feat: Parse labels from Blogger

If they have the category prefix, add them to the slice of categories. Otherwise, add them to the slice of tags

cmd/watch.go

@@ -93,7 +93,7 @@ var watchCmd = &cobra.Command{
 					log.Debug("Destination is not Markdown; not cleaning up posts", "destination", dest.GetName())
 				}
 			}
-			
+
 		}
 		wg.Add(1)
 		go func() {

internal/platforms/blogger.go

@@ -117,6 +117,33 @@ func (b Blogger) Pull(options PushPullOptions) (PostData, error) {
 		return PostData{}, fmt.Errorf("updated date not found in response or is not a string")
 	}
 	// result["labels"].([]interface{})
+	labels := []string{}
+	categories := []string{}
+	tags := []string{}
+	// Assert that the labels key is a slice of interfaces
+	// Type assertions don't work on a per-element basis so we have to loop through the slice
+	labelsAsserted, ok := result["labels"].([]interface{})
+	if !ok {
+		log.Infof("No labels found for post %s", title)
+	} else {
+		for _, tag := range labelsAsserted {
+			if tagStr, ok := tag.(string); ok {
+				labels = append(labels, tagStr)
+			} else {
+				log.Warn("Blogger label is not a string", "tag", tag)
+			}
+		}
+		// For each label, check if it has the category prefix
+		// If it does, add it to the categories slice
+		// Otherwise, add it to the tags slice
+		for _, label := range labels {
+			if strings.HasPrefix(label, b.CategoryPrefix) {
+				categories = append(categories, strings.TrimPrefix(label, b.CategoryPrefix))
+			} else {
+				tags = append(tags, label)
+			}
+		}
+	}
 	dateUpdated, err := time.Parse(time.RFC3339, rfcDateUpdated)
 	if err != nil {
 		return PostData{}, err
@@ -173,6 +200,8 @@ func (b Blogger) Pull(options PushPullOptions) (PostData, error) {
 		Date:         date,
 		DateUpdated:  dateUpdated,
 		Description:  postDescription,
+		Categories:   categories,
+		Tags:         tags,
 		CanonicalUrl: canonicalUrl,
 	}, nil
 

internal/platforms/platforms.go

@@ -28,18 +28,17 @@ type WatchableSource interface {
 }
 
 type PushPullOptions struct {
-	AccessToken    string
-	BlogId         string
-	PostUrl        string
-	Filepath       string
-	RefreshToken   string
-	ClientId       string
-	ClientSecret   string
-	LlmProvider    string
-	LlmBaseUrl     string
-	LlmApiKey      string
-	LlmModel       string
-	CategoryPrefix string
+	AccessToken  string
+	BlogId       string
+	PostUrl      string
+	Filepath     string
+	RefreshToken string
+	ClientId     string
+	ClientSecret string
+	LlmProvider  string
+	LlmBaseUrl   string
+	LlmApiKey    string
+	LlmModel     string
 }
 
 type PostData struct {
@@ -50,6 +49,8 @@ type PostData struct {
 	DateUpdated time.Time
 	// TODO: Add frontmatter descriptions
 	Description string
+	Categories  []string
+	Tags        []string
 	// Other fields that are probably needed are canonical URL, publish date, and description
 	CanonicalUrl string
 }
@@ -63,8 +64,9 @@ type PostData struct {
 // }
 
 type Blogger struct {
-	Name    string
-	BlogUrl string
+	Name           string
+	BlogUrl        string
+	CategoryPrefix string
 	// https://developers.google.com/blogger/docs/3.0/reference/posts/delete
 	Overwrite               bool
 	GenerateLlmDescriptions bool
@@ -133,11 +135,19 @@ func CreateSource(sourceMap map[string]interface{}) (Source, error) {
 		if !ok || blogUrl == "" {
 			return nil, fmt.Errorf("blog_url is required for blogger")
 		}
+		// Check if category_prefix is set, if not, set it to null and move on
+		// If the value is not a string, it will be set to "category::"
+		categoryPrefix, ok := sourceMap["category_prefix"].(string)
+		if !ok || categoryPrefix == "" {
+			log.Warn("category_prefix is not a string or is empty. Using default", "default", "category::")
+			categoryPrefix = "category::"
+		}
 		generateLlmDescriptions, _ := sourceMap["generate_llm_descriptions"].(bool)
 		return &Blogger{
 			Name:                    name,
 			BlogUrl:                 blogUrl,
 			GenerateLlmDescriptions: generateLlmDescriptions,
+			CategoryPrefix:          categoryPrefix,
 		}, nil
 	case "markdown":
 		// If the content_dir is not set, set it to null as its not required