fix: content quality, image format, task monitor, calendar data source, search UI & sort

2026-05-26 11:26:10 +08:00
parent b2d043b231
commit ac8644d752
36 changed files with 2294 additions and 873 deletions
@@ -35,6 +35,28 @@ PLATFORM_RULES = {
    }
 }

+# AI 套话检测模式（一旦出现在正文中，说明写作痕迹明显）
+AI_TELTALES = [
+    "说回到",
+    "一个真实的.*案例很能说明问题",
+    "这就是.*被.*后的样子",
+    "如果你也",
+    "值得注意的是",
+    "首先其次最后",
+    "综上所述",
+    "总的来说",
+    "说到这里",
+    "我们来总结一下",
+    "总而言之",
+    "我们不难发现",
+    "我们可以看出",
+    "从以上分析可以看出",
+    "无可否认",
+    "众所周知",
+    "毋庸置疑",
+    "不知大家有没有发现",
+]
+
 _cached_sensitive_words = None
 _cached_platform_rules = None

@@ -139,15 +161,22 @@ class ComplianceChecker:
        # 6. 最小字数检查
        self._check_min_length(text, platform)

-        # 7. 结构完整性检查（必须包含关键章节）
-        self._check_required_sections(text)
+        # 7. 内容质量检查
+        self._check_ai_telltales(text)
+        self._check_pronoun_consistency(text, platform)
+        self._check_reading_experience(text, platform)
+        self._check_platform_engagement(text, platform)
        self._check_inline_images(text)
        self._check_timeliness(text)

+        hard_types = ('敏感词', '法律法规', '平台规则', '品牌规范', '资源合规')
+        hard_issues = [i for i in self.issues if i['type'] in hard_types]
        return {
-            "passed": len(self.issues) == 0,
+            "passed": len(hard_issues) == 0,
            "issues": self.issues,
-            "score": max(0, 100 - len(self.issues) * 10)
+            "score": max(0, 100 - sum(
+                10 if i['type'] in hard_types else 5 for i in self.issues
+            ))
        }

    def _check_sensitive_words(self, text: str):
@@ -317,23 +346,120 @@ class ComplianceChecker:
                "suggestion": "扩写内容至最低要求"
            })

