webscout 8.2.7__py3-none-any.whl → 8.2.9__py3-none-any.whl

webscout/Provider/TTS/streamElements.py

@@ -0,0 +1,333 @@
+import time
+import requests
+import pathlib
+import urllib.parse
+import tempfile
+from typing import Union
+from io import BytesIO
+from webscout import exceptions
+from webscout.litagent import LitAgent
+from concurrent.futures import ThreadPoolExecutor, as_completed
+from . import utils
+from .base import BaseTTSProvider
+
+class StreamElements(BaseTTSProvider):
+    """
+    Text-to-speech provider using the StreamElements API.
+    """
+
+    # Request headers
+    headers: dict[str, str] = {
+        "User-Agent": LitAgent().random()
+    }
+    all_voices: list[str] = [
+        "Filiz",
+        "Astrid",
+        "Tatyana",
+        "Maxim",
+        "Carmen",
+        "Ines",
+        "Cristiano",
+        "Vitoria",
+        "Ricardo",
+        "Maja",
+        "Jan",
+        "Jacek",
+        "Ewa",
+        "Ruben",
+        "Lotte",
+        "Liv",
+        "Seoyeon",
+        "Takumi",
+        "Mizuki",
+        "Giorgio",
+        "Carla",
+        "Bianca",
+        "Karl",
+        "Dora",
+        "Mathieu",
+        "Celine",
+        "Chantal",
+        "Penelope",
+        "Miguel",
+        "Mia",
+        "Enrique",
+        "Conchita",
+        "Geraint",
+        "Salli",
+        "Matthew",
+        "Kimberly",
+        "Kendra",
+        "Justin",
+        "Joey",
+        "Joanna",
+        "Ivy",
+        "Raveena",
+        "Aditi",
+        "Emma",
+        "Brian",
+        "Amy",
+        "Russell",
+        "Nicole",
+        "Vicki",
+        "Marlene",
+        "Hans",
+        "Naja",
+        "Mads",
+        "Gwyneth",
+        "Zhiyu",
+        "es-ES-Standard-A",
+        "it-IT-Standard-A",
+        "it-IT-Wavenet-A",
+        "ja-JP-Standard-A",
+        "ja-JP-Wavenet-A",
+        "ko-KR-Standard-A",
+        "ko-KR-Wavenet-A",
+        "pt-BR-Standard-A",
+        "tr-TR-Standard-A",
+        "sv-SE-Standard-A",
+        "nl-NL-Standard-A",
+        "nl-NL-Wavenet-A",
+        "en-US-Wavenet-A",
+        "en-US-Wavenet-B",
+        "en-US-Wavenet-C",
+        "en-US-Wavenet-D",
+        "en-US-Wavenet-E",
+        "en-US-Wavenet-F",
+        "en-GB-Standard-A",
+        "en-GB-Standard-B",
+        "en-GB-Standard-C",
+        "en-GB-Standard-D",
+        "en-GB-Wavenet-A",
+        "en-GB-Wavenet-B",
+        "en-GB-Wavenet-C",
+        "en-GB-Wavenet-D",
+        "en-US-Standard-B",
+        "en-US-Standard-C",
+        "en-US-Standard-D",
+        "en-US-Standard-E",
+        "de-DE-Standard-A",
+        "de-DE-Standard-B",
+        "de-DE-Wavenet-A",
+        "de-DE-Wavenet-B",
+        "de-DE-Wavenet-C",
+        "de-DE-Wavenet-D",
+        "en-AU-Standard-A",
+        "en-AU-Standard-B",
+        "en-AU-Wavenet-A",
+        "en-AU-Wavenet-B",
+        "en-AU-Wavenet-C",
+        "en-AU-Wavenet-D",
