Initial commit: Hermes Agent skills collection

- Trading skills (OKX, dividend, lottery, quantitative) - Creative skills (ASCII art, diagrams, video) - Development skills (GitHub, debugging, TDD) - Research skills (arXiv, blog monitoring) - Productivity skills (email, documents, notes) - MCP integration skills - Custom user skills
2026-07-05 02:31:15 -04:00
commit 6770bc9b9d
908 changed files with 239614 additions and 0 deletions
@@ -0,0 +1,365 @@
+#!/usr/bin/env python3
+"""
+用户痛点自动抓取与分析工具
+从 Reddit、Hacker News 等平台采集用户评论，用 LLM 提取痛点。
+
+用法:
+  python3 pain_point_scraper.py --topic "AI写作工具" --sources reddit,hn --limit 20
+  python3 pain_point_scraper.py --topic "OKX交易" --sources reddit --limit 30 --output report.md
+"""
+
+import argparse
+import json
+import sys
+import time
+from datetime import datetime
+from typing import Optional
+
+import requests
+
+# ─────────────────────────────────────────────
+# 配置
+# ─────────────────────────────────────────────
+
+# LLM API (使用本地 Ollama 或 LM Studio)
+LLM_API_URL = "http://192.168.199.100:11434/v1/chat/completions"
+LLM_MODEL = "gemma-4-26b-a4b-it-4bit"
+LLM_API_KEY = ""  # 本地服务通常不需要
+
+# 如果本地不可用，fallback到远程
+LLM_FALLBACK_URL = "https://token-plan-cn.xiaomimimo.com/v1/chat/completions"
+LLM_FALLBACK_MODEL = "mimo-v2.5-pro"
+LLM_FALLBACK_KEY = "tp-cd64i50fo6ifihhiupspq9dypmzeud3u77tia5mlz0rb572u"
+
+HEADERS = {
+    "User-Agent": "Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/537.36"
+}
+
+
+# ─────────────────────────────────────────────
+# Reddit 采集
+# ─────────────────────────────────────────────
+
+def scrape_reddit(topic: str, limit: int = 20) -> list[dict]:
+    """从 Reddit 搜索相关讨论"""
+    results = []
+    
+    # 使用 Reddit JSON API (无需认证)
+    url = "https://www.reddit.com/search.json"
+    params = {
+        "q": topic,
+        "sort": "relevance",
+        "t": "month",  # 最近一个月
+        "limit": min(limit, 25),
+        "type": "link"
+    }
+    
+    try:
+        resp = requests.get(url, params=params, headers=HEADERS, timeout=15)
+        resp.raise_for_status()
+        data = resp.json()
+        
+        for post in data.get("data", {}).get("children", []):
+            d = post.get("data", {})
+            results.append({
+                "source": "Reddit",
+                "title": d.get("title", ""),
+                "text": d.get("selftext", "")[:500],
+                "score": d.get("score", 0),
+                "comments": d.get("num_comments", 0),
+                "url": f"https://reddit.com{d.get('permalink', '')}",
+                "subreddit": d.get("subreddit", ""),
+                "created": datetime.fromtimestamp(d.get("created_utc", 0)).isoformat()
+            })
+        
+        print(f"  ✅ Reddit: 获取 {len(results)} 条", file=sys.stderr)
+    except Exception as e:
+        print(f"  ❌ Reddit 采集失败: {e}", file=sys.stderr)
+    
+    return results
+
+
+# ─────────────────────────────────────────────
+# Hacker News 采集
+# ─────────────────────────────────────────────
+
+def scrape_hn(topic: str, limit: int = 20) -> list[dict]:
+    """从 Hacker News 搜索相关讨论"""
+    results = []
+    
+    url = "https://hn.algolia.com/api/v1/search"
+    params = {
+        "query": topic,
+        "tags": "story",
+        "hitsPerPage": min(limit, 30),
+        "numericFilters": "created_at_i>" + str(int(time.time()) - 30*86400)
+    }
+    
+    try:
+        resp = requests.get(url, params=params, timeout=15)
+        resp.raise_for_status()
+        data = resp.json()
+        
+        for hit in data.get("hits", []):
+            results.append({
+                "source": "Hacker News",
+                "title": hit.get("title", ""),
+                "text": (hit.get("story_text") or "")[:500],
+                "score": hit.get("points", 0),
+                "comments": hit.get("num_comments", 0),
