fix: parse pocket new CSV format (#1112)

* fix pocket parsing error

Signed-off-by: bakito <[email protected]>

* add tests forpocket csv

Signed-off-by: bakito <[email protected]>

* Use file name from test case

* fix lint ant test issues

Signed-off-by: bakito <[email protected]>

---------

Signed-off-by: bakito <[email protected]>
Co-authored-by: Felipe Martin <[email protected]>

Author: bakito

internal/cmd/pocket.go

@@ -10,7 +10,6 @@ import (
 	"regexp"
 	"slices"
 	"strconv"
-	"strings"
 	"time"
 
 	"github.com/PuerkitoBio/goquery"
@@ -123,19 +122,23 @@ func parseCsvExport(ctx context.Context, db model.DB, srcFile *os.File) []model.
 		os.Exit(1)
 	}
 
+	var titleIdx, urlIdx, timeAddedIdx, tagsIdx int
 	for i, cols := range records {
 		// Check and skip header
 		if i == 0 {
-			expected := []string{"title", "url", "time_added", "cursor", "tags", "status"}
-			if slices.Compare(cols, expected) != 0 {
-				cError.Printf("Invalid CSV format. Header must be: %s\n", strings.Join(expected, ","))
+			titleIdx = slices.Index(cols, "title")
+			urlIdx = slices.Index(cols, "url")
+			timeAddedIdx = slices.Index(cols, "time_added")
+			tagsIdx = slices.Index(cols, "tags")
+			if titleIdx == -1 || urlIdx == -1 || timeAddedIdx == -1 || tagsIdx == -1 {
+				cError.Printf("Invalid CSV format. Header must contain: title, url, time_added, tags\n")
 				os.Exit(1)
 			}
 			continue
 		}
 
 		// Get metadata
-		title, url, timeAdded, tags, err := verifyMetadata(cols[0], cols[1], cols[2], cols[4])
+		title, url, timeAdded, tags, err := verifyMetadata(cols[titleIdx], cols[urlIdx], cols[timeAddedIdx], cols[tagsIdx])
 		if err != nil {
 			cError.Printf("Skip %s: %v\n", url, err)
 			continue

internal/cmd/pocket_test.go

@@ -0,0 +1,76 @@
+package cmd
+
+import (
+	"context"
+	"os"
+	"path/filepath"
+	"testing"
+
+	"github.com/go-shiori/shiori/internal/database"
+)
+
+func Test_parseCsvExport_old_format(t *testing.T) {
+	tests := []struct {
+		name     string
+		fileName string
+	}{
+		{
+			name:     "Test old file format",
+			fileName: "pocket-old.csv",
+		},
+		{
+			name:     "Test new file format",
+			fileName: "pocket-new.csv",
+		},
+	}
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			file, err := os.Open("../../testdata/" + tt.fileName)
+			if err != nil {
+				t.Error(err.Error())
+			}
+			defer file.Close()
+			ctx := context.TODO()
+
+			tmpDir, err := os.MkdirTemp("", "shiori-test-*")
+			if err != nil {
+				t.Fatalf("failed to create temp dir: %v", err)
+			}
+			defer os.RemoveAll(tmpDir)
+
+			dbPath := filepath.Join(tmpDir, "shiori.db")
+			db, err := database.OpenSQLiteDatabase(ctx, dbPath)
+			if err != nil {
+				t.Fatalf("failed to open sqlite database: %v", err)
+			}
+
+			if err := db.Migrate(ctx); err != nil {
+				t.Fatalf("failed to migrate sqlite database: %v", err)
+			}
+
+			bookmarks := parseCsvExport(ctx, db, file)
+			if len(bookmarks) != 1 {
+				t.Errorf("Expected 1 bookmarks, got %d", len(bookmarks))
+			}
+			bm := bookmarks[0]
+			if bm.Title != "Shiori" {
+				t.Errorf("Expected Title Shiori got %s", bm.URL)
+			}
+			if bm.URL != "https://github.com/go-shiori/shiori" {
+				t.Errorf("Expected URL https://github.com/go-shiori/shiori, got %s", bm.URL)
+			}
+			if len(bm.Tags) != 1 {
+				t.Errorf("Expected 1 tags, got %d", len(bm.Tags))
+			}
+			if bm.Tags[0].Name != "shiori" {
+				t.Errorf("Expected tag shiori, got %s", bm.Tags[0].Name)
+			}
+			if bm.CreatedAt == "" {
+				t.Error("Expected CreatedAt to be not empty")
+			}
+			if bm.ModifiedAt == "" {
+				t.Error("Expected CreatedAt to be not empty")
+			}
+		})
+	}
+}

testdata/pocket-new.csv

@@ -0,0 +1,2 @@
+title,url,time_added,tags,status
+Shiori,https://github.com/go-shiori/shiori,1541343937,shiori,unread

testdata/pocket-old.csv

@@ -0,0 +1,2 @@
+title,url,time_added,cursor,tags,status
+Shiori,https://github.com/go-shiori/shiori,1541343937,,shiori,unread