init: inspiration collector v1.0

analyzers/daily_digest.py

@@ -0,0 +1,168 @@
+"""Daily digest analyzer - fetches today's memos and generates AI summary."""
+
+import json
+import logging
+import os
+import sys
+from datetime import datetime, timezone
+
+# Add parent dir to path for local imports
+sys.path.insert(0, os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
+
+from tools.config import load_secrets, get_output_dir
+from tools.llm import DeepSeekClient
+from tools.memos_client import MemosClient
+from tools.formatter import format_daily_digest
+
+logger = logging.getLogger(__name__)
+
+# System prompt for daily digest
+DAILY_SYSTEM_PROMPT = """你是一个灵感整理助手。你的任务是将用户零散的灵感记录进行智能整理。
+
+请严格按照要求整理，输出纯 JSON（不要代码块标记，不要额外说明）：
+
+{
+  "summary": "一段简短的今日灵感总结，50字以内",
+  "categories": {
+    "分类名1": ["具体灵感项1", "具体灵感项2"],
+    "分类名2": ["具体灵感项3"]
+  },
+  "connections": ["跨主题关联发现1", "2-3条"],
+  "todos": ["待办事项1", "2-3条"]
+}
+
+分类原则：根据内容自然归类，分类名不超过4个字。
+关联发现：识别不同灵感之间的关联、冲突或可串联的主题。
+待办：从灵感中提取可执行的事项。"""
+
+
+def parse_ai_response(text):
+    """Parse AI response, handling both pure JSON and code-block wrapped JSON."""
+    text = text.strip()
+    # Remove markdown code block wrappers if present
+    if text.startswith("```"):
+        lines = text.split("\n")
+        # Remove first and last ``` lines
+        if lines[0].startswith("```"):
+            lines = lines[1:]
+        if lines and lines[-1].strip() == "```":
+            lines = lines[:-1]
+        text = "\n".join(lines).strip()
+
+    return json.loads(text)
+
+
+def run(memos_client, llm_client, date=None):
+    """Run daily digest analysis."""
+    date = date or datetime.now(timezone.utc)
+
+    logger.info("Daily digest started for %s", date.strftime("%Y-%m-%d"))
+
+    # Step 1: Fetch today's memos
+    memos = memos_client.list_memos(days=1)
+
+    if not memos:
+        logger.info("No memos today, skipping")
+        # Still generate a minimal file
+        content = format_daily_digest(
+            date=date,
+            categories={},
+            summary="今天没有记录灵感。",
+            connections=[],
+            todos=[]
+        )
+        output_dir = get_output_dir("daily")
+        filename = f"{date.strftime('%Y-%m-%d')}_digest.md"
+        filepath = os.path.join(output_dir, filename)
+
+        with open(filepath, "w", encoding="utf-8") as f:
+            f.write(content)
+
+        logger.info("Empty digest written to %s", filepath)
+        return filepath, 0
+
+    # Step 2: Prepare prompt
+    memo_texts = []
+    for m in memos:
+        time_str = m["created_at"][:16].replace("T", " ")
+        memo_texts.append(f"[{time_str}] {m['content']}")
+
+    user_prompt = f"请分析以下灵感记录：\n\n" + "\n".join(memo_texts)
+
+    # Step 3: Call DeepSeek API
+    raw_response = llm_client.ask(
+        system_prompt=DAILY_SYSTEM_PROMPT,
+        user_prompt=user_prompt,
+        temperature=0.3
+    )
+
+    # Step 4: Parse and format
+    try:
+        data = parse_ai_response(raw_response)
+    except (json.JSONDecodeError, KeyError) as e:
+        logger.error("Failed to parse AI response: %s", e)
+        logger.error("Raw response: %s", raw_response[:200])
+        data = {
+            "summary": f"AI 解析失败，请查看原始 Memos。",
+            "categories": {"未分类": [m["content"] for m in memos]},
+            "connections": [],
+            "todos": []
+        }
+
+    categories = data.get("categories", {})
+    connections = data.get("connections", [])
+    todos = data.get("todos", [])
+    summary = data.get("summary", "")
+
+    # Step 5: Format and write output
+    content = format_daily_digest(
+        date=date,
+        categories=categories,
+        summary=summary,
+        connections=connections,
+        todos=todos
+    )
+
+    output_dir = get_output_dir("daily")
+    filename = f"{date.strftime('%Y-%m-%d')}_digest.md"
+    filepath = os.path.join(output_dir, filename)
+
+    with open(filepath, "w", encoding="utf-8") as f:
+        f.write(content)
+
+    logger.info(
+        "Daily digest written to %s | %d memos | %d categories",
+        filepath, len(memos), len(categories)
+    )
+    return filepath, len(memos)
+
+
+def main():
+    """CLI entry point."""
+    logging.basicConfig(
+        level=logging.INFO,
+        format="%(asctime)s [%(levelname)s] %(name)s: %(message)s"
+    )
+
+    try:
+        secrets = load_secrets()
+    except FileNotFoundError as e:
+        print(e)
+        sys.exit(1)
+
+    memos_client = MemosClient(
+        base_url=secrets.get("memos_url", "http://localhost:5230"),
+        access_token=secrets["memos_token"]
+    )
+
+    llm_client = DeepSeekClient(
+        api_key=secrets["deepseek_api_key"],
+        model=secrets.get("deepseek_model", "deepseek-chat")
+    )
+
+    filepath, count = run(memos_client, llm_client)
+    print(f"Done: {filepath} ({count} memos)")
+
+
+if __name__ == "__main__":
+    main()