2 files changed, 38 insertions(+), 17 deletions(-)

M go.mod
M spammer.go

@@ 19,6 19,7 @@ require (
 	github.com/go-ap/jsonld v0.0.0-20221030091449-f2a191312c73 // indirect
 	github.com/go-chi/chi/v5 v5.0.12 // indirect
 	github.com/golang/protobuf v1.5.4 // indirect
+	github.com/grokify/html-strip-tags-go v0.1.0 // indirect
 	github.com/mattn/go-colorable v0.1.13 // indirect
 	github.com/mattn/go-isatty v0.0.20 // indirect
 	github.com/rs/xid v1.5.0 // indirect

@@ 16,6 16,7 @@ import (
 	"net/http"
 	"net/url"
 	"os"
+	"path/filepath"
 	"strings"
 	"sync"
 	"time"


@@ 24,6 25,7 @@ import (
 	ap "github.com/go-ap/activitypub"
 	"github.com/go-ap/client"
 	"github.com/go-ap/errors"
+	strip "github.com/grokify/html-strip-tags-go"
 	"github.com/mariusor/spammy/indieauth"
 	"github.com/mariusor/spammy/internal/names"
 	"golang.org/x/oauth2"


@@ 218,14 220,13 @@ func RandomObject(parent ap.Item) ap.Item {
 		buf := make([]byte, base64.RawStdEncoding.EncodedLen(len(data)))
 		base64.RawStdEncoding.Encode(buf, data)
 		data = buf
+		docTyp, ttyp := filepath.Split(string(mime))
 		ob.Name = ap.NaturalLanguageValues{
-			{ap.NilLangRef, []byte(fmt.Sprintf("random %s", mime))},
+			{ap.NilLangRef, []byte(fmt.Sprintf("Random %s %s", strings.Trim(docTyp, "/"), strings.Trim(ttyp, "/")))},
 		}
 	} else {
-		ob.Name = getNameFromData(data, 5)
-		ob.Summary = ap.NaturalLanguageValues{
-			{ap.NilLangRef, data[:bytes.Index(data, []byte{'.'})]},
-		}
+		ob.Name = getNameFromData(data, mime, 5)
+		ob.Summary = getSummaryFromData(data, mime)
 	}
 	ob.Content = ap.NaturalLanguageValues{
 		{ap.NilLangRef, data},


@@ 234,20 235,39 @@ func RandomObject(parent ap.Item) ap.Item {
 	return ob
 }
 
-func getNameFromData(data []byte, wc int) ap.NaturalLanguageValues {
-	wordCnt := int(math.Min(float64(bytes.Count(data, []byte(" "))), float64(wc)))
-	w := 0
-	idx := bytes.IndexFunc(data, func(r rune) bool {
-		if r == ' ' && w == wordCnt {
-			return true
+func getSummaryFromData(data []byte, mimeType ap.MimeType) ap.NaturalLanguageValues {
+	if strings.Contains(string(mimeType), "html") {
+		data = []byte(strip.StripTags(string(data)))
+	}
+	return ap.NaturalLanguageValues{
+		{ap.NilLangRef, data[:bytes.Index(data, []byte{'.'})]},
+	}
+}
+
+func getNameFromData(data []byte, mimeType ap.MimeType, wc int) ap.NaturalLanguageValues {
+	mime := string(mimeType)
+	name := []byte(fmt.Sprintf("Unknown %q item", mime))
+	if !strings.Contains(mime, "text") {
+		docTyp, typ := filepath.Split(mime)
+		name = []byte(fmt.Sprintf("Random %s %s", strings.Trim(typ, "/"), strings.Trim(docTyp, "/")))
+	} else {
+		if strings.Contains(mime, "html") {
+			data = []byte(strip.StripTags(string(data)))
 		}
-		w++
-		return false
-	})
-	if idx <= 0 {
-		return nil
+		wordCnt := int(math.Min(float64(bytes.Count(data, []byte(" "))), float64(wc)))
+		w := 0
+		idx := bytes.IndexFunc(data, func(r rune) bool {
+			if r == ' ' && w == wordCnt {
+				return true
+			}
+			w++
+			return false
+		})
+		if idx <= 0 {
+			return nil
+		}
+		name = data[:idx]
 	}
-	name := data[:idx]
 	return ap.NaturalLanguageValues{
 		{ap.NilLangRef, name},
 	}