update_audiobooks_ddn()]
 
-        use_custom_model.change(
-            lambda x: [gr.update(visible=x)] * 4,
-            inputs=[use_custom_model],
-            outputs=[custom_model_file, custom_config_file, custom_vocab_file, custom_model_url]
-        )
         ebook_file.change(
             fn=change_ebook_file,
             inputs=[convert_btn, ebook_file],
@@ -1383,9 +1382,8 @@ def web_interface(mode, share):
             fn=process_conversion,
             inputs=[
                 session, device, ebook_file, target_voice_file, language, 
-                use_custom_model, custom_model_file, custom_config_file, 
-                custom_vocab_file, custom_model_url, temperature, length_penalty, repetition_penalty, 
-                top_k, top_p, speed, enable_text_splitting
+                custom_model_file, custom_model_url, temperature, length_penalty,
+                repetition_penalty, top_k, top_p, speed, enable_text_splitting
             ],
             outputs=[conversion_progress, modal_html]           
         ).then(

lib/lang.py

@@ -1,48 +1,33 @@
 import os
 
-language_options = [
-    "en", "es", "fr", "de", "it", "pt", "pl", "tr", "ru", "nl", "cs", "ar", "zh-cn", "ja", "hu", "ko"
-]
-char_limits = {
-    "en": 250,      # English
-    "es": 239,      # Spanish
-    "fr": 273,      # French
-    "de": 253,      # German
-    "it": 213,      # Italian
-    "pt": 203,      # Portuguese
-    "pl": 224,      # Polish
-    "tr": 226,      # Turkish
-    "ru": 182,      # Russian
-    "nl": 251,      # Dutch
-    "cs": 186,      # Czech
-    "ar": 166,      # Arabic
-    "zh-cn": 82,    # Chinese (Simplified)
-    "ja": 71,       # Japanese
-    "hu": 224,      # Hungarian
-    "ko": 95,       # Korean
-}
-
-# Mapping of language codes to NLTK's supported language names
 language_mapping = {
-    "en": "english",
-    "de": "german",
-    "fr": "french",
-    "es": "spanish",
-    "it": "italian",
-    "pt": "portuguese",
-    "nl": "dutch",
-    "pl": "polish",  
-    "cs": "czech",   
-    "ru": "russian",
-    "tr": "turkish",
-    "el": "greek",
-    "et": "estonian",
-    "no": "norwegian",
-    "ml": "malayalam",
-    "sl": "slovene",
-    "da": "danish",
-    "fi": "finnish",
-    "sv": "swedish"
+    "ar": {"name": "Arabic", "native_name": "العربية", "char_limit": 166, "model": "ar_core_news_sm"},
+    "cs": {"name": "Czech", "native_name": "Čeština", "char_limit": 186, "model": "cs_core_news_sm"},
+    "da": {"name": "Danish", "native_name": "Dansk", "char_limit": 220, "model": "da_core_news_sm"},
+    "de": {"name": "German", "native_name": "Deutsch", "char_limit": 253, "model": "de_core_news_sm"},
+    "el": {"name": "Greek", "native_name": "Ελληνικά", "char_limit": 200, "model": "el_core_news_sm"},
+    "en": {"name": "English", "native_name": "English", "char_limit": 250, "model": "en_core_web_sm"},
+    "es": {"name": "Spanish", "native_name": "Español", "char_limit": 239, "model": "es_core_news_md"},
+    "fa": {"name": "Persian", "native_name": "فارسی", "char_limit": 150, "model": "???"},
+    "fi": {"name": "Finnish", "native_name": "Suomi", "char_limit": 230, "model": "fi_core_news_sm"},
+    "fr": {"name": "French", "native_name": "Français", "char_limit": 273, "model": "fr_core_news_sm"},
+    "hi": {"name": "Hindi", "native_name": "हिंदी", "char_limit": 220, "model": "???"},
+    "hr": {"name": "Croatian", "native_name": "Hrvatski", "char_limit": 210, "model": "hr_core_news_sm"},
+    "it": {"name": "Italian", "native_name": "Italiano", "char_limit": 213, "model": "it_core_news_sm"},
+    "ja": {"name": "Japanese", "native_name": "日本語", "char_limit": 71, "model": "ja_core_news_sm"},
+    "ko": {"name": "Korean", "native_name": "한국어", "char_limit": 95, "model": "ko_core_news_sm"},
+    "nb": {"name": "Norwegian", "native_name": "Norsk Bokmål", "char_limit": 225, "model": "nb_core_news_sm"},
+    "nl": {"name": "Dutch", "native_name": "Nederlands", "char_limit": 251, "model": "nl_core_news_sm"},
+    "pl": {"name": "Polish", "native_name": "Polski", "char_limit": 224, "model": "pl_core_news_sm"},
+    "pt": {"name": "Portuguese", "native_name": "Português", "char_limit": 203, "model": "pt_core_news_sm"},
+    "ro": {"name": "Romanian", "native_name": "Română", "char_limit": 190, "model": "ro_core_news_sm"},
+    "ru": {"name": "Russian", "native_name": "Русский", "char_limit": 182, "model": "ru_core_news_sm"},
+    "sl": {"name": "Slovenian", "native_name": "Slovenščina", "char_limit": 210, "model": "sl_core_news_sm"},
+    "sv": {"name": "Swedish", "native_name": "Svenska", "char_limit": 215, "model": "sv_core_news_sm"},
+    "tr": {"name": "Turkish", "native_name": "Türkçe", "char_limit": 200, "model": "???"},
+    "vi": {"name": "Vietnamese", "native_name": "Tiếng Việt", "char_limit": 180, "model": "???"},
+    "yo": {"name": "Yoruba", "native_name": "Yorùbá", "char_limit": 180, "model": "???"},
+    "zh": {"name": "Chinese", "native_name": "中文", "char_limit": 82, "model": "zh_core_web_sm"}
 }
 
 default_language_code = "en"

pyproject.toml

@@ -13,14 +13,20 @@ authors = [
 ]
 dependencies = [
     "beautifulsoup4",
+	"camel-tools",
     "coqui-tts",
-    "ebooklib",
+	"cutlet",
     "docker",
+    "ebooklib",
+	"gensim",
     "gradio>=4.44.0",
+	"jieba",
     "mecab",
     "mecab-python3",
     "nltk>=3.8.2",
     "pydub",
+	"pypinyin",
+	"sentencepiece",
     "translate",
     "tqdm",
     "unidic",

requirements.txt

@@ -1,12 +1,18 @@
 beautifulsoup4
+camel-tools
 coqui-tts
-ebooklib
+cutlet
 docker
+ebooklib
+gensim
 gradio>=4.44.0
+jieba
 mecab
 mecab-python3
 nltk>=3.8.2
 pydub
+pypinyin
+sentencepiece
 translate
 tqdm
 unidic