360 lines
13 KiB
Python
360 lines
13 KiB
Python
"""AI 分析核心服务
|
||
|
||
负责：
|
||
1. 调用 OpenAI 兼容 API 进行分析
|
||
2. Function Calling 循环（AI 可主动获取数据）
|
||
3. 采集当日盘面快照（供次日环比）
|
||
4. 保存报告到数据库
|
||
"""
|
||
|
||
import json
|
||
import re
|
||
import traceback
|
||
|
||
import httpx
|
||
from datetime import datetime, timezone, timedelta
|
||
|
||
from services.ai_config import AI_API_BASE, AI_API_KEY, AI_MODEL, AI_MAX_TOKENS, AI_TEMPERATURE
|
||
from services.ai_tools import TOOLS, execute_tool
|
||
from database import get_connection
|
||
|
||
_CST = timezone(timedelta(hours=8))
|
||
|
||
SYSTEM_PROMPT = """你是一位专业的A股市场分析师，擅长从数据中发现投资机会。
|
||
|
||
你的分析风格：
|
||
- 数据驱动，基于真实数据而非主观臆断
|
||
- 逻辑清晰，先总后分，层层递进
|
||
- 观点明确，给出具体的操作建议
|
||
- 风险提示，每次推荐都需说明风险点
|
||
|
||
可用工具：
|
||
- get_market_dashboard: 获取市场整体数据（指数/涨跌统计/市场温度/连板梯队/行业强度/事件情报）
|
||
- get_theme_history: 获取指定日期的题材涨幅排行
|
||
- get_active_core_stocks: 获取核心股追踪数据（10日涨幅矩阵+所属题材）
|
||
- get_stock_quote: 获取个股实时行情
|
||
- get_fund_flow: 获取个股资金流向
|
||
|
||
重要规则：
|
||
1. 你必须先调用工具获取数据，然后基于数据进行分析
|
||
2. 不要凭空编造数据，所有数据必须来自工具返回
|
||
3. 如果工具返回空数据，如实说明数据不可用
|
||
4. 分析完成后给出明确的结论和建议"""
|
||
|
||
DAILY_ANALYSIS_PROMPT = """请对 {trade_date} 的A股市场进行收盘分析，生成一份完整的分析报告。
|
||
|
||
{prev_report_section}
|
||
{prev_snapshot_section}
|
||
|
||
请先调用以下工具获取数据：
|
||
1. get_market_dashboard - 获取市场整体数据（含涨跌统计、市场温度、连板梯队 limitLadder）
|
||
2. get_theme_history(date="{trade_date}") - 获取今日题材涨幅
|
||
3. get_active_core_stocks - 获取核心股数据
|
||
|
||
生成报告时必须遵守以下格式规则：
|
||
|
||
1. 报告标题（# 一级标题）之后的第一行，必须是一个引用块"定调摘要"，格式严格为：
|
||
> 今日定调：<一句话核心结论，不超过80字，必须包含1-2个关键数字（如成交额、涨停家数、市场温度）>
|
||
|
||
2. 量能与情绪类数字必须给环比：若上方提供了"前一交易日盘面数据快照"，成交额、涨跌家数、涨停数、市场温度等在与昨日对比后表述（如"成交额2.05万亿，较昨日缩量约700亿"）；没有昨日快照则如实说明"暂无昨日数据"。
|
||
|
||
3. 连板梯队必须完整呈现 get_market_dashboard 返回的 limitLadder：从最高连板到2连板逐级列表格，每只标注涨停原因（reason字段）与封单金额（sealWan，单位万，为空则不写）；首板只挑3-5只人气最高的点评。
|
||
|
||
4. 适当使用表格展示数据对比。
|
||
|
||
然后基于数据生成报告，结构如下：
|
||
|
||
## 一、市场总览
|
||
- 主要指数表现（上证、深证、创业板、科创50）
|
||
- 涨跌家数统计（含涨停/跌停/炸板率，须环比）
|
||
- 市场温度评估
|
||
|
||
## 二、题材热点分析
|
||
- 今日涨幅前5题材
|
||
- 持续活跃的题材
|
||
- 新兴热点题材
|
||
- 明显退潮的题材（警示）
|
||
|
||
## 三、核心股追踪
|
||
- 连板梯队分析（按格式规则3完整呈现）
|
||
- 核心股表现
|
||
- 龙头股辨识
|
||
|
||
## 四、关注方向
|
||
- 明日值得关注的题材方向
|
||
- 潜在的交易机会
|
||
|
||
## 五、下个交易日建议
|
||
- 明日大盘预判（支撑/压力位）
|
||
- 建议关注的题材方向（2-3个）
|
||
- 建议关注的核心股（附理由）
|
||
- 操作策略（仓位建议、买卖时机）
|
||
- 需要规避的方向
|
||
|
||
## 六、风险提示
|
||
- 需要警惕的风险因素
|
||
- 操作建议
|
||
|
||
请用 Markdown 格式输出，适当使用表格展示数据对比。"""
