init: inspiration collector v1.0

2026-06-12 23:30:27 +08:00
commit 5a876e5ecd
17 changed files with 881 additions and 0 deletions
--- a/tools/init.py
+++ b/tools/init.py
--- a/tools/config.py
+++ b/tools/config.py
@ -0,0 +1,35 @@
+"""Configuration loader - loads secrets from secrets.json"""
+
+import json
+import os
+from pathlib import Path
+
+# Default location: ~/inspiration-collector/secrets/secrets.json on server
+SECRETS_PATH = os.environ.get(
+    "IC_SECRETS_PATH",
+    str(Path.home() / "inspiration-collector" / "secrets" / "secrets.json")
+)
+
+DEFAULT_OUTPUT_DIR = os.environ.get(
+    "IC_OUTPUT_DIR",
+    str(Path.home() / "inspiration-collector" / "ai-insights")
+)
+
+
+def load_secrets(path=None):
+    """Load secrets from JSON file. Returns dict with api_key, memos_token, memos_url."""
+    path = path or SECRETS_PATH
+    if not os.path.exists(path):
+        raise FileNotFoundError(
+            f"Secrets file not found: {path}\n"
+            f"Copy secrets/secrets.json.template to {path} and fill in your keys."
+        )
+    with open(path) as f:
+        return json.load(f)
+
+
+def get_output_dir(subdir="daily"):
+    """Get output directory for AI-generated insights."""
+    base = DEFAULT_OUTPUT_DIR
+    os.makedirs(os.path.join(base, subdir), exist_ok=True)
+    return os.path.join(base, subdir)
--- a/tools/formatter.py
+++ b/tools/formatter.py
@ -0,0 +1,122 @@
+"""Markdown formatter - converts AI analysis results to beautiful .md files."""
+
+from datetime import datetime
+
+
+def format_daily_digest(date, categories, summary, connections, todos):
+    """Format daily digest markdown file."""
+    date_str = date.strftime("%Y-%m-%d")
+    weekday = ["周一", "周二", "周三", "周四", "周五", "周六", "周日"][date.weekday()]
+
+    lines = []
+    lines.append("---")
+    lines.append(f"date: {date_str}")
+    lines.append("type: daily-digest")
+    lines.append("tags: [灵感收集器, 每日总结]")
+    lines.append("---")
+    lines.append("")
+    lines.append(f"# {date_str} 灵感摘要 · {weekday}")
+    lines.append("")
+
+    # Summary
+    lines.append("## AI 分析（自动生成，请勿编辑）")
+    lines.append("")
+    lines.append(summary)
+    lines.append("")
+
+    # Categories
+    if categories:
+        lines.append("### 分类概览")
+        lines.append("")
+        for cat, items in categories.items():
+            count = len(items) if isinstance(items, list) else items
+            lines.append(f"- **{cat}**：{count}条")
+        lines.append("")
+
+    # Connections
+    if connections:
+        lines.append("### 关联发现")
+        lines.append("")
+        for conn in connections:
+            lines.append(f"- {conn}")
+        lines.append("")
+
+    # Todos
+    if todos:
+        lines.append("### 待办")
+        lines.append("")
+        for todo in todos:
+            lines.append(f"- [ ] {todo}")
+        lines.append("")
+
+    # User annotation area
+    lines.append("---")
+    lines.append("")
+    lines.append("## 我的批注")
+    lines.append("")
+    lines.append("> *在这里写下你的想法、质疑、补充*")
+    lines.append("")
+
+    return "\n".join(lines)
+
+
+def format_weekly_trend(start_date, end_date, stats, highlights, insight):
+    """Format weekly trend markdown file."""
