zclaw_openfang/crates/zclaw-pipeline/src/parser.rs

//! Pipeline DSL Parser
//!
//! Parses YAML pipeline definitions into Pipeline structs.

use std::path::Path;
use serde_yaml;
use thiserror::Error;

use crate::types::{Pipeline, API_VERSION};

/// Parser errors
#[derive(Debug, Error)]
pub enum ParseError {
    #[error("IO error: {0}")]
    Io(#[from] std::io::Error),

    #[error("YAML parse error: {0}")]
    Yaml(#[from] serde_yaml::Error),

    #[error("Invalid API version: expected '{expected}', got '{actual}'")]
    InvalidVersion { expected: String, actual: String },

    #[error("Invalid kind: expected 'Pipeline', got '{0}'")]
    InvalidKind(String),

    #[error("Missing required field: {0}")]
    MissingField(String),

    #[error("Invalid action type: {0}")]
    InvalidAction(String),

    #[error("Validation error: {0}")]
    Validation(String),
}

/// Pipeline parser
pub struct PipelineParser;

impl PipelineParser {
    /// Parse a pipeline from YAML string
    pub fn parse(yaml: &str) -> Result<Pipeline, ParseError> {
        let pipeline: Pipeline = serde_yaml::from_str(yaml)?;

        // Validate API version
        if pipeline.api_version != API_VERSION {
            return Err(ParseError::InvalidVersion {
                expected: API_VERSION.to_string(),
                actual: pipeline.api_version.clone(),
            });
        }

        // Validate kind
        if pipeline.kind != "Pipeline" {
            return Err(ParseError::InvalidKind(pipeline.kind.clone()));
        }

        // Validate required fields
        if pipeline.metadata.name.is_empty() {
            return Err(ParseError::MissingField("metadata.name".to_string()));
        }

        if pipeline.spec.steps.is_empty() {
            return Err(ParseError::Validation("Pipeline must have at least one step".to_string()));
        }

        // Validate step IDs are unique
        let mut seen_ids = std::collections::HashSet::new();
        for step in &pipeline.spec.steps {
            if !seen_ids.insert(&step.id) {
                return Err(ParseError::Validation(
                    format!("Duplicate step ID: {}", step.id)
                ));
            }
        }

        // Validate input names are unique
        let mut seen_inputs = std::collections::HashSet::new();
        for input in &pipeline.spec.inputs {
            if !seen_inputs.insert(&input.name) {
                return Err(ParseError::Validation(
                    format!("Duplicate input name: {}", input.name)
                ));
            }
        }

        Ok(pipeline)
    }

    /// Parse a pipeline from file
    pub fn parse_file(path: &Path) -> Result<Pipeline, ParseError> {
        let content = std::fs::read_to_string(path)?;
        Self::parse(&content)
    }

    /// Parse and validate all pipelines in a directory
    pub fn parse_directory(dir: &Path) -> Result<Vec<(String, Pipeline)>, ParseError> {
        let mut pipelines = Vec::new();

        if !dir.exists() {
            return Ok(pipelines);
        }

        for entry in std::fs::read_dir(dir)? {
            let entry = entry?;
            let path = entry.path();

            if path.extension().map(|e| e == "yaml" || e == "yml").unwrap_or(false) {
                match Self::parse_file(&path) {
                    Ok(pipeline) => {
                        let filename = path.file_stem()
                            .map(|s| s.to_string_lossy().to_string())
                            .unwrap_or_default();
                        pipelines.push((filename, pipeline));
                    }
                    Err(e) => {
                        tracing::warn!("Failed to parse pipeline {:?}: {}", path, e);
                    }
                }
            }
        }

        Ok(pipelines)
    }
}

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

    #[test]
    fn test_parse_valid_pipeline() {
        let yaml = r#"
apiVersion: zclaw/v1
kind: Pipeline
metadata:
  name: test-pipeline
spec:
  steps:
    - id: step1
      action:
        type: llm_generate
        template: "test"
"#;
        let pipeline = PipelineParser::parse(yaml).unwrap();
        assert_eq!(pipeline.metadata.name, "test-pipeline");
    }

    #[test]
    fn test_parse_invalid_version() {
        let yaml = r#"
apiVersion: invalid/v1
kind: Pipeline
metadata:
  name: test
spec:
  steps: []
"#;
        let result = PipelineParser::parse(yaml);
        assert!(matches!(result, Err(ParseError::InvalidVersion { .. })));
    }

    #[test]
    fn test_parse_invalid_kind() {
        let yaml = r#"
apiVersion: zclaw/v1
kind: NotPipeline
metadata:
  name: test
spec:
  steps: []
"#;
        let result = PipelineParser::parse(yaml);
        assert!(matches!(result, Err(ParseError::InvalidKind(_))));
    }

    #[test]
    fn test_parse_empty_steps() {
        let yaml = r#"
apiVersion: zclaw/v1
kind: Pipeline
metadata:
  name: test
spec:
  steps: []
"#;
        let result = PipelineParser::parse(yaml);
        assert!(matches!(result, Err(ParseError::Validation(_))));
    }

    #[test]
    fn test_parse_duplicate_step_ids() {
        let yaml = r#"
apiVersion: zclaw/v1
kind: Pipeline
metadata:
  name: test
spec:
  steps:
    - id: step1
      action:
        type: llm_generate
        template: "test"
    - id: step1
      action:
        type: llm_generate
        template: "test2"
"#;
        let result = PipelineParser::parse(yaml);
        assert!(matches!(result, Err(ParseError::Validation(_))));
    }
}