+                "url": hit.get("url", f"https://news.ycombinator.com/item?id={hit.get('objectID', '')}"),
+                "author": hit.get("author", ""),
+                "created": hit.get("created_at", "")
+            })
+        
+        print(f"  ✅ HN: 获取 {len(results)} 条", file=sys.stderr)
+    except Exception as e:
+        print(f"  ❌ HN 采集失败: {e}", file=sys.stderr)
+    
+    return results
+
+
+# ─────────────────────────────────────────────
+# Product Hunt 采集 (通过搜索)
+# ─────────────────────────────────────────────
+
+def scrape_producthunt(topic: str, limit: int = 10) -> list[dict]:
+    """从 Product Hunt 搜索相关产品评论"""
+    results = []
+    
+    # 使用 Google 搜索 Product Hunt 上的内容
+    url = "https://www.google.com/search"
+    params = {
+        "q": f"site:producthunt.com {topic} review complaints",
+        "num": min(limit, 10)
+    }
+    
+    try:
+        resp = requests.get(url, params=params, headers=HEADERS, timeout=15)
+        # Google 可能会block，所以这里简单处理
+        if resp.status_code == 200:
+            # 简单提取标题
+            import re
+            titles = re.findall(r'<h3[^>]*>(.*?)</h3>', resp.text)
+            for t in titles[:limit]:
+                clean = re.sub(r'<[^>]+>', '', t)
+                if clean and len(clean) > 10:
+                    results.append({
+                        "source": "Product Hunt",
+                        "title": clean,
+                        "text": "",
+                        "score": 0,
+                        "comments": 0,
+                        "url": "",
+                        "created": ""
+                    })
+        
+        print(f"  ✅ Product Hunt: 获取 {len(results)} 条", file=sys.stderr)
+    except Exception as e:
+        print(f"  ❌ Product Hunt 采集失败: {e}", file=sys.stderr)
+    
+    return results
+
+
+# ─────────────────────────────────────────────
+# LLM 分析
+# ─────────────────────────────────────────────
+
+def call_llm(prompt: str, api_url: str = None, model: str = None, api_key: str = None) -> str:
+    """调用 LLM API 进行分析"""
+    api_url = api_url or LLM_API_URL
+    model = model or LLM_MODEL
+    api_key = api_key or LLM_API_KEY
+    
+    headers = {"Content-Type": "application/json"}
+    if api_key:
+        headers["Authorization"] = f"Bearer {api_key}"
+    
+    payload = {
+        "model": model,
+        "messages": [
+            {"role": "system", "content": "你是一个专业的用户研究分析师。你擅长从用户评论中提取痛点、需求和机会。请用中文回答。"},
+            {"role": "user", "content": prompt}
+        ],
+        "temperature": 0.3,
+        "max_tokens": 4000
+    }
+    
+    try:
+        resp = requests.post(api_url, json=payload, headers=headers, timeout=120)
+        resp.raise_for_status()
+        data = resp.json()
+        return data["choices"][0]["message"]["content"]
+    except Exception as e:
+        print(f"  ⚠️ 主API失败 ({e})，尝试fallback...", file=sys.stderr)
+        # Fallback
+        try:
+            fallback_headers = {
+                "Content-Type": "application/json",
+                "Authorization": f"Bearer {LLM_FALLBACK_KEY}"
+            }
+            payload["model"] = LLM_FALLBACK_MODEL
+            resp = requests.post(LLM_FALLBACK_URL, json=payload, headers=fallback_headers, timeout=120)
+            resp.raise_for_status()
+            data = resp.json()
+            return data["choices"][0]["message"]["content"]
+        except Exception as e2:
+            return f"❌ LLM 分析失败: {e2}"
+
+
+def analyze_pain_points(items: list[dict], topic: str) -> str:
+    """用 LLM 分析采集到的数据，提取痛点"""
+    
+    # 构建分析prompt
+    reviews_text = ""
+    for i, item in enumerate(items[:30], 1):  # 最多30条，避免token超限
+        reviews_text += f"\n--- [{i}] {item['source']} | 👍{item['score']} 💬{item['comments']} ---\n"
+        reviews_text += f"标题: {item['title']}\n"
+        if item.get('text'):