-    def _check_required_sections(self, text: str):
-        """检查是否包含必要章节（如引言、核心观点、总结等）"""
-        required_headings = [
-            "引言", "核心观点", "受众痛点", "总结", "行动指南"
-        ]
-        missing = []
-        for heading in required_headings:
-            # 检查 h2 或 h3 中是否出现 heading
-            if not re.search(r'<h[23][^>]*>.*' + re.escape(heading) + r'.*</h[23]>', text, re.IGNORECASE):
-                missing.append(heading)
-        if missing:
+    def _check_ai_telltales(self, text: str):
+        """检查AI套话——正文中出现这些模式说明AI写作痕迹明显"""
+        plain = re.sub(r'<[^>]+>', '', text)
+        for pattern in AI_TELTALES:
+            if re.search(pattern, plain):
+                self.issues.append({
+                    "type": "内容质量",
+                    "category": "AI套话",
+                    "detail": f"正文出现AI套话模式: 「{pattern}」",
+                    "suggestion": "删除或替换为自然表达，不要让读者感觉是AI写的"
+                })
+
+    def _check_pronoun_consistency(self, text: str, platform: str):
+        """检查人称一致性（尤其是微信文章）"""
+        if platform != "wechat":
+            return
+        plain = re.sub(r'<[^>]+>', '', text)
+        has_ni = '你' in plain
+        has_nimen = '你们' in plain
+        has_women = '我们' in plain
+        if has_nimen and has_ni:
            self.issues.append({
-                "type": "结构完整",
-                "category": "章节缺失",
-                "detail": f"缺少必要章节：{', '.join(missing)}",
-                "suggestion": "补充缺失章节"
+                "type": "内容质量",
+                "category": "人称混用",
+                "detail": "微信文章中同时使用「你」和「你们」，建议统一为「你」",
+                "suggestion": "将所有「你们」替换为「你」，保持与读者的单数对话感"
            })
+        if has_women and has_ni:
+            self.issues.append({
+                "type": "内容质量",
+                "category": "人称混用",
+                "detail": "微信文章中同时使用「我们」和「你」，建议统一视角",
+                "suggestion": "将「我们」替换为「你」或「我」，保持与读者对话而非说教"
+            })
+
+    def _check_reading_experience(self, text: str, platform: str = ""):
+        """检查阅读体验（段落长度、配图）"""
+        paragraphs = re.findall(r'<p>(.*?)</p>', text, re.DOTALL)
+        long_paras = [p for p in paragraphs if len(p) > 300]
+        if len(long_paras) > len(paragraphs) * 0.3:
+            self.issues.append({
+                "type": "内容质量",
+                "category": "段落过长",
+                "detail": f"超过30%的段落长度>300字（共{len(paragraphs)}段，{len(long_paras)}段过长），在手机上阅读体验差",
+                "suggestion": "将长段拆分为2-3个短段，每段不超过150-200字"
+            })
+        # 图片检测（各平台阈值不同）
+        imgs = re.findall(r'<img', text)
+        min_imgs = {"zhihu": 1, "wechat": 1, "xiaohongshu": 1}.get(platform, 1)
+        if not imgs:
+            self.issues.append({
+                "type": "内容质量",
+                "category": "缺少配图",
+                "detail": "正文中没有图片，建议插入配图提升阅读体验",
+                "suggestion": "在关键位置插入配图（数据截图、金句卡片、操作步骤图等）"
+            })
+        elif len(imgs) < min_imgs:
+            self.issues.append({
+                "type": "内容质量",
+                "category": "配图不足",
+                "detail": f"只有{len(imgs)}张图，建议至少{min_imgs}张配图提升阅读体验",
+                "suggestion": "在文章中间部分增加配图"
+            })
+
+    def _check_platform_engagement(self, text: str, platform: str):
+        """检查平台特有的互动/传播要素"""
+        plain = re.sub(r'<[^>]+>', '', text)
+
+        if platform == "zhihu":
+            # 知乎需要讨论引导（评论区互动是核心算法权重）
+            if not any(kw in plain for kw in ('你觉得', '你怎么看', '欢迎在评论区', '说说你的', '欢迎讨论', '你怎么想')):
+                self.issues.append({
+                    "type": "内容质量",
+                    "category": "缺少互动引导",
+                    "detail": "知乎文章建议在末尾加讨论引导（如「你觉得呢？欢迎在评论区聊聊」），提升互动率",
+                    "suggestion": "在文章末尾添加一个问题或讨论话题，引导读者评论"
+                })
+            # 检查是否有数据引用（知乎读者看重可信度）
+            if not any(kw in plain for kw in ('据统计', '调研显示', '数据显示', '报告指出', '根据', '研究表明', '调查显示')):
+                self.issues.append({
+                    "type": "内容质量",
+                    "category": "缺少数据引用",
+                    "detail": "知乎文章建议引用具体数据或报告来支撑观点，增强可信度",
+                    "suggestion": "在关键观点处引用权威数据来源"
+                })
+
+        if platform == "xiaohongshu":
+            # 小红书需要收藏引导（收藏率是推荐算法核心指标）
+            if '收藏' not in plain:
+                self.issues.append({
+                    "type": "内容质量",
+                    "category": "缺少收藏引导",
+                    "detail": "小红书笔记建议在末尾加收藏引导（如「觉得有用点个收藏」），提升收藏率",
+                    "suggestion": "在末尾添加收藏引导语"
+                })
+            # 小红书段落需非常短
+            paragraphs = re.findall(r'<p>(.*?)</p>', text, re.DOTALL)
+            long_paras = [p for p in paragraphs if len(p) > 150]
+            if len(long_paras) > len(paragraphs) * 0.2:
+                self.issues.append({
+                    "type": "内容质量",
+                    "category": "段落过长",
+                    "detail": f"小红书建议每段不超过80-100字，当前{len(long_paras)}/{len(paragraphs)}段超过150字",
+                    "suggestion": "将长段拆分为1-2句的短段落，每段不超过100字"
+                })
+            # 小红书需要至少一些 emoji
+            if not re.search(r'[\U0001F300-\U0001F9FF\u2600-\u27BF]', plain):
+                self.issues.append({
+                    "type": "内容质量",
+                    "category": "缺少emoji",
+                    "detail": "小红书笔记建议适当使用emoji来增加视觉吸引力",
+                    "suggestion": "在标题、章节分隔或重点句前添加相关emoji"
+                })
+
    def _check_inline_images(self, html: str):
        """检查图片是否以内联方式嵌入（data:image）"""
        # 提取所有 img 标签的 src 属性值