+        "en-AU-Standard-C",
+        "en-AU-Standard-D",
+        "fr-CA-Standard-A",
+        "fr-CA-Standard-B",
+        "fr-CA-Standard-C",
+        "fr-CA-Standard-D",
+        "fr-FR-Standard-C",
+        "fr-FR-Standard-D",
+        "fr-FR-Wavenet-A",
+        "fr-FR-Wavenet-B",
+        "fr-FR-Wavenet-C",
+        "fr-FR-Wavenet-D",
+        "da-DK-Wavenet-A",
+        "pl-PL-Wavenet-A",
+        "pl-PL-Wavenet-B",
+        "pl-PL-Wavenet-C",
+        "pl-PL-Wavenet-D",
+        "pt-PT-Wavenet-A",
+        "pt-PT-Wavenet-B",
+        "pt-PT-Wavenet-C",
+        "pt-PT-Wavenet-D",
+        "ru-RU-Wavenet-A",
+        "ru-RU-Wavenet-B",
+        "ru-RU-Wavenet-C",
+        "ru-RU-Wavenet-D",
+        "sk-SK-Wavenet-A",
+        "tr-TR-Wavenet-A",
+        "tr-TR-Wavenet-B",
+        "tr-TR-Wavenet-C",
+        "tr-TR-Wavenet-D",
+        "tr-TR-Wavenet-E",
+        "uk-UA-Wavenet-A",
+        "ar-XA-Wavenet-A",
+        "ar-XA-Wavenet-B",
+        "ar-XA-Wavenet-C",
+        "cs-CZ-Wavenet-A",
+        "nl-NL-Wavenet-B",
+        "nl-NL-Wavenet-C",
+        "nl-NL-Wavenet-D",
+        "nl-NL-Wavenet-E",
+        "en-IN-Wavenet-A",
+        "en-IN-Wavenet-B",
+        "en-IN-Wavenet-C",
+        "fil-PH-Wavenet-A",
+        "fi-FI-Wavenet-A",
+        "el-GR-Wavenet-A",
+        "hi-IN-Wavenet-A",
+        "hi-IN-Wavenet-B",
+        "hi-IN-Wavenet-C",
+        "hu-HU-Wavenet-A",
+        "id-ID-Wavenet-A",
+        "id-ID-Wavenet-B",
+        "id-ID-Wavenet-C",
+        "it-IT-Wavenet-B",
+        "it-IT-Wavenet-C",
+        "it-IT-Wavenet-D",
+        "ja-JP-Wavenet-B",
+        "ja-JP-Wavenet-C",
+        "ja-JP-Wavenet-D",
+        "cmn-CN-Wavenet-A",
+        "cmn-CN-Wavenet-B",
+        "cmn-CN-Wavenet-C",
+        "cmn-CN-Wavenet-D",
+        "nb-no-Wavenet-E",
+        "nb-no-Wavenet-A",
+        "nb-no-Wavenet-B",
+        "nb-no-Wavenet-C",
+        "nb-no-Wavenet-D",
+        "vi-VN-Wavenet-A",
+        "vi-VN-Wavenet-B",
+        "vi-VN-Wavenet-C",
+        "vi-VN-Wavenet-D",
+        "sr-rs-Standard-A",
+        "lv-lv-Standard-A",
+        "is-is-Standard-A",
+        "bg-bg-Standard-A",
+        "af-ZA-Standard-A",
+        "Tracy",
+        "Danny",
+        "Huihui",
+        "Yaoyao",
+        "Kangkang",
+        "HanHan",
+        "Zhiwei",
+        "Asaf",
+        "An",
+        "Stefanos",
+        "Filip",
+        "Ivan",
+        "Heidi",
+        "Herena",
+        "Kalpana",
+        "Hemant",
+        "Matej",
+        "Andika",
+        "Rizwan",
+        "Lado",
+        "Valluvar",
+        "Linda",
+        "Heather",
+        "Sean",
+        "Michael",
+        "Karsten",
+        "Guillaume",
+        "Pattara",
+        "Jakub",
+        "Szabolcs",
+        "Hoda",
+        "Naayf",
+    ]
+
+    def __init__(self, timeout: int = 20, proxies: dict = None):
+        """Initializes the StreamElements TTS client."""
