inital commit

parse_helpers.go

@@ -0,0 +1,117 @@
+package fa
+
+import (
+	"strconv"
+	"strings"
+	"time"
+
+	"github.com/PuerkitoBio/goquery"
+)
+
+// parsePopupDate reads a FurAffinity popup_date element. Every popup_date on
+// the site carries an authoritative `data-time` attribute holding the post
+// time as Unix seconds; that is preferred. The visible title attribute is a
+// fallback for older rendered pages that omit data-time.
+//
+// Returns the zero time if neither source is parseable. Callers check IsZero
+// rather than receiving an error, because dates are nice-to-have and
+// shouldn't fail a whole parse on their own.
+func parsePopupDate(s *goquery.Selection) time.Time {
+	if s == nil || s.Length() == 0 {
+		return time.Time{}
+	}
+	if v := trimAttr(s, "data-time"); v != "" {
+		if secs, err := strconv.ParseInt(v, 10, 64); err == nil && secs > 0 {
+			return time.Unix(secs, 0).UTC()
+		}
+	}
+	raw := firstNonEmpty(trimAttr(s, "title"), trimText(s))
+	if t, err := ParseFADate(raw); err == nil {
+		return t
+	}
+	return time.Time{}
+}
+
+// extractIntFromHref pulls the first /<numeric>/ segment from an href.
+// Returns 0 if none is found. Used to lift submission/journal/comment IDs
+// out of links like "/view/12345678/" or "/journal/4567890/".
+func extractIntFromHref(href string) int64 {
+	for _, seg := range strings.Split(href, "/") {
+		if seg == "" {
+			continue
+		}
+		if n, err := strconv.ParseInt(seg, 10, 64); err == nil {
+			return n
+		}
+	}
+	return 0
+}
+
+// trimText is the goquery-friendly equivalent of strings.TrimSpace applied
+// to a selection's text. Returns "" when the selection is empty.
+func trimText(s *goquery.Selection) string {
+	if s == nil || s.Length() == 0 {
+		return ""
+	}
+	return strings.TrimSpace(s.Text())
+}
+
+// trimAttr returns the trimmed value of attr, or "" if missing.
+func trimAttr(s *goquery.Selection, attr string) string {
+	if s == nil || s.Length() == 0 {
+		return ""
+	}
+	v, ok := s.Attr(attr)
+	if !ok {
+		return ""
+	}
+	return strings.TrimSpace(v)
+}
+
+// firstNonEmpty returns the first non-empty (after trim) string.
+func firstNonEmpty(vals ...string) string {
+	for _, v := range vals {
+		if t := strings.TrimSpace(v); t != "" {
+			return t
+		}
+	}
+	return ""
+}
+
+// htmlOf returns the inner HTML of a selection, with leading/trailing
+// whitespace trimmed. Returns "" when the selection is empty or rendering fails.
+func htmlOf(s *goquery.Selection) string {
+	if s == nil || s.Length() == 0 {
+		return ""
+	}
+	h, err := s.Html()
+	if err != nil {
+		return ""
+	}
+	return strings.TrimSpace(h)
+}
+
+// parseStatNumber parses a number that may carry commas or surrounding text
+// (e.g. "1,234 views"). Returns 0 on any failure rather than propagating an
+// error stats are nice-to-have, not load-bearing.
+func parseStatNumber(s string) int {
+	s = strings.TrimSpace(s)
+	if s == "" {
+		return 0
+	}
+	// Strip everything that isn't a digit or comma; keep digits & commas only.
+	var b strings.Builder
+	for _, r := range s {
+		if r >= '0' && r <= '9' {
+			b.WriteRune(r)
+		}
+	}
+	if b.Len() == 0 {
+		return 0
+	}
+	n, err := strconv.Atoi(b.String())
+	if err != nil {
+		return 0
+	}
+	return n
+}