labmcp/internal/gitexplorer/client.go

package gitexplorer

import (
	"errors"
	"fmt"
	"io"
	"strings"

	"github.com/go-git/go-git/v5"
	"github.com/go-git/go-git/v5/plumbing"
	"github.com/go-git/go-git/v5/plumbing/object"
)

var (
	// ErrNotFound is returned when a ref, commit, or file is not found.
	ErrNotFound = errors.New("not found")
	// ErrFileTooLarge is returned when a file exceeds the size limit.
	ErrFileTooLarge = errors.New("file too large")
)

// GitClient provides read-only access to a git repository.
type GitClient struct {
	repo          *git.Repository
	defaultRemote string
}

// NewGitClient opens a git repository at the given path.
func NewGitClient(repoPath string, defaultRemote string) (*GitClient, error) {
	repo, err := git.PlainOpen(repoPath)
	if err != nil {
		return nil, fmt.Errorf("failed to open repository: %w", err)
	}

	if defaultRemote == "" {
		defaultRemote = "origin"
	}

	return &GitClient{
		repo:          repo,
		defaultRemote: defaultRemote,
	}, nil
}

// ResolveRef resolves a ref (branch, tag, or commit hash) to a commit hash.
func (c *GitClient) ResolveRef(ref string) (*ResolveResult, error) {
	result := &ResolveResult{Ref: ref}

	// Try to resolve as a revision
	hash, err := c.repo.ResolveRevision(plumbing.Revision(ref))
	if err != nil {
		return nil, fmt.Errorf("%w: ref '%s'", ErrNotFound, ref)
	}

	result.Commit = hash.String()

	// Determine the type of ref
	// Check if it's a branch
	if _, err := c.repo.Reference(plumbing.NewBranchReferenceName(ref), true); err == nil {
		result.Type = "branch"
		return result, nil
	}

	// Check if it's a remote branch
	if _, err := c.repo.Reference(plumbing.NewRemoteReferenceName(c.defaultRemote, ref), true); err == nil {
		result.Type = "branch"
		return result, nil
	}

	// Check if it's a tag
	if _, err := c.repo.Reference(plumbing.NewTagReferenceName(ref), true); err == nil {
		result.Type = "tag"
		return result, nil
	}

	// Default to commit
	result.Type = "commit"
	return result, nil
}

// GetLog returns the commit log starting from the given ref.
func (c *GitClient) GetLog(ref string, limit int, author string, since string, path string) ([]LogEntry, error) {
	if limit <= 0 || limit > Limits.MaxLogEntries {
		limit = Limits.MaxLogEntries
	}

	// Resolve the ref to a commit hash
	hash, err := c.repo.ResolveRevision(plumbing.Revision(ref))
	if err != nil {
		return nil, fmt.Errorf("%w: ref '%s'", ErrNotFound, ref)
	}

	logOpts := &git.LogOptions{
		From: *hash,
	}

	// Add path filter if specified
	if path != "" {
		if err := ValidatePath(path); err != nil {
			return nil, err
		}
		logOpts.PathFilter = func(p string) bool {
			return strings.HasPrefix(p, path) || p == path
		}
	}

	iter, err := c.repo.Log(logOpts)
	if err != nil {
		return nil, fmt.Errorf("failed to get log: %w", err)
	}
	defer iter.Close()

	var entries []LogEntry
	err = iter.ForEach(func(commit *object.Commit) error {
		// Apply author filter
		if author != "" {
			authorLower := strings.ToLower(author)
			if !strings.Contains(strings.ToLower(commit.Author.Name), authorLower) &&
				!strings.Contains(strings.ToLower(commit.Author.Email), authorLower) {
				return nil
			}
		}

		// Apply since filter
		if since != "" {
			// Parse since as a ref and check if this commit is reachable
			sinceHash, err := c.repo.ResolveRevision(plumbing.Revision(since))
			if err == nil {
				// Stop if we've reached the since commit
				if commit.Hash == *sinceHash {
					return io.EOF
				}
			}
		}