+        super().__init__()
+        self.session = requests.Session()
+        self.session.headers.update(self.headers)
+        if proxies:
+            self.session.proxies.update(proxies)
+        self.timeout = timeout
+
+    def tts(self, text: str, voice: str = "Mathieu", verbose: bool = True) -> str:
+        """
+        Converts text to speech using the StreamElements API and saves it to a file.
+
+        Args:
+            text (str): The text to convert to speech
+            voice (str): The voice to use for TTS (default: "Mathieu")
+            verbose (bool): Whether to print progress messages (default: True)
+
+        Returns:
+            str: Path to the generated audio file
+        """
+        assert (
+            voice in self.all_voices
+        ), f"Voice '{voice}' not one of [{', '.join(self.all_voices)}]"
+
+        filename = pathlib.Path(tempfile.mktemp(suffix=".mp3", dir=self.temp_dir))
+
+        # Split text into sentences
+        sentences = utils.split_sentences(text)
+
+        # Function to request audio for each chunk
+        def generate_audio_for_chunk(part_text: str, part_number: int):
+            while True:
+                try:
+                    # URL encode the text and voice
+                    encoded_text = urllib.parse.quote(part_text)
+                    encoded_voice = urllib.parse.quote(voice)
+
+                    url = f"https://streamelements.com/tts/{encoded_voice}/{encoded_text}"
+
+                    response = self.session.get(url, headers=self.headers, timeout=self.timeout)
+                    response.raise_for_status()
+
+                    # Check if the request was successful
+                    if response.ok and response.status_code == 200:
+                        if verbose:
+                            print(f"[debug] Chunk {part_number} processed successfully")
+                        return part_number, response.content
+                    else:
+                        if verbose:
+                            print(f"[debug] No data received for chunk {part_number}. Retrying...")
+                except requests.RequestException as e:
+                    if verbose:
+                        print(f"[debug] Error for chunk {part_number}: {e}. Retrying...")
+                    time.sleep(1)
+        try:
+            # Using ThreadPoolExecutor to handle requests concurrently
+            with ThreadPoolExecutor() as executor:
+                futures = {executor.submit(generate_audio_for_chunk, sentence.strip(), chunk_num): chunk_num
+                        for chunk_num, sentence in enumerate(sentences, start=1)}
+
+                # Dictionary to store results with order preserved
+                audio_chunks = {}
+
+                for future in as_completed(futures):
+                    chunk_num = futures[future]
+                    try:
+                        part_number, audio_data = future.result()
+                        audio_chunks[part_number] = audio_data  # Store the audio data in correct sequence
+                    except Exception as e:
+                        if verbose:
+                            print(f"[debug] Failed to generate audio for chunk {chunk_num}: {e}")
+
+            # Combine audio chunks in the correct sequence
+            combined_audio = BytesIO()
+            for part_number in sorted(audio_chunks.keys()):
+                combined_audio.write(audio_chunks[part_number])
+                if verbose:
+                    print(f"[debug] Added chunk {part_number} to the combined file.")
+
+            # Save the combined audio data to a single file
+            with open(filename, 'wb') as f:
+                f.write(combined_audio.getvalue())
+            if verbose:
+                print(f"[debug] Final Audio Saved as {filename}")
+            return filename.as_posix()
+
+        except requests.exceptions.RequestException as e:
+            if verbose:
+                print(f"[debug] Failed to perform the operation: {e}")
+            raise exceptions.FailedToGenerateResponseError(
+                f"Failed to perform the operation: {e}"
+            )
+
+# Example usage
+if __name__ == "__main__":
+    streamelements = StreamElements()
+    text = "This is a test of the StreamElements text-to-speech API. It supports multiple sentences and advanced logging."
+
+    print("[debug] Generating audio...")
