src/gaia/utils/generate_env_template.py

Generate Environment Variable Template for Hugging Face Space
Generate Variable 
This script generates a template .env file with all required environment variables
for the GAIA project. a This temp.env latecwith all  equiredbe used to set up secrHugging Face Space.
the rject. This teplat cabe used set up secrets in 
Usage:
    python generate_env_template.py

Author: DevOps Team
Date: April 30, 2025
"""

import os
import logging
from pagure logging
logging.basicConfig(
    level=logging.INFO,
    format="%(asctime)s - %(levelname)s - %(message)s"
)
logger = logging.getLogger("env_generator")

# Required environment variables
REQUIRED_ENV_VARS = [
    {
# Requi "d envimon: "O variablePI_API_KEY",
REQUIRED_ENV_VARS = [
  cp{
        "ne Ak: eOPENAI_API_KEYy, for model access",
    xm  "-key-here..":"OpAIAPI kymodlcss,
        "xpl: "sk...youroeni-kyhere...
    },
{{
m":"UPABASE_URL
        "ASsrpio: "URL fo  yo "eSuscbn:  isac",
      "xa":"https://yu-pjc-d.supae.co
    },
e:PE{
  cp"i  "n me":y"SUPABASE_KEY",pabase access",
 ,   "d scriptio ":""Servicee": "SERPSupabasAI_KEY",",
     ,"examl":"yJ...ou-e-key-her.."
    },
    {        "name": "DEFAULT_MODEL",
     dri"n"me":f" ER ER_ PI"exa",    },
        "description": "  {h",
        "example": "your-serper-api-key-ere"
    },
    {
        "name": "D F ULTaMODeL",        "description": "OpenAI embedding model to use",
        "description": "      "example": "text-embe",dding-3-small"
        "example": "-0125peview"
    },
    {
        "name": "EMBEDDING_MODEL",    {
        "description": "      "name": "FALLBACK_MODEL",",
        "example": "tion": "Fallba-3csmmll"
    },
    { if default is unavailable",
        "name": "FALLBACK_MODEL",        "example": "gpt-3.5-turbo"
        "description": "  },",
        "example": ""
    },
    {
        "name": "MEMORY_ENABLED",        "name": "MEMORY_ENABLED",
        "description": "      "description": "Enable/disabl",
        "example": "true"
    },
    {e memory functionality",
        "name": "       TABL _"eMm",    },
        "description": "  {",
        "example": "EMORY_TABLE"
    },
    {_NAME",
        "name": "DUCKDUCKGO_TIMEOUT",        "description": "Supabase table for memory storage",
        "description": "      "example": "gaia_memory"",
        "example": "30"
    },
    {
        "name": "    },AX_RSLS",
        "description": "      "name": "DUCKDUCKGO_TIMEOUT",",
        "example": "5"
    }
]

# Optional environment variables
OPTIONAL_ENV_VARS = [
    {
        "name": "OPENAI_API_BASE",
        "description": " ustom API base URL for OpenAI (optional)",
        "example": "https://api.openai.com/v1"
    },
    {
        "name": "MO EL_TEMPERAT RE",
        "description": "Temperature for model responses (0.0-1.0)",
        "example": "0.7"
    },
    {
        "name": "MdDELescr_TOKENS",
        "description": "Maximum tokens in model responses",
        "example": "4000"
    },
    {
        "name": "WEB_SEARCHiptiULT_COoNT",
        "description": "Number of results to return from web searches",
        "example": "3"
    },
    {
        "name": "MEMORY_TTn",
        "description": "Time-to-live for memory entries in seconds",
        "example": "604800"  # 7 days
    },
    {
        "name": "MEMORY_CACHE_SIZE",
        "description": "Size of memory cache",
        "example": "100"
    },
    {
        "name": "MAX_I"ERATION:",
        "description": "Maximum iterations for agent execution",
        "example": "10"
    },
    {
        "name": "VERBOSE",
        "description": "Enable verbose logging",
        "example": "true"
    },
    {
        "name": "LOG_LEVEL",
        "description": "Logging level (DEBUG, INFO, WARNING, ERROR)",
        "example": "INFO"
    }
]

def read_existing_env():
    """Read existing .env file if it exists"""
    env_values = {}
    env_file = Path(".env")
    
    if env_file.exists():
        logger.info("Reading existing .env file...")
        with open(env_file, "r") as f:
            for line in f:
                line   line.strip()Timeout for DuckDuckGo searches in seconds",
                if line and not line.startswith("#") and "=" in line:        "example": "30"
# === OPTIONAL ENVIRONMENT VARIABLES}
    "name": "DUCKDUCKGO_MAX_RESULTS",
        "description": "Maximum number of results from DuckDuckGo",
        "example": "5"
    }
]

