config.example.yaml

# MediBrief Configuration

# API Keys and Authentication
api_keys:
  # PubMed/NCBI E-utilities API key
  # Get from: https://ncbiinsights.ncbi.nlm.nih.gov/2017/11/02/new-api-keys-for-the-e-utilities/
  pubmed: "YOUR_PUBMED_API_KEY"

  # Google Cloud Project ID
  gcp_project_id: "YOUR_GCP_PROJECT_ID"

  # Service account key file path (for local development)
  # For cloud deployment, use service account roles instead
  gcp_service_account_key: "path/to/service-account-key.json"

  # Runway API key
  # Get from: https://app.runwayml.com/settings/api-keys
  runway: "YOUR_RUNWAY_API_KEY"

# PubMed API Settings
pubmed:
  # Maximum number of papers to fetch per request
  max_results_per_query: 10

  # Default search fields
  default_fields:
    - "title"
    - "abstract"
    - "authors"
    - "journal"
    - "publication_date"
    - "doi"

  # Medical specialties to search for
  specialties:
    - "cardiology"
    - "neurology"
    - "oncology"
    - "pediatrics"
    - "psychiatry"
    - "infectious disease"
    - "endocrinology"
    - "gastroenterology"
    - "rheumatology"
    - "pulmonology"

  # Time period for paper search (in days)
  time_period_days: 7

  # Rate limiting (requests per second)
  rate_limit: 3

# PDF Processing
pdf_processing:
  # Temporary storage for downloaded PDFs
  temp_storage_path: "temp/pdfs"

  # OCR settings for image text extraction
  ocr:
    language: "eng"
    config: "--oem 3 --psm 6"

  # Figure extraction settings
  figure_extraction:
    min_figure_size: 100  # Minimum pixel size to consider as figure
    caption_keywords:
      - "Figure"
      - "Fig."
      - "Table"
      - "Chart"
      - "Graph"
      - "Image"

# AI Processing
ai_processing:
  # Vertex AI model settings
  vertex_ai:
    location: "us-central1"
    model_name: "gemini-1.5-pro"

  # Summarization settings
  summarization:
    max_length: 800
    min_length: 500
    temperature: 0.2
    top_p: 0.95

  # Key takeaways settings
  key_takeaways:
    count: 5
    max_length_each: 100

# Image Analysis
image_analysis:
  # Vision AI settings
  vision_ai:
    max_results: 10
    feature_types:
      - "TEXT_DETECTION"
      - "IMAGE_PROPERTIES"
      - "OBJECT_LOCALIZATION"

  # Figure selection settings
  figure_selection:
    max_figures: 5
    min_quality_score: 0.7

# Text-to-Speech
tts:
  # Voice settings
  voice:
    language_code: "en-US"
    name: "en-US-Neural2-F"  # Professional female voice
    speaking_rate: 0.9
    pitch: 0.0

  # Audio settings
  audio:
    encoding: "MP3"
    sample_rate_hertz: 24000

  # Chunk settings for API limits
  max_chunk_length: 5000  # Characters per API request

# Video Generation
video_generation:
  # Output settings
  output:
    resolution: "1080p"  # 1920x1080
    fps: 30
    format: "mp4"

  # Style settings
  style:
    background_color: "#FFFFFF"
    text_color: "#333333"
    highlight_color: "#4285F4"  # Google Blue
    font: "Arial"
    title_font_size: 48
    body_font_size: 32

  # Timing settings
  timing:
    intro_duration: 5  # seconds
    slide_duration: 10  # seconds
    transition_duration: 0.5  # seconds
    outro_duration: 5  # seconds

  # Video generation method
  # Options: "moviepy" or "runway"
  method: "moviepy"

  # Runway Gen-2 API settings
  runway:
    # API settings
    max_retries: 5
    retry_delay: 10  # seconds
    timeout: 120  # seconds

    # Scene generation settings
    scene_duration: 4  # seconds per scene
    max_scenes: 10  # maximum number of scenes to generate

    # Style settings
    style_preset: "medical"  # style preset for scene generation

    # Prompt settings
    prompt_templates:
      title: "Professional medical video title sequence showing '{title}'. Clean, modern design with blue and white color scheme."
      abstract: "Visual representation of medical research about {title}. Professional laboratory setting with researchers analyzing data."
      methods: "Medical laboratory with scientists conducting experiments. Clinical trial visualization with modern equipment and data analysis."
      results: "Data visualization of medical research results. Charts, graphs, and statistical analysis in a professional setting."
      conclusion: "Medical professionals discussing research findings in a conference room. Clinical implementation of research results."

    # Post-processing settings
    add_subtitles: true
    subtitle_font: "Arial"
    subtitle_size: 24
    subtitle_color: "#FFFFFF"
    subtitle_background: "#000000AA"  # Semi-transparent black

# Cloud Storage
cloud_storage:
  # Bucket names
  buckets:
    videos: "medbrief-videos"
    pdfs: "medbrief-pdfs"
    images: "medbrief-images"
    audio: "medbrief-audio"

  # Storage classes
  storage_class: "STANDARD"

  # Retention policy (days)
  retention_days: 30

# YouTube Upload
youtube:
  # Channel settings
  channel_id: "YOUR_YOUTUBE_CHANNEL_ID"

  # Video settings
  video:
    category_id: "27"  # Education
    privacy_status: "public"  # public, unlisted, private
    tags:
      - "medical research"
      - "healthcare"
      - "medicine"
      - "research summary"

  # Description template
  description_template: |
    This video summarizes the research paper: {title}

    Authors: {authors}
    Journal: {journal}
    Publication Date: {publication_date}
    DOI: {doi}

    Key Takeaways:
    {key_takeaways}

    This summary was automatically generated by MediBrief.
    For more medical research summaries, subscribe to our channel.

# Automation Pipeline
pipeline:
  # Schedule settings (cron format)
  schedule: "0 0 * * *"  # Daily at midnight

  # Parallel processing
  max_concurrent_papers: 5

  # Error handling
  max_retries: 3
  retry_delay_seconds: 60

  # Notification settings
  notifications:
    email: "your.email@example.com"
    error_reporting: true
    success_reporting: true

# Logging
logging:
  level: "INFO"  # DEBUG, INFO, WARNING, ERROR, CRITICAL
  format: "%(asctime)s - %(name)s - %(levelname)s - %(message)s"
  file: "logs/medbrief.log"