+    audio_file = streamelements.tts(text, voice="Mathieu")
+    print(f"Audio saved to: {audio_file}")

webscout/Provider/TTS/utils.py

@@ -0,0 +1,280 @@
+"""
+Text processing utilities for TTS providers.
+"""
+from typing import Union, List, Dict, Tuple, Set, Optional, Pattern
+import re
+
+
+class SentenceTokenizer:
+    """Advanced sentence tokenizer with support for complex cases and proper formatting."""
+    
+    def __init__(self) -> None:
+        # Common abbreviations by category
+        self.TITLES: Set[str] = {
+            'mr', 'mrs', 'ms', 'dr', 'prof', 'rev', 'sr', 'jr', 'esq',
+            'hon', 'pres', 'gov', 'atty', 'supt', 'det', 'rev', 'col','maj', 'gen', 'capt', 'cmdr',
+            'lt', 'sgt', 'cpl', 'pvt'
+        }
+        
+        self.ACADEMIC: Set[str] = {
+            'ph.d', 'phd', 'm.d', 'md', 'b.a', 'ba', 'm.a', 'ma', 'd.d.s', 'dds',
+            'm.b.a', 'mba', 'b.sc', 'bsc', 'm.sc', 'msc', 'llb', 'll.b', 'bl'
+        }
+        
+        self.ORGANIZATIONS: Set[str] = {
+            'inc', 'ltd', 'co', 'corp', 'llc', 'llp', 'assn', 'bros', 'plc', 'cos',
+            'intl', 'dept', 'est', 'dist', 'mfg', 'div'
+        }
+        
+        self.MONTHS: Set[str] = {
+            'jan', 'feb', 'mar', 'apr', 'jun', 'jul', 'aug', 'sep', 'oct', 'nov', 'dec'
+        }
+        
+        self.UNITS: Set[str] = {
+            'oz', 'pt', 'qt', 'gal', 'ml', 'cc', 'km', 'cm', 'mm', 'ft', 'in',
+            'kg', 'lb', 'lbs', 'hz', 'khz', 'mhz', 'ghz', 'kb', 'mb', 'gb', 'tb'
+        }
+        
+        self.TECHNOLOGY: Set[str] = {
+            'v', 'ver', 'app', 'sys', 'dir', 'exe', 'lib', 'api', 'sdk', 'url',
+            'cpu', 'gpu', 'ram', 'rom', 'hdd', 'ssd', 'lan', 'wan', 'sql', 'html'
+        }
+        
+        self.MISC: Set[str] = {
+            'vs', 'etc', 'ie', 'eg', 'no', 'al', 'ca', 'cf', 'pp', 'est', 'st',
+            'approx', 'appt', 'apt', 'dept', 'depts', 'min', 'max', 'avg'
+        }
+
+        # Combine all abbreviations
+        self.all_abbreviations: Set[str] = (
+            self.TITLES | self.ACADEMIC | self.ORGANIZATIONS |
+            self.MONTHS | self.UNITS | self.TECHNOLOGY | self.MISC
+        )
+
+        # Special patterns
+        self.ELLIPSIS: str = r'\.{2,}|…'
+        self.URL_PATTERN: str = (
+            r'(?:https?:\/\/|www\.)[\w\-\.]+\.[a-zA-Z]{2,}(?:\/[^\s]*)?'
+        )
+        self.EMAIL_PATTERN: str = r'[\w\.-]+@[\w\.-]+\.\w+'
+        self.NUMBER_PATTERN: str = (
+            r'\d+(?:\.\d+)?(?:%|°|km|cm|mm|m|kg|g|lb|ft|in|mph|kmh|hz|mhz|ghz)?'
