ralph-coder 0.2.1

use crate::errors::{RalphError, Result};
use ignore::overrides::OverrideBuilder;
use ignore::WalkBuilder;
use regex::Regex;
use std::path::Path;

pub fn read_file(path: &Path) -> Result<String> {
    if !path.exists() {
        return Err(RalphError::FileNotFound(path.display().to_string()));
    }
    std::fs::read_to_string(path).map_err(RalphError::Io)
}

/// Read a file with optional pagination.
///
/// - `offset`: 0-based line index to start from (default: 0).
/// - `limit`:  maximum number of lines to return (default: 150).
///
/// Returns a truncation notice when more lines remain so the caller
/// knows the exact offset to pass in a follow-up call.
pub fn read_file_ranged(
    path: &Path,
    offset: Option<usize>,
    limit: Option<usize>,
) -> Result<String> {
    if !path.exists() {
        return Err(RalphError::FileNotFound(path.display().to_string()));
    }
    let content = std::fs::read_to_string(path).map_err(RalphError::Io)?;

    let all_lines: Vec<&str> = content.lines().collect();
    let total = all_lines.len();

    let start = offset.unwrap_or(0).min(total);
    let max_lines = limit.unwrap_or(150);
    let end = (start + max_lines).min(total);

    let mut out = String::new();
    for line in &all_lines[start..end] {
        out.push_str(line);
        out.push('\n');
    }

    if end < total {
        out.push_str(&format!(
            "\n[...truncated after line {} of {}. \
             To read further use `read_file` with `offset={}`.]",
            end, total, end
        ));
    }

    Ok(out)
}

pub fn list_dir(path: &Path) -> Result<String> {
    if !path.exists() {
        return Err(RalphError::FileNotFound(path.display().to_string()));
    }

    let mut entries = Vec::new();
    let walker = WalkBuilder::new(path)
        .hidden(false)
        .ignore(true)
        .git_ignore(true)
        .max_depth(Some(4))
        .build();

    for entry in walker.flatten() {
        let rel = entry
            .path()
            .strip_prefix(path)
            .unwrap_or(entry.path())
            .display()
            .to_string();
        if rel.is_empty() {
            continue;
        }
        let suffix = if entry.path().is_dir() { "/" } else { "" };
        entries.push(format!("{}{}", rel, suffix));
    }
    entries.sort();
    Ok(entries.join("\n"))
}

pub fn write_file(path: &Path, content: &str) -> Result<()> {
    if let Some(parent) = path.parent() {
        std::fs::create_dir_all(parent)?;
    }
    std::fs::write(path, content).map_err(RalphError::Io)
}

pub fn edit_file(path: &Path, old_string: &str, new_string: &str) -> Result<()> {
    let content = read_file(path)?;
    if !content.contains(old_string) {
        return Err(RalphError::EditNotFound {
            path: path.display().to_string(),
        });
    }
    let updated = content.replacen(old_string, new_string, 1);
    std::fs::write(path, updated).map_err(RalphError::Io)
}

/// Apply multiple (old_string → new_string) replacements to a file in one call.
/// All replacements are validated before any are applied: if any `old_string` is
/// not found, the entire operation fails with an index hint.
pub fn edit_file_multi(path: &Path, edits: &[(String, String)]) -> Result<Vec<String>> {
    let mut content = read_file(path)?;
    let mut applied = Vec::new();
    for (i, (old, new)) in edits.iter().enumerate() {
        if !content.contains(old.as_str()) {
            let hint = find_closest_match_hint(&content, old);
            return Err(RalphError::ToolFailed {
                tool: "edit_file_multi".to_string(),
                message: format!(
                    "edit[{i}]: old_string not found in {}.\n{hint}",
                    path.display()
                ),
            });
        }
        content = content.replacen(old.as_str(), new.as_str(), 1);
        applied.push(format!("edit[{i}] applied"));
    }
    std::fs::write(path, content).map_err(RalphError::Io)?;
    Ok(applied)
}

