review-bot/vcs/util.go

package vcs

import (
	"context"
	"fmt"
	"strings"
)

// GetAllFilesInPath recursively fetches all file contents under a path using the
// provided FileReader. Returns a map of filepath -> content for all files found.
// If the path points to an empty directory, returns an empty map.
func GetAllFilesInPath(ctx context.Context, client FileReader, owner, repo, path string) (map[string]string, error) {
	results := make(map[string]string)

	entries, err := client.ListContents(ctx, owner, repo, path)
	if err != nil {
		return nil, fmt.Errorf("list contents %q: %w", path, err)
	}

	for _, entry := range entries {
		switch entry.Type {
		case "file":
			content, err := client.GetFileContent(ctx, owner, repo, entry.Path, "")
			if err != nil {
				return nil, fmt.Errorf("get file %q: %w", entry.Path, err)
			}
			results[entry.Path] = content
		case "dir":
			subResults, err := GetAllFilesInPath(ctx, client, owner, repo, entry.Path)
			if err != nil {
				return nil, fmt.Errorf("recurse into %q: %w", entry.Path, err)
			}
			for k, v := range subResults {
				results[k] = v
			}
		}
	}

	return results, nil
}

// BuildLineToPositionMap parses a unified diff and returns a map of
// filename -> (new line number -> diff position). The diff position is a
// 1-indexed offset from the @@ hunk header line for each file.
// Only lines that appear in the new file (context lines and additions) are mapped.
// Deletion-only lines are not included.
func BuildLineToPositionMap(diff string) map[string]map[int]int {
	result := make(map[string]map[int]int)

	lines := strings.Split(diff, "\n")
	var currentFile string
	var position int
	var newLine int

	for _, line := range lines {
		// Detect new file in diff
		if strings.HasPrefix(line, "+++ b/") {
			currentFile = strings.TrimPrefix(line, "+++ b/")
			position = 0
			newLine = 0
			if result[currentFile] == nil {
				result[currentFile] = make(map[int]int)
			}
			continue
		}

		// Skip --- lines (old file header)
		if strings.HasPrefix(line, "--- ") {
			continue
		}

		// Skip diff --git lines
		if strings.HasPrefix(line, "diff --git") {
			continue
		}

		// Skip index lines
		if strings.HasPrefix(line, "index ") {
			continue
		}

		// Parse hunk headers
		if strings.HasPrefix(line, "@@") {
			position++
			// Extract new file start line from @@ -a,b +c,d @@
			newLine = parseHunkNewStart(line)
			continue
		}

		// We need a current file to map lines
		if currentFile == "" {
			continue
		}

		// Process diff content lines
		if strings.HasPrefix(line, "+") {
			position++
			result[currentFile][newLine] = position
			newLine++
		} else if strings.HasPrefix(line, "-") {
			position++
			// Deletion lines don't map to new line numbers
		} else if strings.HasPrefix(line, " ") {
			// Context line (space-prefixed)
			if position > 0 {
				position++
				result[currentFile][newLine] = position
				newLine++
			}
		}
	}

	return result
}

// parseHunkNewStart extracts the new-file starting line number from a hunk header.
// Format: @@ -old_start[,old_count] +new_start[,new_count] @@
func parseHunkNewStart(hunkLine string) int {
	// Find the +N part
	plusIdx := strings.Index(hunkLine, "+")
	if plusIdx < 0 {
		return 1
	}
	rest := hunkLine[plusIdx+1:]

	// Read digits until comma or space
	var numStr string
	for _, ch := range rest {
		if ch >= '0' && ch <= '9' {
			numStr += string(ch)
		} else {
			break
		}
	}

	if numStr == "" {
		return 1
	}

	n := 0
	for _, ch := range numStr {
		n = n*10 + int(ch-'0')
	}
	return n
}