charybdis/src/config.rs

/// Configuration management for Charybdis
///
/// This module handles loading configuration from a TOML file with support for
/// environment variable substitution using ${VAR_NAME} syntax.
use anyhow::{Context, Result};
use serde::{Deserialize, Serialize};
use std::collections::HashMap;
use std::fs;
use std::path::Path;
use tracing::{debug, info};

/// Main Charybdis configuration
#[derive(Debug, Clone, Serialize, Deserialize)]
pub struct Config {
    /// Server configuration
    pub server: ServerConfig,

    /// Database configuration
    pub database: DatabaseConfig,

    /// Security configuration (mTLS and RBAC)
    #[serde(default)]
    pub security: crate::security::config::SecurityConfig,

    /// OpenTelemetry configuration
    #[serde(default)]
    pub telemetry: crate::telemetry::config::TelemetryConfig,

    /// Plugin configuration
    #[serde(default)]
    pub plugins: PluginsConfig,
}

/// Server configuration
#[derive(Debug, Clone, Serialize, Deserialize)]
pub struct ServerConfig {
    /// gRPC server host
    #[serde(default = "default_grpc_host")]
    pub grpc_host: String,

    /// gRPC server port
    #[serde(default = "default_grpc_port")]
    pub grpc_port: u16,

    /// YAML adapter configuration
    #[serde(default)]
    pub yaml_adapter: YamlAdapterConfig,
}

/// YAML adapter configuration for Backstage integration
#[derive(Debug, Clone, Serialize, Deserialize)]
pub struct YamlAdapterConfig {
    /// Enable YAML adapter
    #[serde(default = "default_yaml_enabled")]
    pub enabled: bool,

    /// YAML adapter host
    #[serde(default = "default_yaml_host")]
    pub host: String,

    /// YAML adapter port
    #[serde(default = "default_yaml_port")]
    pub port: u16,
}

/// Database configuration
#[derive(Debug, Clone, Serialize, Deserialize)]
pub struct DatabaseConfig {
    /// PostgreSQL connection URL
    /// Supports environment variable substitution: ${DATABASE_URL}
    pub url: String,

    /// Maximum number of connections in the pool
    #[serde(default = "default_max_connections")]
    pub max_connections: u32,

    /// Connection timeout in seconds
    #[serde(default = "default_connection_timeout")]
    pub connection_timeout_secs: u64,
}

/// Plugin system configuration
#[derive(Debug, Clone, Serialize, Deserialize, Default)]
pub struct PluginsConfig {
    /// DefectDojo plugin configuration
    #[serde(default)]
    pub defectdojo: Option<HashMap<String, toml::Value>>,

    /// Dependency-Track plugin configuration
    #[serde(default)]
    pub dependencytrack: Option<HashMap<String, toml::Value>>,

    /// Keycloak plugin configuration
    #[serde(default)]
    pub keycloak: Option<HashMap<String, toml::Value>>,

    /// Custom plugin configurations
    #[serde(flatten)]
    pub custom: HashMap<String, HashMap<String, toml::Value>>,
}

// Default values
fn default_grpc_host() -> String {
    "[::1]".to_string()
}

fn default_grpc_port() -> u16 {
    50051
}

fn default_yaml_enabled() -> bool {
    true
}

fn default_yaml_host() -> String {
    "0.0.0.0".to_string()
}

fn default_yaml_port() -> u16 {
    8080
}

fn default_max_connections() -> u32 {
    10
}

fn default_connection_timeout() -> u64 {
    30
}

impl Default for ServerConfig {
    fn default() -> Self {
        Self {
            grpc_host: default_grpc_host(),
            grpc_port: default_grpc_port(),
            yaml_adapter: YamlAdapterConfig::default(),
        }
    }
}

impl Default for YamlAdapterConfig {
    fn default() -> Self {
        Self {
            enabled: default_yaml_enabled(),
            host: default_yaml_host(),
            port: default_yaml_port(),
        }
    }
}

impl Config {
    /// Load configuration from a TOML file
    ///
    /// Environment variables in the format ${VAR_NAME} will be substituted
    /// with their values from the environment.
    pub fn from_file<P: AsRef<Path>>(path: P) -> Result<Self> {
        let path = path.as_ref();
        info!("Loading configuration from: {}", path.display());

        // Read the file
        let content = fs::read_to_string(path)
            .with_context(|| format!("Failed to read config file: {}", path.display()))?;

        // Substitute environment variables
        let content = Self::substitute_env_vars(&content)?;

        debug!("Parsing configuration");