/// Given a file's content and a failed `old_string`, return a human-readable
/// hint pointing at the closest matching block — helps the LLM fix whitespace
/// or indentation mismatches without re-reading the entire file.
pub fn find_closest_match_hint(content: &str, old_string: &str) -> String {
    let content_lines: Vec<&str> = content.lines().collect();
    let search_lines: Vec<&str> = old_string.lines().collect();

    // Find the first non-blank search line to use as the anchor.
    let anchor = match search_lines.iter().find(|l| !l.trim().is_empty()) {
        Some(l) => l.trim(),
        None => return String::new(),
    };

    // Extract words of length ≥ 3 as "significant tokens".
    let sig_words: Vec<&str> = anchor
        .split(|c: char| !c.is_alphanumeric() && c != '_')
        .filter(|w| w.len() >= 3)
        .collect();

    if sig_words.is_empty() {
        return String::new();
    }

    // Score each content line by how many significant words it shares with the anchor.
    let mut best_score = 0usize;
    let mut best_line = usize::MAX;
    for (i, line) in content_lines.iter().enumerate() {
        let lc = line.to_lowercase();
        let score = sig_words.iter().filter(|w| lc.contains(*w)).count();
        if score > best_score {
            best_score = score;
            best_line = i;
        }
    }

    if best_score == 0 || best_line == usize::MAX {
        return String::new();
    }

    let start = best_line.saturating_sub(2);
    let end = (best_line + search_lines.len() + 3).min(content_lines.len());
    let mut hint = format!("Closest match near line {}:\n", best_line + 1);
    for i in start..end {
        let marker = if i == best_line { ">>>" } else { "   " };
        hint.push_str(&format!("{} {:4}: {}\n", marker, i + 1, content_lines[i]));
    }
    hint
}

/// Return a structural outline of a file — function/class/struct signatures
/// with line numbers, without the bodies. Ideal for surveying a large file
/// before deciding which section to read in full.
pub fn read_file_outline(path: &Path) -> Result<String> {
    if !path.exists() {
        return Err(RalphError::FileNotFound(path.display().to_string()));
    }
    let content = std::fs::read_to_string(path).map_err(RalphError::Io)?;
    let lines: Vec<&str> = content.lines().collect();
    let ext = path.extension().and_then(|e| e.to_str()).unwrap_or("");

    // Per-language signature patterns.
    let pats: &[&str] = match ext {
        "rs" => &[
            r"^\s*(?:pub(?:\([^)]*\))?\s+)?(?:async\s+)?fn\s+\w+",
            r"^\s*(?:pub(?:\([^)]*\))?\s+)?struct\s+\w+",
            r"^\s*(?:pub(?:\([^)]*\))?\s+)?enum\s+\w+",
            r"^\s*(?:pub(?:\([^)]*\))?\s+)?trait\s+\w+",
            r"^\s*(?:pub(?:\([^)]*\))?\s+)?impl(?:<[^>]*>)?\s+\w+",
        ],
        "py" => &[r"^\s*(?:async\s+)?def\s+\w+", r"^\s*class\s+\w+"],
        "js" | "ts" | "tsx" | "jsx" | "mjs" | "cjs" => &[
            r"^\s*(?:export\s+)?(?:default\s+)?(?:async\s+)?function\s+\w+",
            r"^\s*(?:export\s+)?(?:default\s+)?class\s+\w+",
            r"^\s*(?:export\s+)?(?:const|let)\s+\w+\s*[=:]",
        ],
        "go" => &[
            r"^\s*func\s+(?:\([^)]*\)\s+)?\w+",
            r"^\s*type\s+\w+\s+(?:struct|interface)",
        ],
        "java" | "kt" => &[
            r"^\s*(?:(?:public|private|protected|static|final|abstract|override)\s+)*\w+\s+\w+\s*\(",
            r"^\s*(?:public\s+|private\s+|protected\s+)?(?:class|interface|enum|object)\s+\w+",
        ],
        "rb" => &[r"^\s*def\s+\w+", r"^\s*class\s+\w+", r"^\s*module\s+\w+"],
        _ => &[],
    };

    let compiled: Vec<Regex> = pats.iter().filter_map(|p| Regex::new(p).ok()).collect();

