this repo has no description
Improve wikipedia preview handling
+42
-2
+3
-1
internal/assets/css/screen.css
+3
-1
internal/assets/css/screen.css
···
303
303
.og-description {
304
304
font-size: 15px;
305
305
line-height: 22px;
306
-
color: #1d1c1d;
306
+
color: #2d2d2d; /* Slightly softer than #1d1c1d */
307
307
margin-bottom: 8px;
308
+
font-family: -apple-system, BlinkMacSystemFont, "Segoe UI", Roboto, Helvetica, Arial, sans-serif;
309
+
letter-spacing: normal;
308
310
}
309
311
310
312
.og-image {
+39
-1
internal/handler/preview.go
+39
-1
internal/handler/preview.go
···
29
29
}
30
30
31
31
// Fetch data
32
-
resp, err := http.Get(urlParam)
32
+
req, err := http.NewRequest("GET", urlParam, nil)
33
+
if err != nil {
34
+
json.NewEncoder(w).Encode(map[string]string{"error": "Invalid URL"})
35
+
return
36
+
}
37
+
// Use a standard browser UA to avoid 403s (e.g. Wikipedia)
38
+
req.Header.Set("User-Agent", "Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/120.0.0.0 Safari/537.36")
39
+
40
+
client := &http.Client{}
41
+
resp, err := client.Do(req)
33
42
if err != nil {
34
43
json.NewEncoder(w).Encode(map[string]string{"error": "Failed to fetch URL"})
35
44
return
···
44
53
}
45
54
46
55
metadata := make(map[string]string)
56
+
57
+
// Helper to extract text from a node's children
58
+
var extractText func(*html.Node) string
59
+
extractText = func(n *html.Node) string {
60
+
var sb strings.Builder
61
+
if n.Type == html.TextNode {
62
+
sb.WriteString(n.Data)
63
+
}
64
+
for c := n.FirstChild; c != nil; c = c.NextSibling {
65
+
sb.WriteString(extractText(c))
66
+
}
67
+
return sb.String()
68
+
}
69
+
47
70
var f func(*html.Node)
48
71
f = func(n *html.Node) {
49
72
if n.Type == html.ElementNode && n.Data == "meta" {
···
83
106
if n.FirstChild != nil {
84
107
if _, ok := metadata["title"]; !ok {
85
108
metadata["title"] = n.FirstChild.Data
109
+
}
110
+
}
111
+
}
112
+
113
+
// Look for first paragraph if no description yet
114
+
if n.Type == html.ElementNode && n.Data == "p" {
115
+
if _, hasDesc := metadata["description"]; !hasDesc {
116
+
text := strings.TrimSpace(extractText(n))
117
+
// Wikipedia paragraphs often have citations [1] or are empty/short
118
+
// Simple heuristic: length > 50
119
+
if len(text) > 50 {
120
+
// Check for "Coordinates:" which matches length but isn't intro
121
+
if !strings.HasPrefix(text, "Coordinates:") {
122
+
metadata["description"] = text
123
+
}
86
124
}
87
125
}
88
126
}