		// Get first line of commit message as subject
		subject := commit.Message
		if idx := strings.Index(subject, "\n"); idx != -1 {
			subject = subject[:idx]
		}
		subject = strings.TrimSpace(subject)

		entries = append(entries, LogEntry{
			Hash:      commit.Hash.String(),
			ShortHash: commit.Hash.String()[:7],
			Author:    commit.Author.Name,
			Email:     commit.Author.Email,
			Date:      commit.Author.When,
			Subject:   subject,
		})

		if len(entries) >= limit {
			return io.EOF
		}
		return nil
	})

	// io.EOF is expected when we hit the limit
	if err != nil && err != io.EOF {
		return nil, fmt.Errorf("failed to iterate log: %w", err)
	}

	return entries, nil
}

// GetCommitInfo returns full details about a commit.
func (c *GitClient) GetCommitInfo(ref string, includeStats bool) (*CommitInfo, error) {
	hash, err := c.repo.ResolveRevision(plumbing.Revision(ref))
	if err != nil {
		return nil, fmt.Errorf("%w: ref '%s'", ErrNotFound, ref)
	}

	commit, err := c.repo.CommitObject(*hash)
	if err != nil {
		return nil, fmt.Errorf("failed to get commit: %w", err)
	}

	info := &CommitInfo{
		Hash:       commit.Hash.String(),
		Author:     commit.Author.Name,
		Email:      commit.Author.Email,
		Date:       commit.Author.When,
		Committer:  commit.Committer.Name,
		CommitDate: commit.Committer.When,
		Message:    commit.Message,
	}

	for _, parent := range commit.ParentHashes {
		info.Parents = append(info.Parents, parent.String())
	}

	if includeStats {
		stats, err := c.getCommitStats(commit)
		if err == nil {
			info.Stats = stats
		}
	}

	return info, nil
}

// getCommitStats computes file change statistics for a commit.
func (c *GitClient) getCommitStats(commit *object.Commit) (*FileStats, error) {
	stats, err := commit.Stats()
	if err != nil {
		return nil, err
	}

	result := &FileStats{
		FilesChanged: len(stats),
	}

	for _, s := range stats {
		result.Additions += s.Addition
		result.Deletions += s.Deletion
	}

	return result, nil
}

// GetDiffFiles returns the files changed between two commits.
func (c *GitClient) GetDiffFiles(fromRef, toRef string) (*DiffResult, error) {
	fromHash, err := c.repo.ResolveRevision(plumbing.Revision(fromRef))
	if err != nil {
		return nil, fmt.Errorf("%w: ref '%s'", ErrNotFound, fromRef)
	}

	toHash, err := c.repo.ResolveRevision(plumbing.Revision(toRef))
	if err != nil {
		return nil, fmt.Errorf("%w: ref '%s'", ErrNotFound, toRef)
	}

	fromCommit, err := c.repo.CommitObject(*fromHash)
	if err != nil {
		return nil, fmt.Errorf("failed to get from commit: %w", err)
	}

	toCommit, err := c.repo.CommitObject(*toHash)
	if err != nil {
		return nil, fmt.Errorf("failed to get to commit: %w", err)
	}

	patch, err := fromCommit.Patch(toCommit)
	if err != nil {
		return nil, fmt.Errorf("failed to get patch: %w", err)
	}

	result := &DiffResult{
		FromCommit: fromHash.String(),
		ToCommit:   toHash.String(),
	}

	for i, filePatch := range patch.FilePatches() {
		if i >= Limits.MaxDiffFiles {
			break
		}

		from, to := filePatch.Files()

		df := DiffFile{}

		// Determine status and paths
		switch {
		case from == nil && to != nil:
			df.Status = "added"
			df.Path = to.Path()
		case from != nil && to == nil:
			df.Status = "deleted"
			df.Path = from.Path()
		case from != nil && to != nil && from.Path() != to.Path():
			df.Status = "renamed"
			df.Path = to.Path()
			df.OldPath = from.Path()
		default:
			df.Status = "modified"
			if to != nil {
				df.Path = to.Path()
			} else if from != nil {
				df.Path = from.Path()
			}
		}