+        )
+        
+        # Quote and bracket pairs
+        self.QUOTE_PAIRS: Dict[str, str] = {
+            '"': '"', "'": "'", '"': '"', "「": "」", "『": "』",
+            "«": "»", "‹": "›", "'": "'", "‚": "'"
+        }
+        
+        self.BRACKETS: Dict[str, str] = {
+            '(': ')', '[': ']', '{': '}', '⟨': '⟩', '「': '」',
+            '『': '』', '【': '】', '〖': '〗', '｢': '｣'
+        }
+
+        # Compile regex patterns
+        self._compile_patterns()
+
+    def _compile_patterns(self) -> None:
+        """Compile regex patterns for better performance."""
+        # Pattern for finding potential sentence boundaries
+        self.SENTENCE_END: Pattern = re.compile(
+            r'''
+            # Group for sentence endings
+            (?:
+                # Standard endings with optional quotes/brackets
+                (?<=[.!?])[\"\'\)\]\}»›」』\s]*
+                
+                # Ellipsis
+                |(?:\.{2,}|…)
+                
+                # Asian-style endings
+                |(?<=[。！？」』】\s])
+            )
+            
+            # Must be followed by whitespace and capital letter or number
+            (?=\s+(?:[A-Z0-9]|["'({[\[「『《‹〈][A-Z]))
+            ''',
+            re.VERBOSE
+        )
+
+        # Pattern for abbreviations
+        abbrev_pattern = '|'.join(re.escape(abbr) for abbr in self.all_abbreviations)
+        self.ABBREV_PATTERN: Pattern = re.compile(
+            fr'\b(?:{abbrev_pattern})\.?',
+            re.IGNORECASE
+        )
+
+    def _protect_special_cases(self, text: str) -> Tuple[str, Dict[str, str]]:
+        """Protect URLs, emails, and other special cases from being split."""
+        protected = text
+        placeholders: Dict[str, str] = {}
+        counter = 0
+
+        # Protect URLs and emails
+        for pattern in [self.URL_PATTERN, self.EMAIL_PATTERN]:
+            for match in re.finditer(pattern, protected):
+                placeholder = f'__PROTECTED_{counter}__'
+                placeholders[placeholder] = match.group()
+                protected = protected.replace(match.group(), placeholder)
+                counter += 1
+
+        # Protect quoted content
+        stack = []
+        protected_chars = list(protected)
+        i = 0
+        while i < len(protected_chars):
+            char = protected_chars[i]
+            if char in self.QUOTE_PAIRS:
+                stack.append((char, i))
+            elif stack and char == self.QUOTE_PAIRS[stack[-1][0]]:
+                start_quote, start_idx = stack.pop()
+                content = ''.join(protected_chars[start_idx:i + 1])
+                placeholder = f'__PROTECTED_{counter}__'
+                placeholders[placeholder] = content
+                protected_chars[start_idx:i + 1] = list(placeholder)
+                counter += 1
+            i += 1
+
+        return ''.join(protected_chars), placeholders
+
+    def _restore_special_cases(self, text: str, placeholders: Dict[str, str]) -> str:
+        """Restore protected content."""
+        restored = text
+        for placeholder, original in placeholders.items():
+            restored = restored.replace(placeholder, original)
+        return restored
+
+    def _handle_abbreviations(self, text: str) -> str:
+        """Handle abbreviations to prevent incorrect sentence splitting."""
+        def replace_abbrev(match: re.Match) -> str:
+            abbr = match.group().lower().rstrip('.')
+            if abbr in self.all_abbreviations:
+                return match.group().replace('.', '__DOT__')
+            return match.group()
+
+        return self.ABBREV_PATTERN.sub(replace_abbrev, text)
+
+    def _normalize_whitespace(self, text: str) -> str:
+        """Normalize whitespace while preserving paragraph breaks."""
+        # Replace multiple newlines with special marker
+        text = re.sub(r'\n\s*\n', ' __PARA__ ', text)
+        # Normalize remaining whitespace
+        text = re.sub(r'\s+', ' ', text)
+        return text.strip()
+
+    def _restore_formatting(self, sentences: List[str]) -> List[str]:
+        """Restore original formatting and clean up sentences."""