    if compiled.is_empty() {
        return Ok(format!(
            "{} ({} lines) — unknown file type; use read_file with offset/limit to navigate",
            path.display(),
            lines.len()
        ));
    }

    let mut out = format!("{} ({} lines)\n", path.display(), lines.len());
    for (i, line) in lines.iter().enumerate() {
        if compiled.iter().any(|re| re.is_match(line)) {
            out.push_str(&format!("{:5}: {}\n", i + 1, line.trim_end()));
        }
    }

    if out.trim_end().ends_with(')') {
        // Nothing matched after the header
        out.push_str("(no recognisable top-level definitions found)\n");
    }

    Ok(out)
}

pub fn delete_file(path: &Path) -> Result<()> {
    if !path.exists() {
        return Err(RalphError::FileNotFound(path.display().to_string()));
    }
    std::fs::remove_file(path).map_err(RalphError::Io)
}

/// Load all files matching a glob pattern (e.g. `src/**/*.rs`), returning
/// their contents concatenated with path headers.  Caps at 100 files and
/// 500 KB total to keep context manageable.
pub fn load_files(pattern: &str, root: &Path) -> Result<String> {
    let mut overrides = OverrideBuilder::new(root);
    overrides.add(pattern).map_err(|e| RalphError::ToolFailed {
        tool: "load_files".to_string(),
        message: format!("Invalid glob pattern: {}", e),
    })?;
    let overrides = overrides.build().map_err(|e| RalphError::ToolFailed {
        tool: "load_files".to_string(),
        message: format!("Failed to build override: {}", e),
    })?;

    let walker = WalkBuilder::new(root)
        .hidden(false)
        .ignore(true)
        .git_ignore(true)
        .overrides(overrides)
        .build();

    let mut output = String::new();
    let mut file_count = 0usize;
    const MAX_FILES: usize = 100;
    const MAX_BYTES: usize = 512 * 1024; // 500 KB

    for entry in walker.flatten() {
        let path = entry.path();
        if !path.is_file() {
            continue;
        }
        let rel = path.strip_prefix(root).unwrap_or(path);
        let content = match std::fs::read_to_string(path) {
            Ok(c) => c,
            Err(_) => continue, // skip binary files
        };

        let header = format!("=== {} ===\n", rel.display());
        if output.len() + header.len() + content.len() > MAX_BYTES {
            output.push_str(&format!(
                "\n[load_files] Output truncated at {} files / 500 KB limit.",
                file_count
            ));
            break;
        }

        output.push_str(&header);
        output.push_str(&content);
        output.push('\n');
        file_count += 1;

        if file_count >= MAX_FILES {
            output.push_str("\n[load_files] Reached 100-file limit.");
            break;
        }
    }

    if output.is_empty() {
        Ok(format!("No files matched pattern: {}", pattern))
    } else {
        Ok(output)
    }
}

// ── Project type detection helpers ───────────────────────────────────────────

struct ProjectInfo {
    kind: &'static str,
    manifest: &'static str,
    src_ext: &'static str,
}

const PROJECT_SIGNATURES: &[ProjectInfo] = &[
    ProjectInfo {
        kind: "Rust",
        manifest: "Cargo.toml",
        src_ext: "rs",
    },
    ProjectInfo {
        kind: "Node.js",
        manifest: "package.json",
        src_ext: "js",
    },
    ProjectInfo {
        kind: "TypeScript",
        manifest: "tsconfig.json",
        src_ext: "ts",
    },
    ProjectInfo {
        kind: "Python",
        manifest: "pyproject.toml",
        src_ext: "py",
    },
    ProjectInfo {
        kind: "Python",
        manifest: "setup.py",
        src_ext: "py",
    },
    ProjectInfo {
        kind: "Go",
        manifest: "go.mod",
        src_ext: "go",
    },
    ProjectInfo {
        kind: "Java",
        manifest: "pom.xml",
        src_ext: "java",
    },
    ProjectInfo {
        kind: "Java",
        manifest: "build.gradle",
        src_ext: "java",
    },
    ProjectInfo {
        kind: "Ruby",
        manifest: "Gemfile",
        src_ext: "rb",
    },
    ProjectInfo {
        kind: "PHP",
        manifest: "composer.json",
        src_ext: "php",
    },
];

