sm-seo-plugin/classes/csv_test.go

package classes

import (
	"errors"
	"reflect"
	"strings"
	"testing"
)

// TestCSV: the export writes a BOM, the delimiter and escaped formula
// cells; the import strips the BOM, detects or takes the delimiter, reads
// quoting, unescapes formula cells and refuses what is not CSV.
func TestCSV(t *testing.T) {
	header := []string{"source", "target", "note"}
	rows := [][]string{
		{"/a", "/b", "plain"},
		{"/ł", "/ż", "zażółć; gęślą, jaźń"},
		{"/q", "/r", `she said "hi"` + "\nnext line"},
		{"/f", "/g", "=HYPERLINK(\"x\")"},
		{"/p", "/s", "+1"},
		{"/m", "/n", "-1"},
		{"/at", "/x", "@SUM(A1)"},
		{"/tab", "/x", "\tlead"},
		{"/cr", "/x", "\rlead"},
		{"/ok", "/x", "a=b"},
		{"/ap", "/x", "'=x"},
		{"/ap2", "/x", "''-)"},
		{"/ap3", "/x", "'plain"},
	}

	t.Run("round trip with each delimiter", func(t *testing.T) {
		for _, delim := range []string{",", ";", ""} {
			out, err := WriteCSV(header, rows, delim)
			if err != nil {
				t.Fatal(err)
			}
			if !strings.HasPrefix(out, BOM) {
				t.Fatalf("%q: no BOM", delim)
			}
			sep := ","
			if delim == ";" {
				sep = ";"
			}
			if !strings.HasPrefix(strings.TrimPrefix(out, BOM), strings.Join(header, sep)+"\n") {
				t.Fatalf("%q: header line %q", delim, strings.SplitN(out, "\n", 2)[0])
			}
			gotHeader, got, err := ReadCSV(out, delim)
			if err != nil || !reflect.DeepEqual(gotHeader, header) || len(got) != len(rows) {
				t.Fatalf("%q: ReadCSV = %v, %d rows, %v", delim, gotHeader, len(got), err)
			}
			for i, row := range rows {
				// Escaped cells keep their leading tab or CR: the apostrophe
				// protects them from the import's trimming.
				for j, col := range header {
					if want := row[j]; got[i].Fields[col] != want {
						t.Fatalf("%q row %d %s = %q, want %q", delim, i, col, got[i].Fields[col], want)
					}
				}
			}
		}
	})

	t.Run("formula cells are escaped on export", func(t *testing.T) {
		out, _ := WriteCSV([]string{"note"}, [][]string{{"=1+1"}, {"+x"}, {"-x"}, {"@x"}, {"a=b"}, {""}}, ",")
		lines := strings.Split(strings.TrimPrefix(out, BOM), "\n")
		if lines[1] != "'=1+1" || lines[2] != "'+x" || lines[3] != "'-x" || lines[4] != "'@x" || lines[5] != "a=b" {
			t.Fatalf("export = %q", lines)
		}
		for in, want := range map[string]string{"'=x": "=x", "'+x": "+x", "'abc": "'abc", "''=x": "'=x", "plain": "plain", "'": "'"} {
			if got := unescapeCell(in); got != want {
				t.Errorf("unescapeCell(%q) = %q, want %q", in, got, want)
			}
		}
	})

	t.Run("delimiter detection and the setting", func(t *testing.T) {
		if CSVDelimiter(";") != ';' || CSVDelimiter(" ; ") != ';' || CSVDelimiter(",") != ',' || CSVDelimiter("") != ',' || CSVDelimiter("|") != ',' {
			t.Fatal("CSVDelimiter")
		}
		h, rows, err := ReadCSV("Source;Target\n/a;/b\n", "")
		if err != nil || !reflect.DeepEqual(h, []string{"source", "target"}) || rows[0].Fields["target"] != "/b" {
			t.Fatalf("detected ; = %v %+v %v", h, rows, err)
		}
		h, rows, err = ReadCSV(BOM+" SOURCE , Target \n/a,/b\n", "")
		if err != nil || !reflect.DeepEqual(h, []string{"source", "target"}) || rows[0].Line != 2 {
			t.Fatalf("detected , = %v %+v %v", h, rows, err)
		}
	})

	t.Run("quoting, blank rows and extra cells", func(t *testing.T) {
		data := "source,target,note\n\"/a,1\",\"/b\"\"x\",\"two\nlines\"\n,,\n\n/c,/d,n,extra\n"
		_, rows, err := ReadCSV(data, ",")
		if err != nil || len(rows) != 2 {
			t.Fatalf("rows = %+v, %v", rows, err)
		}
		if rows[0].Fields["source"] != "/a,1" || rows[0].Fields["target"] != `/b"x` || rows[0].Fields["note"] != "two\nlines" || rows[0].Line != 2 {
			t.Fatalf("quoted row = %+v", rows[0])
		}
		if rows[1].Fields["source"] != "/c" || len(rows[1].Fields) != 3 || rows[1].Line != 6 {
			t.Fatalf("row with an extra cell = %+v", rows[1])
		}
	})

	t.Run("unreadable files", func(t *testing.T) {
		for name, data := range map[string]string{
			"empty":               "",
			"a BOM only":          BOM,
			"blank":               " \n\t\n",
			"a bare quote":        "source,target\n\"/a,/b\n",
			"the wrong delimiter": "source;target\n/a;/b\n",
		} {
			delim := ""
			if name == "the wrong delimiter" {
				delim = ","
			}
			if _, _, err := ReadCSV(data, delim); !errors.Is(err, ErrUnreadable) {
				t.Errorf("%s: err = %v", name, err)
			}
		}
	})
}