+    week_start = start_date.strftime("%m/%d")
+    week_end = end_date.strftime("%m/%d")
+    week_number = start_date.isocalendar()[1]
+
+    lines = []
+    lines.append("---")
+    lines.append(f"date: {start_date.year}-W{week_number}")
+    lines.append("type: weekly-trend")
+    lines.append("tags: [灵感收集器, 每周趋势]")
+    lines.append("---")
+    lines.append("")
+    lines.append(f"# 第 {week_number} 周灵感趋势（{week_start} - {week_end}）")
+    lines.append("")
+
+    # Stats
+    lines.append("## AI 分析（自动生成，请勿编辑）")
+    lines.append("")
+    lines.append(f"本周共记录 **{stats.get('total', 0)}** 条灵感。")
+    lines.append("")
+
+    if stats.get("daily_counts"):
+        lines.append("### 每日活跃度")
+        lines.append("")
+        for day, count in stats["daily_counts"].items():
+            bar = "█" * count if count > 0 else "▏"
+            lines.append(f"- {day}：{bar} {count}条")
+        lines.append("")
+
+    if stats.get("categories"):
+        lines.append("### 主题分布")
+        lines.append("")
+        for cat, count in stats["categories"].items():
+            pct = count / stats["total"] * 100 if stats["total"] else 0
+            lines.append(f"- **{cat}**：{count}条（{pct:.0f}%）")
+        lines.append("")
+
+    if highlights:
+        lines.append("### 本周亮点")
+        lines.append("")
+        for h in highlights:
+            lines.append(f"- {h}")
+        lines.append("")
+
+    if insight:
+        lines.append("### 值得关注的模式")
+        lines.append("")
+        lines.append(insight)
+        lines.append("")
+
+    # User annotation
+    lines.append("---")
+    lines.append("")
+    lines.append("## 我的批注")
+    lines.append("")
+    lines.append("> *有什么想补充的写在下面*")
+    lines.append("")
+
+    return "\n".join(lines)
--- a/tools/llm.py
+++ b/tools/llm.py
@ -0,0 +1,71 @@
+"""DeepSeek API client - OpenAI-compatible, zero extra dependencies."""
+
+import json
+import logging
+import time
+import requests
+
+logger = logging.getLogger(__name__)
+
+
+class DeepSeekClient:
+    """Client for DeepSeek API (OpenAI-compatible format)."""
+
+    BASE_URL = "https://api.deepseek.com/v1/chat/completions"
+
+    def __init__(self, api_key, model="deepseek-chat", temperature=0.7, max_retries=3):
+        self.api_key = api_key
+        self.model = model
+        self.temperature = temperature
+        self.max_retries = max_retries
+        self.session = requests.Session()
+        self.session.headers.update({
+            "Authorization": f"Bearer {self.api_key}",
+            "Content-Type": "application/json"
+        })
+
+    def ask(self, system_prompt, user_prompt, temperature=None):
+        """Send a chat completion request. Returns response text."""
+        messages = [
+            {"role": "system", "content": system_prompt},
+            {"role": "user", "content": user_prompt}
+        ]
+        payload = {
+            "model": self.model,
+            "messages": messages,
+            "temperature": temperature or self.temperature
+        }
+
+        last_error = None
+        for attempt in range(self.max_retries):
+            try:
+                resp = self.session.post(self.BASE_URL, json=payload, timeout=60)
+                resp.raise_for_status()
+                data = resp.json()
+                result = data["choices"][0]["message"]["content"]
+                logger.info(
+                    "DeepSeek API OK | model=%s | input=%d chars | output=%d chars",
+                    self.model, len(user_prompt), len(result)
+                )
+                return result
+
+            except requests.exceptions.Timeout:
+                last_error = "Timeout"
+                logger.warning("Attempt %d/%d timed out", attempt + 1, self.max_retries)
+            except requests.exceptions.HTTPError as e:
+                last_error = str(e)
+                logger.warning("Attempt %d/%d HTTP error: %s", attempt + 1, self.max_retries, e)
+                if "insufficient_quota" in str(e).lower() or "invalid_api_key" in str(e).lower():
+                    raise  # don't retry these
+            except (json.JSONDecodeError, KeyError) as e:
+                last_error = f"Parse error: {e}"
+                logger.warning("Attempt %d/%d parse error: %s", attempt + 1, self.max_retries, e)
+
+            if attempt < self.max_retries - 1:
+                sleep_time = 2 ** attempt
+                logger.info("Retrying in %ds...", sleep_time)
+                time.sleep(sleep_time)
+
+        raise RuntimeError(
+            f"DeepSeek API failed after {self.max_retries} retries. Last error: {last_error}"
+        )
--- a/tools/memos_client.py
+++ b/tools/memos_client.py
@ -0,0 +1,138 @@
+"""Memos API client - fetch memos using Connect RPC protocol."""
