Refactor conversion pipeline by responsibility

cmd/dataxl/conversion.go

@@ -0,0 +1,71 @@
+package main
+
+import (
+	"fmt"
+	"path/filepath"
+	"strings"
+)
+
+// convert routes data through either the structured-value or table internal
+// representation. Keeping this matrix in one place makes format adapters
+// independent from CLI and file I/O concerns.
+func convert(input []byte, from, to, sheet string, pretty bool) ([]byte, error) {
+	switch {
+	case isTabular(from) && isTabular(to):
+		t, err := parseTable(input, from, sheet)
+		if err != nil {
+			return nil, err
+		}
+		return encodeTable(t, to, sheet)
+	case isTabular(from):
+		t, err := parseTable(input, from, sheet)
+		if err != nil {
+			return nil, err
+		}
+		return encodeStructured(tableToRecords(t), to, pretty)
+	case isTabular(to):
+		value, err := parseStructured(input, from)
+		if err != nil {
+			return nil, err
+		}
+		return encodeTable(valueToTable(value), to, sheet)
+	default:
+		value, err := parseStructured(input, from)
+		if err != nil {
+			return nil, err
+		}
+		return encodeStructured(value, to, pretty)
+	}
+}
+
+func normalizeFormat(format string) string {
+	format = strings.ToLower(strings.TrimPrefix(strings.TrimSpace(format), "."))
+	switch format {
+	case "yml":
+		return "yaml"
+	case "xlsm", "xls":
+		return "xlsx"
+	default:
+		return format
+	}
+}
+
+func inferFormat(path string) string {
+	if path == "" || path == "-" {
+		return ""
+	}
+	return normalizeFormat(filepath.Ext(path))
+}
+
+func validateFormat(format string) error {
+	switch format {
+	case "json", "yaml", "toml", "csv", "tsv", "xlsx":
+		return nil
+	default:
+		return fmt.Errorf("unsupported format %q", format)
+	}
+}
+
+func isTabular(format string) bool {
+	return format == "csv" || format == "tsv" || format == "xlsx"
+}