writing more tests for multimedia. Writing factory functions for test standardization.

Author: the-ride-never-ends

ipfs_datasets_py/multimedia/media_processor.py

@@ -14,7 +14,7 @@
 
 
 def make_media_processor(
-    default_output_dir: Optional[str] = None,
+    default_output_dir: Optional[str|Path] = None,
     enable_logging: bool = True,
     logger: logging.Logger = logging.getLogger(__name__),
     ytdlp: Optional[YtDlpWrapper] = None,
@@ -112,7 +112,7 @@ class MediaProcessor:
         
         # Check available capabilities
         capabilities = processor.get_capabilities()
-        if capabilities["supported_operations"]["download"]:
+        if capabilities["download"]:
             print("Video downloading is available")
         
         # Custom output directory and logging
@@ -130,7 +130,7 @@ class MediaProcessor:
     """
 
     def __init__(self, 
-                default_output_dir: Optional[str] = None,
+                default_output_dir: Optional[str|Path] = None,
                 enable_logging: bool = True,
                 logger: logging.Logger =  logging.getLogger(__name__),
                 ytdlp: Optional[YtDlpWrapper] = None,
@@ -145,7 +145,7 @@ def __init__(self,
         and configures the processor accordingly.
 
         Args:
-            default_output_dir (Optional[str], optional): Default directory path for output files.
+            default_output_dir (Optional[str|Path], optional): Default directory path for output files.
                 Can be relative or absolute path. If relative, it will be resolved relative
                 to the current working directory. If None, uses current working directory.
                 The directory will be created if it doesn't exist. Defaults to None.
@@ -317,7 +317,8 @@ async def download_and_convert(self,
                 "error": str(e)
             }
 
-    def get_capabilities(self) -> Dict[str, Any]:
+    @classmethod
+    def get_capabilities(cls) -> Dict[str, Any]:
         """
         Get comprehensive information about available capabilities and supported operations.
 
@@ -332,35 +333,30 @@ def get_capabilities(self) -> Dict[str, Any]:
 
         Returns:
             Dict[str, Any]: Comprehensive capabilities information with the following structure:
-            {
-                "ytdlp_available": bool,      # Whether yt-dlp backend is available
-                "ffmpeg_available": bool,     # Whether FFmpeg backend is available
-                "supported_operations": {
+                {
                     "download": bool,                    # Video downloading capability
                     "convert": bool,                     # Media conversion capability
-                    "download_and_convert": bool         # Complete workflow capability
                 }
-            }
 
         Examples:
-            >>> capabilities = processor.get_capabilities()
+            >>> processor_can = processor.get_capabilities()
             >>> 
             >>> # Check if downloading is supported
-            >>> if capabilities["supported_operations"]["download"]:
+            >>> if processor_can["download"]:
             ...     print("Video downloading is available")
             ... else:
             ...     print("yt-dlp is required for video downloading")
             >>> 
             >>> # Adapt UI based on capabilities
-            >>> if capabilities["ytdlp_available"] and capabilities["ffmpeg_available"]:
+            >>> if processor_can["download"] and processor_can["convert"]:
             ...     show_full_interface()
-            ... elif capabilities["ytdlp_available"]:
+            ... elif processor_can["download"]:
             ...     show_download_only_interface()
             ... else:
             ...     show_no_capabilities_message()
             >>> 
             >>> # Validate operation before attempting
-            >>> if not capabilities["supported_operations"]["convert"]:
+            >>> if not processor_can["convert"]:
             ...     raise RuntimeError("Conversion not supported - FFmpeg required")
 
         Note:
@@ -371,11 +367,7 @@ def get_capabilities(self) -> Dict[str, Any]:
             - This method is synchronous and performs no external operations
         """
         return {
-            "ytdlp_available": YTDLP_AVAILABLE,
-            "ffmpeg_available": FFMPEG_AVAILABLE,
-            "supported_operations": {
-                "download": YTDLP_AVAILABLE,
-                "convert": FFMPEG_AVAILABLE,
-                "download_and_convert": YTDLP_AVAILABLE and FFMPEG_AVAILABLE
-            }
+            "download": YTDLP_AVAILABLE,
+            "convert": FFMPEG_AVAILABLE,
         }
+

ipfs_datasets_py/multimedia/ytdlp_wrapper.py

@@ -11,7 +11,7 @@
 import time
 import uuid
 from pathlib import Path
-from typing import Any, Callable, Dict, List, Optional
+from typing import Any, Callable, Dict, List, Optional, Literal
 
 
 logger = logging.getLogger(__name__)
@@ -24,6 +24,25 @@
     logger.warning("yt-dlp not available. Install with: pip install yt-dlp")
 
