fix: hybrid SVG+PIL PNG image rendering with CJK fonts, prompt URL reference

scripts/image_generator.py

@@ -3,9 +3,67 @@
 多平台文章配图生成器
 为知乎/公众号/小红书生成平台风格的 SVG 配图（base64 内联，无需外部资源）
 """
-import base64, math, textwrap, re
+import base64, math, textwrap, re, io, os
 from typing import List, Optional
 
+CJK_FONT = "/usr/share/fonts/google-noto-cjk/NotoSansCJKsc-Regular.otf"
+CJK_FONT_BOLD = "/usr/share/fonts/google-noto-cjk/NotoSansCJKsc-Bold.otf"
+
+
+def _svg_to_png_with_pil(svg: str) -> bytes:
+    """用 cairosvg 渲染背景 + Pillow 叠加中文文字"""
+    from PIL import Image, ImageDraw, ImageFont
+    import cairosvg
+
+    text_elements = []
+    def _extract_text(m):
+        attrs = dict(re.findall(r'([\w-]+)="([^"]*)"', m.group(1)))
+        text_elements.append({
+            'x': int(float(attrs.get('x', 0))),
+            'y': int(float(attrs.get('y', 0))),
+            'size': int(float(attrs.get('font-size', 14))),
+            'weight': attrs.get('font-weight', 'normal'),
+            'fill': attrs.get('fill', '#000'),
+            'opacity': float(attrs.get('opacity', 1) or 1),
+            'text': m.group(2),
+        })
+        return ''
+
+    svg_no_text = re.sub(r'<text\b([^>]*)>([^<]*)</text>', _extract_text, svg)
+
+    png = cairosvg.svg2png(bytestring=svg_no_text.encode('utf-8'))
+    img = Image.open(io.BytesIO(png)).convert('RGBA')
+    draw = ImageDraw.Draw(img, 'RGBA')
+
+    for el in text_elements:
+        if not el['text'].strip():
+            continue
+        font_path = CJK_FONT_BOLD if el['weight'] in ('bold', '700', '800', '900') else CJK_FONT
+        try:
+            font = ImageFont.truetype(font_path, el['size'])
+        except Exception:
+            font = ImageFont.load_default()
+        fill_color = el['fill'].lstrip('#')
+        try:
+            r, g, b = tuple(int(fill_color[i:i+2], 16) for i in (0, 2, 4))
+        except Exception:
+            r, g, b = 0, 0, 0
+        a = int(255 * el['opacity'])
+        _, _, tw, th = draw.textbbox((0, 0), el['text'], font=font)
+        draw.text((el['x'], el['y'] - int(el['size'] * 0.85)), el['text'], font=font, fill=(r, g, b, a))
+
+    buf = io.BytesIO()
+    img.save(buf, 'PNG')
+    return buf.getvalue()
+
+
+def _to_base64(svg: str) -> str:
+    try:
+        png = _svg_to_png_with_pil(svg)
+        return 'data:image/png;base64,' + base64.b64encode(png).decode('ascii')
+    except Exception:
+        return 'data:image/svg+xml;base64,' + base64.b64encode(svg.encode('utf-8')).decode('ascii')
+
 PLATFORM_STYLES = {
     "zhihu": {
         "primary": "#0084ff",
@@ -39,7 +97,7 @@ PLATFORM_STYLES = {
     },
 }
 
-FONT = "-apple-system, BlinkMacSystemFont, 'PingFang SC', 'Microsoft YaHei', 'Helvetica Neue', sans-serif"
+FONT = "-apple-system, BlinkMacSystemFont, 'Noto Sans CJK SC', 'PingFang SC', 'Microsoft YaHei', 'Helvetica Neue', sans-serif"
 
 
 def _wrap_chinese(text: str, chars_per_line: int = 14) -> List[str]:
@@ -61,15 +119,6 @@ def _wrap_chinese(text: str, chars_per_line: int = 14) -> List[str]:
     return lines
 
 
-def _to_base64(svg: str) -> str:
-    try:
-        import cairosvg
-        png = cairosvg.svg2png(bytestring=svg.encode('utf-8'))
-        return 'data:image/png;base64,' + base64.b64encode(png).decode('ascii')
-    except Exception:
-        return 'data:image/svg+xml;base64,' + base64.b64encode(svg.encode('utf-8')).decode('ascii')
-
-
 def _alt_attr(text: str) -> str:
     return text.replace('&', '&amp;').replace('<', '&lt;').replace('>', '&gt;').replace('"', '&quot;').replace("'", '&#39;')
 
@@ -166,11 +215,11 @@ def generate_data_highlight(platform: str, number: str, label: str) -> str:
 
 
 def insert_lead_image(html: str, platform: str, title: str, field: str) -> str:
-    """在 HTML 正文开头插入头图（仅必要环节）"""
+    """在 HTML 正文开头插入头图（位于 h1 之前）"""
     lead = generate_lead(platform, title, field)
-    h1_end = html.find('</h1>')
-    if h1_end != -1:
-        html = html[:h1_end + 5] + '\n' + lead + html[h1_end + 5:]
+    h1_start = html.find('<h1>')
+    if h1_start != -1:
+        html = html[:h1_start] + lead + '\n' + html[h1_start:]
     else:
         html = lead + html
     return html