		// Count additions and deletions
		for _, chunk := range filePatch.Chunks() {
			content := chunk.Content()
			lines := strings.Split(content, "\n")
			switch chunk.Type() {
			case 1: // Add
				df.Additions += len(lines)
			case 2: // Delete
				df.Deletions += len(lines)
			}
		}

		result.Files = append(result.Files, df)
	}

	return result, nil
}

// GetFileAtCommit returns the content of a file at a specific commit.
func (c *GitClient) GetFileAtCommit(ref, path string) (*FileContent, error) {
	if err := ValidatePath(path); err != nil {
		return nil, err
	}

	hash, err := c.repo.ResolveRevision(plumbing.Revision(ref))
	if err != nil {
		return nil, fmt.Errorf("%w: ref '%s'", ErrNotFound, ref)
	}

	commit, err := c.repo.CommitObject(*hash)
	if err != nil {
		return nil, fmt.Errorf("failed to get commit: %w", err)
	}

	file, err := commit.File(path)
	if err != nil {
		return nil, fmt.Errorf("%w: file '%s'", ErrNotFound, path)
	}

	// Check file size
	if file.Size > Limits.MaxFileContent {
		return nil, fmt.Errorf("%w: %d bytes (max %d)", ErrFileTooLarge, file.Size, Limits.MaxFileContent)
	}

	content, err := file.Contents()
	if err != nil {
		return nil, fmt.Errorf("failed to read file: %w", err)
	}

	return &FileContent{
		Path:    path,
		Commit:  hash.String(),
		Size:    file.Size,
		Content: content,
	}, nil
}

// IsAncestor checks if ancestor is an ancestor of descendant.
func (c *GitClient) IsAncestor(ancestorRef, descendantRef string) (*AncestryResult, error) {
	ancestorHash, err := c.repo.ResolveRevision(plumbing.Revision(ancestorRef))
	if err != nil {
		return nil, fmt.Errorf("%w: ref '%s'", ErrNotFound, ancestorRef)
	}

	descendantHash, err := c.repo.ResolveRevision(plumbing.Revision(descendantRef))
	if err != nil {
		return nil, fmt.Errorf("%w: ref '%s'", ErrNotFound, descendantRef)
	}

	ancestorCommit, err := c.repo.CommitObject(*ancestorHash)
	if err != nil {
		return nil, fmt.Errorf("failed to get ancestor commit: %w", err)
	}

	descendantCommit, err := c.repo.CommitObject(*descendantHash)
	if err != nil {
		return nil, fmt.Errorf("failed to get descendant commit: %w", err)
	}

	isAncestor, err := ancestorCommit.IsAncestor(descendantCommit)
	if err != nil {
		return nil, fmt.Errorf("failed to check ancestry: %w", err)
	}

	return &AncestryResult{
		Ancestor:   ancestorHash.String(),
		Descendant: descendantHash.String(),
		IsAncestor: isAncestor,
	}, nil
}

// CommitsBetween returns commits between two refs (exclusive of from, inclusive of to).
func (c *GitClient) CommitsBetween(fromRef, toRef string, limit int) (*CommitRange, error) {
	if limit <= 0 || limit > Limits.MaxLogEntries {
		limit = Limits.MaxLogEntries
	}

	fromHash, err := c.repo.ResolveRevision(plumbing.Revision(fromRef))
	if err != nil {
		return nil, fmt.Errorf("%w: ref '%s'", ErrNotFound, fromRef)
	}

	toHash, err := c.repo.ResolveRevision(plumbing.Revision(toRef))
	if err != nil {
		return nil, fmt.Errorf("%w: ref '%s'", ErrNotFound, toRef)
	}

	iter, err := c.repo.Log(&git.LogOptions{
		From: *toHash,
	})
	if err != nil {
		return nil, fmt.Errorf("failed to get log: %w", err)
	}
	defer iter.Close()

	result := &CommitRange{
		FromCommit: fromHash.String(),
		ToCommit:   toHash.String(),
	}

	err = iter.ForEach(func(commit *object.Commit) error {
		// Stop when we reach the from commit (exclusive)
		if commit.Hash == *fromHash {
			return io.EOF
		}

		subject := commit.Message
		if idx := strings.Index(subject, "\n"); idx != -1 {
			subject = subject[:idx]
		}
		subject = strings.TrimSpace(subject)

		result.Commits = append(result.Commits, LogEntry{
			Hash:      commit.Hash.String(),
			ShortHash: commit.Hash.String()[:7],
			Author:    commit.Author.Name,
			Email:     commit.Author.Email,
			Date:      commit.Author.When,
			Subject:   subject,
		})

		if len(result.Commits) >= limit {
			return io.EOF
		}
		return nil
	})

	if err != nil && err != io.EOF {
		return nil, fmt.Errorf("failed to iterate log: %w", err)
	}

	result.Count = len(result.Commits)
	return result, nil
}

// ListBranches returns all branches in the repository.
func (c *GitClient) ListBranches(includeRemote bool) (*BranchList, error) {
	result := &BranchList{}