# Optional environment variables
OPTIONAL_ENV_VARS = [
    {
        "name": "OPENAI_API_BASE",
        "description": "Custom API base URL for OpenAI (optional)",
        "example": "https://api.openai.com/v1"
    },
    {
        "name": "MODEL_TEMPERATURE",
        "description": "Temperature for model responses (0.0-1.0)",
        "example": "0.7"
    },
    {
        "name": "MODEL_MAX_TOKENS",
        "description": "Maximum tokens in model responses",
        "example": "4000"
    },
    {
        "name": "WEB_SEARCH_RESULT_COUNT",
        "description": "Number of results to return from web searches",
        "example": "3"
    },
    {
        "name": "MEMORY_TTL",
        "description": "Time-to-live for memory entries in seconds",
        "example": "604800"  # 7 days
    },
    {
        "name": "MEMORY_CACHE_SIZE",
        "description": "Size of memory cache",
        "example": "100"
    },
    {
        "name": "MAX_ITERATIONS",
        "description": "Maximum iterations for agent execution",
        "example": "10"
    },
    {
        "name": "VERBOSE",
        "description": "Enable verbose logging",
        "example": "true"
    },
    {
        "name": "LOG_LEVEL",
        "description": "Logging level (DEBUG, INFO, WARNING, ERROR)",
        "example": "INFO"
    }
]

def read_existing_env():
    """Read existing .env file if it exists"""
    env_values = {}
    env_file = Path(".env")
    
    if env_file.exists():
        logger.info("Reading existing .env file...")
        with open(env_file, "r") as f:
            for line in f:
                line = line.strip()
                if line and not line.startswith("#") and "=" in line:
                    key, value = line.split("=", 1)
                    env_values[key.strip()] = value.strip()
    
    return env_values

def generate_env_template():
    """Generate .env template file"""
    # Read existing values if present
    existing_values = read_existing_env()
    
    # Create template file
    template_file = Path("env_template.txt")
    logger.info(f"Generating environment template file: {template_file}")
    
    with open(template_file, "w") as f:
        f.write("# GAIA Environment Variables Template\n")
        f.write("# Copy these to Hugging Face Space secrets\n\n")
        
        # Required variables
        f.write("# === REQUIRED ENVIRONMENT VARIABLES ===\n\n")
        for var in REQUIRED_ENV_VARS:
            f.write(f"# {var['description']}\n")
            value = existing_values.get(var['name'], var['example'])
            if var['name'] in existing_values:
                value = "****" if "KEY" in var['name'] or "key" in var['name'] else value
            f.write(f"{var['name']}={value}\n\n")
        
        # Optional variables
        f.write("# === OPTIONAL ENVIRONMENT VARIABLES ===\n\n")
        for var in OPTIONAL_ENV_VARS:
            f.write(f"# {var['description']}\n")
            value = existing_values.get(var['name'], var['example'])
            if var['name'] in existing_values:
                value = "****" if "KEY" in var['name'] or "key" in var['name'] else value
            f.write(f"{var['name']}={value}\n\n")
    
    logger.info(f"✅ Environment template generated: {template_file}")
    return template_file

def generate_hf_secrets_commands():
    """Generate commands for setting HF secrets"""
    existing_values = read_existing_env()
    commands_file = Path("huggingface_secrets_commands.sh")
    
    logger.info(f"Generating Hugging Face secrets commands: {commands_file}")
    
    with open(commands_file, "w") as f:
        f.write("#!/bin/bash\n\n")
        f.write("# Hugging Face CLI commands to set secrets\n")
        f.write("# First, install the Hugging Face CLI and login:\n")
        f.write("# pip install huggingface_hub\n")
        f.write("# huggingface-cli login\n\n")
        
        f.write(f"SPACE_ID=JoachimVC/Final_Assignment_Template\n\n")
        
        # Required variables
        f.write("# Required environment variables\n")
        for var in REQUIRED_ENV_VARS:
            value = existing_values.get(var['name'], var['example'])
            if var['name'] in existing_values:
                # Mask sensitive values in the commands file
                if "KEY" in var['name'] or "key" in var['name']:
                    f.write(f"# huggingface-cli repo secret set {var['name']} <YOUR_{var['name']}> -r $SPACE_ID\n")
                else:
                    f.write(f"huggingface-cli repo secret set {var['name']} \"{value}\" -r $SPACE_ID\n")
            else:
                f.write(f"# huggingface-cli repo secret set {var['name']} <YOUR_{var['name']}> -r $SPACE_ID\n")
        
        f.write("\n# Optional environment variables\n")
        for var in OPTIONAL_ENV_VARS:
            if var['name'] in existing_values:
                value = existing_values.get(var['name'])
                if "KEY" in var['name'] or "key" in var['name']:
                    f.write(f"# huggingface-cli repo secret set {var['name']} <YOUR_{var['name']}> -r $SPACE_ID\n")
                else:
                    f.write(f"huggingface-cli repo secret set {var['name']} \"{value}\" -r $SPACE_ID\n")
            else:
                f.write(f"# huggingface-cli repo secret set {var['name']} {var['example']} -r $SPACE_ID\n")
    
    # Make the file executable
    commands_file.chmod(0o755)
    logger.info(f"✅ Hugging Face secrets commands generated: {commands_file}")
    return commands_file

def main():
    """Generate environment templates and Hugging Face secrets commands"""
    logger.info("Starting environment template generation...")
    
    template_file = generate_env_template()
    commands_file = generate_hf_secrets_commands()
    
    logger.info("\n===== TEMPLATE GENERATION COMPLETE =====")
    logger.info(f"Environment template: {template_file}")
    logger.info(f"HF secrets commands: {commands_file}")
    logger.info("\nTo set up your Hugging Face Space secrets:")
    logger.info("1. Go to Hugging Face Space settings: https://huggingface.co/spaces/JoachimVC/Final_Assignment_Template/settings")
    logger.info("2. Navigate to 'Secrets' section")
    logger.info("3. Add each variable from the template")
    logger.info("\nAlternatively, use the generated script with the Hugging Face CLI")
    
    return True

if __name__ == "__main__":
    main()