compliance-scanner-agent/compliance-graph/src/parsers/registry.rs

use std::collections::HashMap;
use std::path::Path;

use compliance_core::error::CoreError;
use compliance_core::traits::graph_builder::{LanguageParser, ParseOutput};
use tracing::info;

use super::javascript::JavaScriptParser;
use super::python::PythonParser;
use super::rust_parser::RustParser;
use super::typescript::TypeScriptParser;

/// Registry of language parsers, indexed by file extension
pub struct ParserRegistry {
    parsers: Vec<Box<dyn LanguageParser>>,
    extension_map: HashMap<String, usize>,
}

impl ParserRegistry {
    /// Create a registry with all built-in parsers
    pub fn new() -> Self {
        let parsers: Vec<Box<dyn LanguageParser>> = vec![
            Box::new(RustParser::new()),
            Box::new(PythonParser::new()),
            Box::new(JavaScriptParser::new()),
            Box::new(TypeScriptParser::new()),
        ];

        let mut extension_map = HashMap::new();
        for (idx, parser) in parsers.iter().enumerate() {
            for ext in parser.extensions() {
                extension_map.insert(ext.to_string(), idx);
            }
        }

        Self {
            parsers,
            extension_map,
        }
    }

    /// Check if a file extension is supported
    pub fn supports_extension(&self, ext: &str) -> bool {
        self.extension_map.contains_key(ext)
    }

    /// Get supported extensions
    pub fn supported_extensions(&self) -> Vec<&str> {
        self.extension_map.keys().map(|s| s.as_str()).collect()
    }

    /// Parse a file, selecting the appropriate parser by extension
    pub fn parse_file(
        &self,
        file_path: &Path,
        source: &str,
        repo_id: &str,
        graph_build_id: &str,
    ) -> Result<Option<ParseOutput>, CoreError> {
        let ext = file_path.extension().and_then(|e| e.to_str()).unwrap_or("");

        let parser_idx = match self.extension_map.get(ext) {
            Some(idx) => *idx,
            None => return Ok(None),
        };

        let parser = &self.parsers[parser_idx];
        info!(
            file = %file_path.display(),
            language = parser.language(),
            "Parsing file"
        );

        let output = parser.parse_file(file_path, source, repo_id, graph_build_id)?;
        Ok(Some(output))
    }

    /// Parse all supported files in a directory tree
    pub fn parse_directory(
        &self,
        dir: &Path,
        repo_id: &str,
        graph_build_id: &str,
        max_nodes: u32,
    ) -> Result<ParseOutput, CoreError> {
        let mut combined = ParseOutput::default();
        let mut node_count: u32 = 0;

        self.walk_directory(
            dir,
            dir,
            repo_id,
            graph_build_id,
            max_nodes,
            &mut node_count,
            &mut combined,
        )?;

        info!(
            nodes = combined.nodes.len(),
            edges = combined.edges.len(),
            "Directory parsing complete"
        );

        Ok(combined)
    }

    fn walk_directory(
        &self,
        base: &Path,
        dir: &Path,
        repo_id: &str,
        graph_build_id: &str,
        max_nodes: u32,
        node_count: &mut u32,
        combined: &mut ParseOutput,
    ) -> Result<(), CoreError> {
        let entries = std::fs::read_dir(dir).map_err(|e| {
            CoreError::Graph(format!("Failed to read directory {}: {e}", dir.display()))
        })?;

        for entry in entries {
            let entry = entry.map_err(|e| CoreError::Graph(format!("Dir entry error: {e}")))?;
            let path = entry.path();

            // Skip hidden directories and common non-source dirs
            if let Some(name) = path.file_name().and_then(|n| n.to_str()) {
                if name.starts_with('.')
                    || name == "node_modules"
                    || name == "target"
                    || name == "__pycache__"
                    || name == "vendor"
                    || name == "dist"
                    || name == "build"
                    || name == ".git"
                {
                    continue;
                }
            }

            if path.is_dir() {
                self.walk_directory(
                    base,
                    &path,
                    repo_id,
                    graph_build_id,
                    max_nodes,
                    node_count,
                    combined,
                )?;
            } else if path.is_file() {
                if *node_count >= max_nodes {
                    info!(max_nodes, "Reached node limit, stopping parse");
                    return Ok(());
                }

                let ext = path.extension().and_then(|e| e.to_str()).unwrap_or("");
                if !self.supports_extension(ext) {
                    continue;
                }

                // Use relative path from base
                let rel_path = path.strip_prefix(base).unwrap_or(&path);

                let source = match std::fs::read_to_string(&path) {
                    Ok(s) => s,
                    Err(_) => continue, // Skip binary/unreadable files
                };

                if let Some(output) = self.parse_file(rel_path, &source, repo_id, graph_build_id)? {
                    *node_count += output.nodes.len() as u32;
                    combined.nodes.extend(output.nodes);
                    combined.edges.extend(output.edges);
                }
            }
        }

        Ok(())
    }
}

impl Default for ParserRegistry {
    fn default() -> Self {
        Self::new()
    }
}