config.py

from pathlib import Path
from typing import Optional

# Root and main directories
ROOT_DIR = Path(__file__).parent.resolve()
TRANSCRIPTS_DIR = ROOT_DIR / "transcripts"
PROMPTS_DIR = ROOT_DIR / "prompts"
DATA_DIR = ROOT_DIR / "data"

# Model configuration
MODEL_NAME = "gemini-2.5-pro-latest"
PROMPT_VERSION = "v1"
VERSION_ID = f"{MODEL_NAME}_{PROMPT_VERSION}"

# Default channel for testing
DEFAULT_CHANNEL_ID = "UC3l6H774PloAg3XEz46-KQg"

# Batch processing configuration
BATCH_SIZE = 500  # WebShare's concurrency limit

# File naming patterns
TRANSCRIPT_FILENAME = "transcript.txt"
ERROR_FILE_PATTERN = "{video_id}_error.txt"
DOWNLOAD_REPORT_FILENAME = "download_report.json"

def get_transcript_path(channel_id: str, video_id: str) -> Path:
    """Get the path to a video's transcript file"""
    return TRANSCRIPTS_DIR / channel_id / video_id / TRANSCRIPT_FILENAME

def get_analysis_path(channel_id: str, video_id: str) -> Path:
    """Get the path to a video's analysis file"""
    return TRANSCRIPTS_DIR / channel_id / video_id / f"{VERSION_ID}.txt"

def get_channel_analysis_path(channel_id: str) -> Path:
    """Get the path to a channel's aggregated analysis file"""
    return TRANSCRIPTS_DIR / channel_id / f"{VERSION_ID}.txt"

def get_global_analysis_path() -> Path:
    """Get the path to the global aggregated analysis file"""
    return TRANSCRIPTS_DIR / f"{VERSION_ID}.txt"

def get_error_path(channel_id: str, video_id: str) -> Path:
    """Get the path to a video's error log file"""
    error_dir = TRANSCRIPTS_DIR / channel_id / "errors"
    return error_dir / ERROR_FILE_PATTERN.format(video_id=video_id)

def get_download_report_path(channel_id: str) -> Path:
    """Get the path to a channel's download report file"""
    return TRANSCRIPTS_DIR / channel_id / DOWNLOAD_REPORT_FILENAME

def get_prompt_path(version: Optional[str] = None) -> Path:
    """Get the path to a prompt file"""
    version = version or PROMPT_VERSION
    return PROMPTS_DIR / f"{version}.txt"

# Create necessary directories
def ensure_directories():
    """Create all necessary directories if they don't exist"""
    TRANSCRIPTS_DIR.mkdir(parents=True, exist_ok=True)
    PROMPTS_DIR.mkdir(parents=True, exist_ok=True)
    DATA_DIR.mkdir(parents=True, exist_ok=True)

# Call this when the module is imported
ensure_directories()