feat(scripts): support external file embedding

pkg/assetloader/embed.go

@@ -0,0 +1,165 @@
+package assetloader
+
+import (
+	"encoding/base64"
+	"fmt"
+	"os"
+	"path/filepath"
+	"regexp"
+	"strconv"
+	"strings"
+	"unicode/utf8"
+)
+
+const ScriptSourcePathMetadataKey = "source_path"
+
+var (
+	embedDirectivePattern = regexp.MustCompile(`(?m)^\s*--\s*@embed:\s*(.+?)\s*$`)
+	embedAttrPattern      = regexp.MustCompile(`([a-zA-Z_][a-zA-Z0-9_]*)=("[^"]*"|'[^']*'|\S+)`)
+	embedVarPattern       = regexp.MustCompile(`^:[a-zA-Z_][a-zA-Z0-9_]*$`)
+)
+
+// ProcessEmbedDirectives expands SQL comments in the form:
+//
+//	-- @embed: path=... var=:... mode=text|base64
+//
+// Paths are resolved relative to sqlPath. Text mode embeds a quoted UTF-8 SQL
+// string literal. Base64 mode embeds a quoted base64 literal suitable for
+// decode(:var, 'base64').
+func ProcessEmbedDirectives(sqlPath, sql string) (string, error) {
+	directives := embedDirectivePattern.FindAllStringSubmatch(sql, -1)
+	if len(directives) == 0 {
+		return sql, nil
+	}
+
+	if sqlPath == "" {
+		return "", fmt.Errorf("sql path is required for embed directives")
+	}
+
+	result := embedDirectivePattern.ReplaceAllString(sql, "")
+	for i, directive := range directives {
+		literal, placeholder, err := embedDirectiveLiteral(sqlPath, directive[1], i+1)
+		if err != nil {
+			return "", err
+		}
+		if !embedPlaceholderPattern(placeholder).MatchString(result) {
+			return "", fmt.Errorf("%s embed directive %d: placeholder %s not found", sqlPath, i+1, placeholder)
+		}
+		result = replaceEmbedPlaceholder(result, placeholder, literal)
+	}
+
+	return result, nil
+}
+
+func embedDirectiveLiteral(sqlPath, raw string, directiveNumber int) (string, string, error) {
+	attrs, err := parseEmbedAttrs(raw)
+	if err != nil {
+		return "", "", fmt.Errorf("%s embed directive %d: %w", sqlPath, directiveNumber, err)
+	}
+
+	pathValue := attrs["path"]
+	varValue := attrs["var"]
+	modeValue := attrs["mode"]
+	if pathValue == "" {
+		return "", "", fmt.Errorf("%s embed directive %d: missing path", sqlPath, directiveNumber)
+	}
+	if !embedVarPattern.MatchString(varValue) {
+		return "", "", fmt.Errorf("%s embed directive %d: var must be a named placeholder like :asset", sqlPath, directiveNumber)
+	}
+	if modeValue != "text" && modeValue != "base64" {
+		return "", "", fmt.Errorf("%s embed directive %d: mode must be text or base64", sqlPath, directiveNumber)
+	}
+
+	resolved := filepath.Join(filepath.Dir(sqlPath), filepath.Clean(pathValue))
+	data, err := os.ReadFile(resolved)
+	if err != nil {
+		return "", "", fmt.Errorf("%s embed directive %d: reading %s: %w", sqlPath, directiveNumber, resolved, err)
+	}
+
+	switch modeValue {
+	case "text":
+		if !utf8.Valid(data) {
+			return "", "", fmt.Errorf("%s embed directive %d: %s is not valid UTF-8", sqlPath, directiveNumber, resolved)
+		}
+		literal, err := sqlStringLiteral(string(data))
+		if err != nil {
+			return "", "", fmt.Errorf("%s embed directive %d: %w", sqlPath, directiveNumber, err)
+		}
+		return literal, varValue, nil
+	case "base64":
+		literal, err := sqlStringLiteral(base64.StdEncoding.EncodeToString(data))
+		if err != nil {
+			return "", "", fmt.Errorf("%s embed directive %d: %w", sqlPath, directiveNumber, err)
+		}
+		return literal, varValue, nil
+	default:
+		return "", "", fmt.Errorf("%s embed directive %d: mode must be text or base64", sqlPath, directiveNumber)
+	}
+}
+
+func parseEmbedAttrs(raw string) (map[string]string, error) {
+	attrs := map[string]string{}
+	matches := embedAttrPattern.FindAllStringSubmatchIndex(raw, -1)
+	if len(matches) == 0 {
+		return nil, fmt.Errorf("expected path, var, and mode attributes")
+	}
+
+	lastEnd := 0
+	for _, match := range matches {
+		gap := strings.TrimSpace(raw[lastEnd:match[0]])
+		if gap != "" {
+			return nil, fmt.Errorf("invalid attribute syntax near %q", gap)
+		}
+
+		key := raw[match[2]:match[3]]
+		value := raw[match[4]:match[5]]
+		if _, exists := attrs[key]; exists {
+			return nil, fmt.Errorf("duplicate attribute %q", key)
+		}
+		unquoted, err := unquoteEmbedValue(value)
+		if err != nil {
+			return nil, fmt.Errorf("invalid %s value: %w", key, err)
+		}
+		attrs[key] = unquoted
+		lastEnd = match[1]
+	}
+	if tail := strings.TrimSpace(raw[lastEnd:]); tail != "" {
+		return nil, fmt.Errorf("invalid attribute syntax near %q", tail)
+	}
+
+	for key := range attrs {
+		if key != "path" && key != "var" && key != "mode" {
+			return nil, fmt.Errorf("unknown attribute %q", key)
+		}
+	}
+
+	return attrs, nil
+}
+
+func unquoteEmbedValue(value string) (string, error) {
+	if len(value) < 2 {
+		return value, nil
+	}
+	if value[0] == '"' {
+		return strconv.Unquote(value)
+	}
+	if value[0] == '\'' && value[len(value)-1] == '\'' {
+		return value[1 : len(value)-1], nil
+	}
+	return value, nil
+}
+
+func sqlStringLiteral(value string) (string, error) {
+	if strings.ContainsRune(value, '\x00') {
+		return "", fmt.Errorf("embedded text contains NUL byte")
+	}
+	return "'" + strings.ReplaceAll(value, "'", "''") + "'", nil
+}
+
+func replaceEmbedPlaceholder(sql, placeholder, literal string) string {
+	return embedPlaceholderPattern(placeholder).ReplaceAllString(sql, "${1}"+literal+"${2}")
+}
+
+func embedPlaceholderPattern(placeholder string) *regexp.Regexp {
+	return regexp.MustCompile(`(^|[^a-zA-Z0-9_:])` + regexp.QuoteMeta(placeholder) + `([^a-zA-Z0-9_]|$)`)
+}