+
+import logging
+from datetime import datetime, timezone, timedelta
+
+import requests
+
+logger = logging.getLogger(__name__)
+
+
+class MemosClient:
+    """Client for self-hosted Memos API (Connect RPC protocol)."""
+
+    def __init__(self, base_url, access_token):
+        self.base_url = base_url.rstrip("/")
+        self.session = requests.Session()
+        self.session.headers.update({
+            "Authorization": f"Bearer {access_token}",
+            "Content-Type": "application/json",
+        })
+        self._user_id = None
+
+    def get_user(self):
+        """Get current user info via Connect RPC."""
+        resp = self.session.post(
+            f"{self.base_url}/memos.api.v1.AuthService/GetCurrentUser",
+            json={},
+            timeout=10
+        )
+        resp.raise_for_status()
+        data = resp.json()
+        user = data.get("user", data)
+        self._user_id = user.get("name", "")
+        logger.info("Memos user: %s (%s)", user.get("username"), self._user_id)
+        return user
+
+    def get_user_id(self):
+        """Get current user's resource name (e.g. 'users/FXY')."""
+        if not self._user_id:
+            self.get_user()
+        return self._user_id
+
+    def list_memos(self, days=1, page_size=100):
+        """Fetch memos from the last N days via Connect RPC."""
+        user = self.get_user_id()
+        since = (datetime.now(timezone.utc) - timedelta(days=days)).isoformat()
+
+        # Connect RPC uses POST with JSON body
+        payload = {
+            "pageSize": page_size,
+            "filter": f"creator == '{user}'",
+        }
+
+        resp = self.session.post(
+            f"{self.base_url}/memos.api.v1.MemoService/ListMemos",
+            json=payload,
+            timeout=15
+        )
+        resp.raise_for_status()
+        data = resp.json()
+
+        memos = data.get("memos", [])
+        logger.info("Memos API OK | fetched %d memos", len(memos))
+
+        # Filter by time client-side (Connect RPC filter might not work as expected)
+        results = []
+        since_dt = datetime.now(timezone.utc) - timedelta(days=days)
+        for m in memos:
+            created = m.get("createTime", "")
+            try:
+                created_dt = datetime.fromisoformat(created.replace("Z", "+00:00"))
+            except (ValueError, AttributeError):
+                created_dt = datetime.now(timezone.utc)
+
+            if created_dt < since_dt:
+                continue
+
+            results.append({
+                "id": m.get("name", "").split("/")[-1],
+                "content": m.get("content", ""),
+                "created_at": created,
+                "tags": self._extract_tags(m.get("content", "")),
+                "visibility": m.get("visibility", ""),
+            })
+
+        logger.info(
+            "Filtered to %d memos since %s", len(results), since[:10]
+        )
+        return results
+
+    def list_all_memos_from_range(self, start_date, end_date, page_size=200):
+        """Fetch memos within a date range via Connect RPC."""
+        user = self.get_user_id()
+
+        payload = {"pageSize": page_size}
+
+        resp = self.session.post(
+            f"{self.base_url}/memos.api.v1.MemoService/ListMemos",
+            json=payload,
+            timeout=15
+        )
+        resp.raise_for_status()
+        data = resp.json()
+
+        memos = data.get("memos", [])
+        logger.info("Memos API OK | fetched %d memos total", len(memos))
+
+        # Client-side date range filtering
+        results = []
+        for m in memos:
+            created = m.get("createTime", "")
+            try:
+                created_dt = datetime.fromisoformat(created.replace("Z", "+00:00"))
+            except (ValueError, AttributeError):
+                continue
+
+            if created_dt < start_date or created_dt > end_date:
+                continue
+
+            results.append({
+                "id": m.get("name", "").split("/")[-1],
+                "content": m.get("content", ""),
+                "created_at": created,
+                "tags": self._extract_tags(m.get("content", "")),
+                "visibility": m.get("visibility", ""),
+            })
+
+        logger.info(
+            "Filtered to %d memos in range %s - %s",
+            len(results), start_date.strftime("%m/%d"), end_date.strftime("%m/%d")
+        )
+        return results
+
+    @staticmethod
+    def _extract_tags(content):
+        """Extract #tags from memo content."""
+        import re
+        return re.findall(r"#(\w[\w\-]*)", content)