dd0c/products/06-runbook-automation/agent/src/parser.rs

use serde::{Deserialize, Serialize};
use anyhow::{Context, Result};

/// Parsed runbook step.
#[derive(Debug, Clone, Serialize, Deserialize)]
pub struct RunbookStep {
    pub index: usize,
    pub description: String,
    pub command: String,
    #[serde(default = "default_timeout")]
    pub timeout_seconds: u64,
    #[serde(default)]
    pub on_failure: FailureAction,
    pub condition: Option<String>,
}

fn default_timeout() -> u64 { 300 }

#[derive(Debug, Clone, Serialize, Deserialize)]
#[serde(tag = "action", rename_all = "snake_case")]
pub enum FailureAction {
    Abort,
    Continue,
    Retry { max_attempts: u32 },
}

impl Default for FailureAction {
    fn default() -> Self { FailureAction::Abort }
}

/// Parsed runbook.
#[derive(Debug, Clone, Serialize, Deserialize)]
pub struct Runbook {
    pub name: String,
    #[serde(default)]
    pub description: String,
    #[serde(default = "default_version")]
    pub version: String,
    pub steps: Vec<RunbookStep>,
    #[serde(default)]
    pub variables: std::collections::HashMap<String, VariableSpec>,
}

fn default_version() -> String { "0.1.0".into() }

#[derive(Debug, Clone, Serialize, Deserialize)]
pub struct VariableSpec {
    pub description: String,
    pub default: Option<String>,
    #[serde(default)]
    pub required: bool,
}

/// Parse a YAML runbook into structured steps.
///
/// Example YAML:
/// ```yaml
/// name: restart-service
/// description: Restart a stuck ECS service
/// variables:
///   cluster:
///     description: ECS cluster name
///     required: true
///   service:
///     description: ECS service name
///     required: true
/// steps:
///   - description: Check current task count
///     command: "aws ecs describe-services --cluster {{cluster}} --services {{service}}"
///     timeout_seconds: 30
///   - description: Force new deployment
///     command: "aws ecs update-service --cluster {{cluster}} --service {{service}} --force-new-deployment"
///     timeout_seconds: 60
///     on_failure:
///       action: abort
///   - description: Wait for stable
///     command: "aws ecs wait services-stable --cluster {{cluster}} --services {{service}}"
///     timeout_seconds: 300
///     on_failure:
///       action: retry
///       max_attempts: 3
/// ```
pub fn parse_yaml(content: &str) -> Result<Runbook> {
    let mut runbook: Runbook = serde_yaml::from_str(content)
        .context("Failed to parse runbook YAML")?;

    // Assign step indices
    for (i, step) in runbook.steps.iter_mut().enumerate() {
        step.index = i;
    }

    // Validate: at least one step
    if runbook.steps.is_empty() {
        anyhow::bail!("Runbook must have at least one step");
    }

    Ok(runbook)
}

/// Substitute variables in a command string.
/// Variables use `{{name}}` syntax.
pub fn substitute_variables(
    command: &str,
    variables: &std::collections::HashMap<String, String>,
) -> Result<String> {
    let mut result = command.to_string();
    for (key, value) in variables {
        result = result.replace(&format!("{{{{{}}}}}", key), value);
    }

    // Check for unresolved variables
    if result.contains("{{") {
        let unresolved: Vec<&str> = result
            .match_indices("{{")
            .filter_map(|(start, _)| {
                result[start..].find("}}").map(|end| &result[start..start + end + 2])
            })
            .collect();
        anyhow::bail!("Unresolved variables: {:?}", unresolved);
    }

    Ok(result)
}

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

    const SAMPLE_YAML: &str = r#"
name: restart-service
description: Restart a stuck ECS service
variables:
  cluster:
    description: ECS cluster name
    required: true
  service:
    description: ECS service name
    required: true
steps:
  - description: Check current task count
    command: "aws ecs describe-services --cluster {{cluster}} --services {{service}}"
    timeout_seconds: 30
  - description: Force new deployment
    command: "aws ecs update-service --cluster {{cluster}} --service {{service}} --force-new-deployment"
    timeout_seconds: 60
    on_failure:
      action: abort
  - description: Wait for stable
    command: "aws ecs wait services-stable --cluster {{cluster}} --services {{service}}"
    timeout_seconds: 300
    on_failure:
      action: retry
      max_attempts: 3
"#;

    #[test]
    fn test_parse_valid_yaml() {
        let runbook = parse_yaml(SAMPLE_YAML).unwrap();
        assert_eq!(runbook.name, "restart-service");
        assert_eq!(runbook.steps.len(), 3);
        assert_eq!(runbook.steps[0].index, 0);
        assert_eq!(runbook.steps[2].timeout_seconds, 300);
        assert_eq!(runbook.variables.len(), 2);
        assert!(runbook.variables.get("cluster").unwrap().required);
    }

    #[test]
    fn test_parse_empty_steps_fails() {
        let yaml = "name: empty\nsteps: []";
        assert!(parse_yaml(yaml).is_err());
    }

    #[test]
    fn test_parse_invalid_yaml_fails() {
        assert!(parse_yaml("not: [valid: yaml: {{").is_err());
    }

    #[test]
    fn test_substitute_variables() {
        let mut vars = HashMap::new();
        vars.insert("cluster".into(), "prod".into());
        vars.insert("service".into(), "api".into());

        let result = substitute_variables(
            "aws ecs describe-services --cluster {{cluster}} --services {{service}}",
            &vars,
        ).unwrap();

        assert_eq!(result, "aws ecs describe-services --cluster prod --services api");
    }

    #[test]
    fn test_unresolved_variable_fails() {
        let vars = HashMap::new();
        let result = substitute_variables("echo {{missing}}", &vars);
        assert!(result.is_err());
    }

    #[test]
    fn test_default_failure_action_is_abort() {
        let yaml = r#"
name: simple
steps:
  - description: test
    command: echo hello
"#;
        let runbook = parse_yaml(yaml).unwrap();
        assert!(matches!(runbook.steps[0].on_failure, FailureAction::Abort));
    }

    #[test]
    fn test_default_timeout() {
        let yaml = r#"
name: simple
steps:
  - description: test
    command: echo hello
"#;
        let runbook = parse_yaml(yaml).unwrap();
        assert_eq!(runbook.steps[0].timeout_seconds, 300);
    }
}