	// Get HEAD to determine current branch
	head, err := c.repo.Head()
	if err == nil && head.Name().IsBranch() {
		result.Current = head.Name().Short()
	}

	// List local branches
	branchIter, err := c.repo.Branches()
	if err != nil {
		return nil, fmt.Errorf("failed to list branches: %w", err)
	}

	err = branchIter.ForEach(func(ref *plumbing.Reference) error {
		if len(result.Branches) >= Limits.MaxBranches {
			return io.EOF
		}

		branch := Branch{
			Name:     ref.Name().Short(),
			Commit:   ref.Hash().String(),
			IsRemote: false,
			IsHead:   ref.Name().Short() == result.Current,
		}

		result.Branches = append(result.Branches, branch)
		return nil
	})
	if err != nil && err != io.EOF {
		return nil, fmt.Errorf("failed to iterate branches: %w", err)
	}

	// List remote branches if requested
	if includeRemote {
		refs, err := c.repo.References()
		if err != nil {
			return nil, fmt.Errorf("failed to list references: %w", err)
		}

		err = refs.ForEach(func(ref *plumbing.Reference) error {
			if len(result.Branches) >= Limits.MaxBranches {
				return io.EOF
			}

			if ref.Name().IsRemote() {
				branch := Branch{
					Name:     ref.Name().Short(),
					Commit:   ref.Hash().String(),
					IsRemote: true,
					IsHead:   false,
				}
				result.Branches = append(result.Branches, branch)
			}
			return nil
		})
		if err != nil && err != io.EOF {
			return nil, fmt.Errorf("failed to iterate references: %w", err)
		}
	}

	result.Total = len(result.Branches)
	return result, nil
}

// SearchCommits searches commit messages for a pattern.
func (c *GitClient) SearchCommits(ref, query string, limit int) (*SearchResult, error) {
	if limit <= 0 || limit > Limits.MaxSearchResult {
		limit = Limits.MaxSearchResult
	}

	hash, err := c.repo.ResolveRevision(plumbing.Revision(ref))
	if err != nil {
		return nil, fmt.Errorf("%w: ref '%s'", ErrNotFound, ref)
	}

	iter, err := c.repo.Log(&git.LogOptions{
		From: *hash,
	})
	if err != nil {
		return nil, fmt.Errorf("failed to get log: %w", err)
	}
	defer iter.Close()

	result := &SearchResult{
		Query: query,
	}
	queryLower := strings.ToLower(query)

	// We need to scan more commits to find matches
	scanned := 0
	maxScan := limit * 100 // Scan up to 100x the limit

	err = iter.ForEach(func(commit *object.Commit) error {
		scanned++
		if scanned > maxScan {
			return io.EOF
		}

		// Search in message (case-insensitive)
		if !strings.Contains(strings.ToLower(commit.Message), queryLower) {
			return nil
		}

		subject := commit.Message
		if idx := strings.Index(subject, "\n"); idx != -1 {
			subject = subject[:idx]
		}
		subject = strings.TrimSpace(subject)

		result.Commits = append(result.Commits, LogEntry{
			Hash:      commit.Hash.String(),
			ShortHash: commit.Hash.String()[:7],
			Author:    commit.Author.Name,
			Email:     commit.Author.Email,
			Date:      commit.Author.When,
			Subject:   subject,
		})

		if len(result.Commits) >= limit {
			return io.EOF
		}
		return nil
	})

	if err != nil && err != io.EOF {
		return nil, fmt.Errorf("failed to search commits: %w", err)
	}

	result.Count = len(result.Commits)
	return result, nil
}