/// Analyze the code structure at `root` and return a structured summary that
/// the LLM can use to explain the codebase to the user.
pub fn explain_code(root: &Path) -> Result<String> {
    if !root.exists() {
        return Err(RalphError::FileNotFound(root.display().to_string()));
    }

    let mut report = String::new();

    // ── Project type ──────────────────────────────────────────────────────────
    let mut detected_kind = "Unknown";
    let mut src_ext = "";

    for sig in PROJECT_SIGNATURES {
        let manifest_path = root.join(sig.manifest);
        if manifest_path.exists() {
            detected_kind = sig.kind;
            src_ext = sig.src_ext;
            let manifest_content = std::fs::read_to_string(&manifest_path).unwrap_or_default();
            report.push_str(&format!("## Project Type\n{}\n\n", sig.kind));
            report.push_str(&format!(
                "## {} ({})\n```\n{}\n```\n\n",
                sig.manifest,
                sig.kind,
                manifest_content
                    .lines()
                    .take(50)
                    .collect::<Vec<_>>()
                    .join("\n")
            ));
            break;
        }
    }

    if detected_kind == "Unknown" {
        report.push_str("## Project Type\nUnknown (no recognized manifest file found)\n\n");
    }

    // ── Directory tree ────────────────────────────────────────────────────────
    report.push_str("## Directory Structure\n```\n");
    let walker = WalkBuilder::new(root)
        .hidden(false)
        .ignore(true)
        .git_ignore(true)
        .max_depth(Some(3))
        .build();

    let mut entries: Vec<String> = walker
        .flatten()
        .filter_map(|e| {
            let rel = e.path().strip_prefix(root).ok()?.display().to_string();
            if rel.is_empty() {
                return None;
            }
            let suffix = if e.path().is_dir() { "/" } else { "" };
            Some(format!("{}{}", rel, suffix))
        })
        .collect();
    entries.sort();
    report.push_str(&entries.join("\n"));
    report.push_str("\n```\n\n");

    // ── Source file summaries ─────────────────────────────────────────────────
    if !src_ext.is_empty() {
        report.push_str("## Source Files\n");

        let walker2 = WalkBuilder::new(root)
            .hidden(false)
            .ignore(true)
            .git_ignore(true)
            .build();

        let mut source_files: Vec<std::path::PathBuf> = walker2
            .flatten()
            .filter(|e| {
                e.path().is_file()
                    && e.path()
                        .extension()
                        .and_then(|x| x.to_str())
                        .map(|x| x == src_ext)
                        .unwrap_or(false)
            })
            .map(|e| e.into_path())
            .collect();
        source_files.sort();

        for file in &source_files {
            let rel = file.strip_prefix(root).unwrap_or(file.as_path());
            let content = std::fs::read_to_string(file).unwrap_or_default();
            let lines: Vec<&str> = content.lines().collect();
            let preview: String = lines
                .iter()
                .take(30)
                .cloned()
                .collect::<Vec<_>>()
                .join("\n");
            report.push_str(&format!(
                "\n### {} ({} lines)\n```{}\n{}\n```\n",
                rel.display(),
                lines.len(),
                src_ext,
                preview
            ));
        }
        report.push('\n');
    }

    // ── Entry points ──────────────────────────────────────────────────────────
    let entry_candidates = [
        "src/main.rs",
        "src/lib.rs",
        "main.py",
        "app.py",
        "index.js",
        "index.ts",
        "main.go",
        "main.java",
        "app.rb",
        "index.php",
    ];
    let mut found_entries = Vec::new();
    for candidate in &entry_candidates {
        if root.join(candidate).exists() {
            found_entries.push(*candidate);
        }
    }
    if !found_entries.is_empty() {
        report.push_str("## Entry Points\n");
        for ep in found_entries {
            report.push_str(&format!("- `{}`\n", ep));
        }
        report.push('\n');
    }

    Ok(report)
}