+            reviews_text += f"内容: {item['text'][:300]}\n"
+    
+    prompt = f"""请分析以下关于"{topic}"的用户评论/讨论，提取用户痛点、需求和机会。
+
+## 原始数据
+{reviews_text}
+
+## 分析要求
+
+请按以下格式输出：
+
+### 📊 概览
+- 总评论数：N
+- 主要痛点数：N
+- 整体情绪：正面/负面/混合
+
+### 🔥 TOP 痛点（按严重程度排序）
+
+对每个痛点：
+1. **痛点标题**（一句话概括）
+2. **用户原话**（引用1-2条最能说明问题的原文）
+3. **痛点描述**（为什么这是问题）
+4. **影响人群**（谁会遇到这个问题）
+5. **现有解决方案**（有没有竞品在解决）
+6. **机会评估**（⭐1-5星，这个痛点值不值得做）
+
+### 💡 机会洞察
+- 基于以上痛点，有哪些产品/功能机会
+- 有哪些未被满足的需求
+
+### 📈 趋势观察
+- 近期讨论的热点变化
+- 用户关注点的转移
+
+请用中文回答，保持分析的客观性和实用性。"""
+
+    return call_llm(prompt)
+
+
+# ─────────────────────────────────────────────
+# 报告生成
+# ─────────────────────────────────────────────
+
+def generate_report(topic: str, items: list[dict], analysis: str) -> str:
+    """生成完整的Markdown报告"""
+    
+    sources = {}
+    for item in items:
+        src = item["source"]
+        sources[src] = sources.get(src, 0) + 1
+    
+    source_summary = "、".join([f"{k}({v}条)" for k, v in sources.items()])
+    
+    report = f"""# 🔍 痛点分析报告：{topic}
+> 数据来源：{source_summary} | 采集时间：{datetime.now().strftime('%Y-%m-%d %H:%M')} | 分析评论数：{len(items)}
+
+---
+
+{analysis}
+
+---
+
+## 📋 原始数据摘要
+
+| # | 来源 | 标题 | 👍 | 💬 |
+|---|------|------|-----|-----|
+"""
+    
+    for i, item in enumerate(items[:20], 1):
+        title = item["title"][:50] + ("..." if len(item["title"]) > 50 else "")
+        report += f"| {i} | {item['source']} | {title} | {item['score']} | {item['comments']} |\n"
+    
+    report += f"\n*共 {len(items)} 条数据，显示前 20 条*\n"
+    
+    return report
+
+
+# ─────────────────────────────────────────────
+# 主流程
+# ─────────────────────────────────────────────
+
+def main():
+    parser = argparse.ArgumentParser(description="用户痛点自动抓取与分析")
+    parser.add_argument("--topic", "-t", required=True, help="分析主题/关键词")
+    parser.add_argument("--sources", "-s", default="reddit,hn", help="数据源 (逗号分隔: reddit,hn,producthunt)")
+    parser.add_argument("--limit", "-l", type=int, default=20, help="每个源采集数量")
+    parser.add_argument("--output", "-o", help="输出文件路径 (默认stdout)")
+    parser.add_argument("--raw", action="store_true", help="只输出原始数据，不做LLM分析")
+    
+    args = parser.parse_args()
+    
+    print(f"\n🔍 痛点分析：{args.topic}", file=sys.stderr)
+    print(f"📡 数据源：{args.sources}", file=sys.stderr)
+    print(f"📊 采集数量：每个源 {args.limit} 条\n", file=sys.stderr)
+    
+    # 1. 采集数据
+    all_items = []
+    source_map = {
+        "reddit": scrape_reddit,
+        "hn": scrape_hn,
+        "producthunt": scrape_producthunt
+    }
+    
+    for source in args.sources.split(","):
+        source = source.strip().lower()
+        if source in source_map:
+            items = source_map[source](args.topic, args.limit)
+            all_items.extend(items)
+        else:
+            print(f"  ⚠️ 未知数据源: {source}", file=sys.stderr)
+    
+    if not all_items:
+        print("\n❌ 没有采集到任何数据", file=sys.stderr)
+        sys.exit(1)
+    
+    print(f"\n📊 总计采集: {len(all_items)} 条\n", file=sys.stderr)
+    
+    # 2. 如果只输出原始数据
+    if args.raw:
+        output = json.dumps(all_items, ensure_ascii=False, indent=2)
+        if args.output:
+            with open(args.output, "w") as f:
+                f.write(output)
+            print(f"✅ 原始数据已保存到: {args.output}", file=sys.stderr)
+        else:
+            print(output)
+        return
+    
+    # 3. LLM 分析
+    print("🧠 正在分析痛点...", file=sys.stderr)
+    analysis = analyze_pain_points(all_items, args.topic)
+    
+    # 4. 生成报告
+    report = generate_report(args.topic, all_items, analysis)
+    
+    if args.output:
+        with open(args.output, "w") as f:
+            f.write(report)
+        print(f"\n✅ 报告已保存到: {args.output}", file=sys.stderr)
+    else:
+        print(report)
+    
+    print(f"\n✅ 分析完成！", file=sys.stderr)
+
+
+if __name__ == "__main__":
+    main()