        // Parse TOML
        let config: Config = toml::from_str(&content)
            .with_context(|| format!("Failed to parse config file: {}", path.display()))?;

        info!("Configuration loaded successfully");
        Ok(config)
    }

    /// Load configuration from default location
    ///
    /// Looks for config in the following order:
    /// 1. ./config.toml (current directory)
    /// 2. ./charybdis.toml
    /// 3. /etc/charybdis/config.toml (Linux/Unix)
    ///
    /// Falls back to environment variables if no config file is found.
    pub fn load() -> Result<Self> {
        let candidates = vec![
            "./config.toml",
            "./charybdis.toml",
            "/etc/charybdis/config.toml",
        ];

        for path in candidates {
            if Path::new(path).exists() {
                return Self::from_file(path);
            }
        }

        info!("No config file found, using environment variables");
        Self::from_env()
    }

    /// Create configuration from environment variables (legacy support)
    pub fn from_env() -> Result<Self> {
        use crate::security::config::SecurityConfig;
        use crate::telemetry::config::TelemetryConfig;

        let database_url = std::env::var("DATABASE_URL")
            .context("DATABASE_URL must be set (or provide config.toml)")?;

        Ok(Config {
            server: ServerConfig {
                grpc_host: std::env::var("GRPC_HOST").unwrap_or_else(|_| default_grpc_host()),
                grpc_port: std::env::var("GRPC_PORT")
                    .ok()
                    .and_then(|s| s.parse().ok())
                    .unwrap_or_else(default_grpc_port),
                yaml_adapter: YamlAdapterConfig::default(),
            },
            database: DatabaseConfig {
                url: database_url,
                max_connections: default_max_connections(),
                connection_timeout_secs: default_connection_timeout(),
            },
            security: SecurityConfig::from_env(),
            telemetry: TelemetryConfig::from_env(),
            plugins: PluginsConfig::default(),
        })
    }

    /// Substitute environment variables in the format ${VAR_NAME} or ${VAR_NAME:-default}
    fn substitute_env_vars(content: &str) -> Result<String> {
        let mut result = content.to_string();
        // Match ${VAR_NAME} or ${VAR_NAME:-default_value}
        let var_pattern =
            regex::Regex::new(r"\$\{([A-Z_][A-Z0-9_]*)(?::-((?:[^}])*))?\}").unwrap();

        for capture in var_pattern.captures_iter(content) {
            let full_match = &capture[0];
            let var_name = &capture[1];
            let default_value = capture.get(2).map(|m| m.as_str());

            match std::env::var(var_name) {
                Ok(value) => {
                    debug!("Substituting ${{{}}}", var_name);
                    result = result.replace(full_match, &value);
                }
                Err(_) => {
                    if let Some(default) = default_value {
                        debug!(
                            "Environment variable not set: ${{{}}}, using default: {}",
                            var_name, default
                        );
                        result = result.replace(full_match, default);
                    } else {
                        debug!("Environment variable not set: ${{{}}}", var_name);
                    }
                }
            }
        }

        Ok(result)
    }
}

#[cfg(test)]
mod tests {
    use super::*;

    #[test]
    fn test_env_var_substitution() {
        // Use PATH which always exists in test environment
        let path_value = std::env::var("PATH").unwrap();

        let input = "some_path = \"${PATH}\"";
        let result = Config::substitute_env_vars(input).unwrap();

        assert!(result.contains(&path_value));
        assert!(!result.contains("${PATH}"));
    }

    #[test]
    fn test_env_var_substitution_missing() {
        // Test that missing env vars are left as-is (allowing optional substitution)
        let input = "url = \"postgresql://${NONEXISTENT_VAR_12345}:password@localhost\"";
        let result = Config::substitute_env_vars(input).unwrap();

        // Variable should remain unchanged when not found
        assert!(result.contains("${NONEXISTENT_VAR_12345}"));
    }

    #[test]
    fn test_env_var_substitution_multiple() {
        // Test multiple variable substitutions
        let path_value = std::env::var("PATH").unwrap();

        let input = "path = \"${PATH}\" and user = \"${USER}\"";
        let result = Config::substitute_env_vars(input).unwrap();

        assert!(result.contains(&path_value));
        assert!(!result.contains("${PATH}"));
    }

    #[test]
    fn test_default_values() {
        let server = ServerConfig::default();
        assert_eq!(server.grpc_port, 50051);
        assert_eq!(server.grpc_host, "[::1]");
    }

    #[test]
    fn test_env_var_with_default_value() {
        let input = "url = \"${NONEXISTENT_VAR_99999:-http://localhost:8080}\"";
        let result = Config::substitute_env_vars(input).unwrap();
        assert_eq!(result, "url = \"http://localhost:8080\"");
    }

    #[test]
    fn test_env_var_with_default_value_overridden() {
        // PATH always exists — use it with a default that should be ignored
        let path_value = std::env::var("PATH").unwrap();
        let input = "url = \"${PATH:-http://fallback:8080}\"";
        let result = Config::substitute_env_vars(input).unwrap();
        assert_eq!(result, format!("url = \"{}\"", path_value));
    }
}