 
+def make_ytdlp_wrapper(
+    default_output_dir: Optional[str] = None,
+    enable_logging: bool = True,
+    default_quality: Literal['best', 'worst'] = "best",
+    logger: logging.Logger = logging.getLogger(__name__),
+    ) -> 'YtDlpWrapper':
+
+    resources = {
+        "default_output_dir": default_output_dir,
+        "enable_logging": enable_logging,
+        "default_quality": default_quality,
+        logger: logger
+    }
+
+    return YtDlpWrapper(**resources)
+
+def type_name(obj: Any) -> str:
+    return type(obj).__name__
+
 class YtDlpWrapper:
     """
     YT-DLP Wrapper for Multi-Platform Media Downloads
@@ -69,6 +88,7 @@ class YtDlpWrapper:
         default_quality (str): Default quality setting for downloads
         downloads (Dict[str, Dict[str, Any]]): Registry tracking all download operations,
             keyed by unique download IDs containing status, timing, and result information
+        logger (logging.Logger): Logger instance for outputting messages
 
     Public Methods:
         download_video(url, output_path=None, quality=None, **kwargs) -> Dict[str, Any]:
@@ -161,7 +181,9 @@ def progress_callback(download_id, progress_data):
     def __init__(self, 
                  default_output_dir: Optional[str] = None,
                  enable_logging: bool = True,
-                 default_quality: str = "best"):
+                 default_quality: Literal['best', 'worst'] = "best",
+                 logger: logging.Logger = logging.getLogger(__name__)
+                 ):
         """
         Initialize YT-DLP wrapper with configuration options and dependency validation.
 
@@ -182,8 +204,6 @@ def __init__(self,
                 not explicitly specified. Supported values include:
                 - 'best': Highest available quality
                 - 'worst': Lowest available quality  
-                - Resolution strings: '720p', '1080p', '480p'
-                - Custom format selectors: 'best[height<=720]'
                 Defaults to "best".
 
         Attributes initialized:
@@ -207,14 +227,14 @@ def __init__(self,
             >>> wrapper = YtDlpWrapper(
             ...     default_output_dir="/home/user/downloads",
             ...     enable_logging=False,
-            ...     default_quality="720p"
+            ...     default_quality="best"
             ... )
             
             >>> # Production configuration
             >>> wrapper = YtDlpWrapper(
             ...     default_output_dir="/var/media/downloads",
             ...     enable_logging=True,
-            ...     default_quality="best[height<=1080]"
+            ...     default_quality="best"
             ... )
 
         Notes:
@@ -223,14 +243,30 @@ def __init__(self,
             - Download tracking dictionary starts empty and accumulates data over time
             - All paths are converted to Path objects for cross-platform compatibility
         """
-        self.default_output_dir = Path(default_output_dir or tempfile.gettempdir())
-        self.enable_logging = enable_logging
-        self.default_quality = default_quality
-        self.downloads = {}  # Track active downloads
-        
         if not YTDLP_AVAILABLE:
             raise ImportError("yt-dlp is required but not installed. Install with: pip install yt-dlp")
-    
+        
+        if default_output_dir is not None and not isinstance(default_output_dir, str):
+            raise TypeError(f"default_output_dir must be a string, got {type_name(default_output_dir)}")
+
+        if not isinstance(default_quality, str):
+            raise TypeError(f"default_quality must be a string, got {type_name(default_quality)}")
+
+        if default_quality not in ['best', 'worst']:
+            raise ValueError(f"default_quality must be 'best' or 'worst', got {default_quality}")
+
+        if not isinstance(enable_logging, bool):
+            raise TypeError(f"enable_logging must be a boolean, got {type_name(enable_logging)}")
+
+        if logger is not None and not isinstance(logger, logging.Logger):
+            raise TypeError(f"logger must be a logging.Logger instance, got {type_name(logger)}")
+
+        self.default_output_dir: Path = Path(default_output_dir or tempfile.gettempdir())
+        self.enable_logging: bool = enable_logging
+        self.default_quality: Literal['best','worst'] = default_quality
+        self.logger: logging.Logger = logger
+        self.downloads = {}  # Track active downloads
+
     async def download_video(self, 
                            url: str,
                            output_path: Optional[str] = None,

tests/_test_utils.py

@@ -18,6 +18,19 @@
 from typing import Literal
 
 
+try:
+    from nltk import word_tokenize, sent_tokenize
+except ImportError:
+    raise ImportError(
+        "The 'nltk' package is required for this module. "
+        "Please install it using 'pip install nltk'."
+    )
+
+
+
+
+
+
 class FalsifiedCodeError(NotImplementedError):
     """Custom exception for falsified code."""
     pass