gitea-mcp/operation/repo/archive.go

// archive.go - download_repo_archive tool
//
// מטרה: הורדת ריפו שלם כ-zip/tar.gz בקריאה אחת.
// קוראים ל-Gitea archive endpoint ושומרים ל-disk.

package repo

import (
	"context"
	"crypto/sha256"
	"encoding/hex"
	"fmt"
	"net/url"
	"os"
	"path/filepath"
	"strings"

	"gitea.com/gitea/gitea-mcp/pkg/gitea"
	"gitea.com/gitea/gitea-mcp/pkg/log"
	"gitea.com/gitea/gitea-mcp/pkg/params"
	"gitea.com/gitea/gitea-mcp/pkg/to"

	"github.com/mark3labs/mcp-go/mcp"
	"github.com/mark3labs/mcp-go/server"
)

const (
	DownloadRepoArchiveToolName = "download_repo_archive"

	// MaxArchiveSizeBytes — תקרת ביטחון לגודל ארכיון (500MB).
	MaxArchiveSizeBytes = 500 * 1024 * 1024
)

var DownloadRepoArchiveTool = mcp.NewTool(
	DownloadRepoArchiveToolName,
	mcp.WithDescription(
		"Download an entire repository as a zip or tar.gz archive in a single call. "+
			"Much more efficient than calling get_file_contents for each file. "+
			"Saves to the local filesystem; returns the saved path, size and sha256.",
	),
	mcp.WithString("owner", mcp.Required(), mcp.Description("repository owner")),
	mcp.WithString("repo", mcp.Required(), mcp.Description("repository name")),
	mcp.WithString("ref", mcp.Required(), mcp.Description("ref: branch / tag / commit sha")),
	mcp.WithString("format",
		mcp.Description("archive format: zip (default) or tar.gz"),
		mcp.DefaultString("zip"),
	),
	mcp.WithString("output_path",
		mcp.Required(),
		mcp.Description("absolute path on the host where the archive should be saved"),
	),
)

func init() {
	Tool.RegisterRead(server.ServerTool{
		Tool:    DownloadRepoArchiveTool,
		Handler: DownloadRepoArchiveFn,
	})
}

// ArchiveResult — הפלט שמוחזר ל-LLM.
type ArchiveResult struct {
	SavedTo    string `json:"saved_to"`
	SizeBytes  int64  `json:"size_bytes"`
	SHA256     string `json:"sha256"`
	Format     string `json:"format"`
	Repository string `json:"repository"`
	Ref        string `json:"ref"`
}

func DownloadRepoArchiveFn(ctx context.Context, req mcp.CallToolRequest) (*mcp.CallToolResult, error) {
	log.Debugf("Called DownloadRepoArchiveFn")
	args := req.GetArguments()

	owner, err := params.GetString(args, "owner")
	if err != nil {
		return to.ErrorResult(err)
	}
	repo, err := params.GetString(args, "repo")
	if err != nil {
		return to.ErrorResult(err)
	}
	ref, err := params.GetString(args, "ref")
	if err != nil {
		return to.ErrorResult(err)
	}
	outputPath, err := params.GetString(args, "output_path")
	if err != nil {
		return to.ErrorResult(err)
	}

	format, _ := args["format"].(string)
	if format == "" {
		format = "zip"
	}
	if format != "zip" && format != "tar.gz" {
		return to.ErrorResult(fmt.Errorf("invalid format %q: must be 'zip' or 'tar.gz'", format))
	}

	// אבטחה: ודא ש-output_path absolute (לא relative שיוכל לדלוף).
	if !filepath.IsAbs(outputPath) {
		return to.ErrorResult(fmt.Errorf("output_path must be absolute, got: %s", outputPath))
	}

	// ודא שתיקיית היעד קיימת
	parent := filepath.Dir(outputPath)
	if err := os.MkdirAll(parent, 0o755); err != nil {
		return to.ErrorResult(fmt.Errorf("create output dir %q: %w", parent, err))
	}

	// בנה את הנתיב ל-API של Gitea
	// GET /repos/{owner}/{repo}/archive/{ref}.{format}
	apiPath := fmt.Sprintf("repos/%s/%s/archive/%s.%s",
		url.PathEscape(owner),
		url.PathEscape(repo),
		ref,
		format,
	)

	// קבע accept header לפי format
	accept := "application/zip"
	if format == "tar.gz" {
		accept = "application/x-gzip"
	}

	// השתמש ב-DoBytes הקיים ב-rest.go
	archiveBytes, statusCode, err := gitea.DoBytes(ctx, "GET", apiPath, nil, nil, accept)
	if err != nil {
		return to.ErrorResult(fmt.Errorf("download archive (status %d): %w", statusCode, err))
	}

	// בדיקת גודל
	if int64(len(archiveBytes)) > MaxArchiveSizeBytes {
		return to.ErrorResult(fmt.Errorf(
			"archive too large (%d bytes > %d limit). "+
				"Use a more specific ref or fetch individual files via get_files_batch.",
			len(archiveBytes), MaxArchiveSizeBytes,
		))
	}

	// כתיבה ל-disk
	if err := os.WriteFile(outputPath, archiveBytes, 0o644); err != nil {
		return to.ErrorResult(fmt.Errorf("write archive to %q: %w", outputPath, err))
	}

	// חישוב SHA-256 לוידוא integrity
	hash := sha256.Sum256(archiveBytes)
	sha := hex.EncodeToString(hash[:])

	result := ArchiveResult{
		SavedTo:    outputPath,
		SizeBytes:  int64(len(archiveBytes)),
		SHA256:     sha,
		Format:     format,
		Repository: fmt.Sprintf("%s/%s", owner, repo),
		Ref:        ref,
	}

	msg := fmt.Sprintf(
		"Repository archive downloaded successfully:\n"+
			"  Repository: %s\n"+
			"  Ref:        %s\n"+
			"  Format:     %s\n"+
			"  Saved to:   %s\n"+
			"  Size:       %s\n"+
			"  SHA-256:    %s\n\n"+
			"Next steps: extract with %s\n",
		result.Repository,
		result.Ref,
		result.Format,
		result.SavedTo,
		formatBytes(result.SizeBytes),
		result.SHA256,
		extractCommand(format, outputPath),
	)

	return to.TextResult(msg)
}

func formatBytes(b int64) string {
	const unit = 1024
	if b < unit {
		return fmt.Sprintf("%d B", b)
	}
	div, exp := int64(unit), 0
	for n := b / unit; n >= unit; n /= unit {
		div *= unit
		exp++
	}
	return fmt.Sprintf("%.1f %cB",
		float64(b)/float64(div),
		"KMGTPE"[exp],
	)
}

func extractCommand(format, path string) string {
	if strings.HasSuffix(format, "zip") {
		return fmt.Sprintf("unzip %q -d <target_dir>", path)
	}
	return fmt.Sprintf("tar xzf %q -C <target_dir>", path)
}