+        restored = []
+        for sentence in sentences:
+            # Restore dots in abbreviations
+            sentence = sentence.replace('__DOT__', '.')
+            
+            # Restore paragraph breaks
+            sentence = sentence.replace('__PARA__', '\n\n')
+            
+            # Clean up whitespace
+            sentence = re.sub(r'\s+', ' ', sentence).strip()
+            
+            # Capitalize first letter if it's lowercase and not an abbreviation
+            words = sentence.split()
+            if words and words[0].lower() not in self.all_abbreviations:
+                sentence = sentence[0].upper() + sentence[1:]
+            
+            if sentence:
+                restored.append(sentence)
+        
+        return restored
+
+    def tokenize(self, text: str) -> List[str]:
+        """
+        Split text into sentences while handling complex cases.
+        
+        Args:
+            text (str): Input text to split into sentences.
+            
+        Returns:
+            List[str]: List of properly formatted sentences.
+        """
+        if not text or not text.strip():
+            return []
+
+        # Step 1: Protect special cases
+        protected_text, placeholders = self._protect_special_cases(text)
+        
+        # Step 2: Normalize whitespace
+        protected_text = self._normalize_whitespace(protected_text)
+        
+        # Step 3: Handle abbreviations
+        protected_text = self._handle_abbreviations(protected_text)
+        
+        # Step 4: Split into potential sentences
+        potential_sentences = self.SENTENCE_END.split(protected_text)
+        
+        # Step 5: Process and restore formatting
+        sentences = self._restore_formatting(potential_sentences)
+        
+        # Step 6: Restore special cases
+        sentences = [self._restore_special_cases(s, placeholders) for s in sentences]
+        
+        # Step 7: Post-process sentences
+        final_sentences = []
+        current_sentence = []
+        
+        for sentence in sentences:
+            # Skip empty sentences
+            if not sentence.strip():
+                continue
+                
+            # Check if sentence might be continuation of previous
+            if current_sentence and sentence[0].islower():
+                current_sentence.append(sentence)
+            else:
+                if current_sentence:
+                    final_sentences.append(' '.join(current_sentence))
+                current_sentence = [sentence]
+        
+        # Add last sentence if exists
+        if current_sentence:
+            final_sentences.append(' '.join(current_sentence))
+        
+        return final_sentences
+
+
+def split_sentences(text: str) -> List[str]:
+    """
+    Convenience function to split text into sentences using SentenceTokenizer.
+    
+    Args:
+        text (str): Input text to split into sentences.
+    
+    Returns:
+        List[str]: List of properly formatted sentences.
+    """
+    tokenizer = SentenceTokenizer()
+    return tokenizer.tokenize(text)
+
+
+if __name__ == "__main__":
+    # Test text with various challenging cases
+    test_text: str = """
+    Dr. Smith (Ph.D., M.D.) visited Washington D.C. on Jan. 20, 2024! He met with Prof. Johnson at 3:30 p.m. 
+    They discussed A.I. and machine learning... "What about the U.S. market?" asked Dr. Smith. 
+    The meeting ended at 5 p.m. Later, they went to Mr. Wilson's house (located at 123 Main St.) for dinner.
+    
+    Visit our website at https://www.example.com or email us at test@example.com!
+    The temperature was 72.5°F (22.5°C). The company's Q3 2023 revenue was $12.5M USD.
+    
+    「これは日本語の文章です。」This is a mixed-language text! How cool is that?
+    
+    Some technical specs: CPU: 3.5GHz, RAM: 16GB, Storage: 2TB SSD.
+    Common abbreviations: etc., i.e., e.g., vs., cf., approx. 100 units.
+    """
+    
+    # Process and print each sentence
+    sentences: List[str] = split_sentences(test_text)
+    print("Detected sentences:")
+    print("-" * 80)
+    for i, sentence in enumerate(sentences, 1):
+        print(f"{i}. {sentence}")
+        print("-" * 80)