|
||
|
||
# 报告头部的"今日定调"引用行，保存时提取为 summary
|
||
_TONE_LINE_RE = re.compile(r"^>\s*今日定调[：:]\s*(.+)$", re.MULTILINE)
|
||
|
||
|
||
async def call_llm(messages: list, tools: list = None) -> dict:
|
||
"""调用 OpenAI 兼容 API"""
|
||
async with httpx.AsyncClient() as client:
|
||
payload = {
|
||
"model": AI_MODEL,
|
||
"messages": messages,
|
||
}
|
||
if AI_MAX_TOKENS is not None:
|
||
payload["max_tokens"] = AI_MAX_TOKENS
|
||
if AI_TEMPERATURE is not None:
|
||
payload["temperature"] = AI_TEMPERATURE
|
||
if tools:
|
||
payload["tools"] = tools
|
||
payload["tool_choice"] = "auto"
|
||
|
||
resp = await client.post(
|
||
f"{AI_API_BASE}/chat/completions",
|
||
headers={
|
||
"Authorization": f"Bearer {AI_API_KEY}",
|
||
"Content-Type": "application/json",
|
||
},
|
||
json=payload,
|
||
timeout=120,
|
||
)
|
||
resp.raise_for_status()
|
||
return resp.json()
|
||
|
||
|
||
async def collect_ai_analysis(trade_date: str) -> dict:
|
||
"""AI 分析主流程
|
||
|
||
Args:
|
||
trade_date: 交易日 YYYY-MM-DD
|
||
|
||
Returns:
|
||
{"id": int, "tokens_used": int, "tools_used": list}
|
||
"""
|
||
prev_report_section = _get_prev_report_section(trade_date)
|
||
# 采集当日盘面快照（供次日环比），并读取上一交易日快照注入 prompt
|
||
await _capture_market_snapshot(trade_date)
|
||
prev_snapshot_section = _get_prev_snapshot_section(trade_date)
|
||
|
||
messages = [
|
||
{"role": "system", "content": SYSTEM_PROMPT},
|
||
{"role": "user", "content": DAILY_ANALYSIS_PROMPT.format(
|
||
trade_date=trade_date,
|
||
prev_report_section=prev_report_section,
|
||
prev_snapshot_section=prev_snapshot_section,
|
||
)}
|
||
]
|
||
|
||
tools_used = []
|
||
total_tokens = 0
|
||
final_content = ""
|
||
was_truncated = False
|
||
max_rounds = 30 # 安全上限，正常分析约 3-8 轮
|
||
|
||
for i in range(max_rounds):
|
||
response = await call_llm(messages, tools=TOOLS)
|
||
total_tokens += response.get("usage", {}).get("total_tokens", 0)
|
||
|
||
choice = response["choices"][0]
|
||
message = choice["message"]
|
||
messages.append(message)
|
||
|
||
finish_reason = choice.get("finish_reason", "")
|
||
|
||
if finish_reason == "stop":
|
||
final_content = message.get("content") or ""
|
||
break
|
||
|
||
if finish_reason == "length":
|
||
was_truncated = True
|
||
final_content = message.get("content") or ""
|
||
if not final_content:
|
||
for msg in reversed(messages):
|
||
if msg.get("role") == "assistant" and msg.get("content"):
|
||
final_content = msg["content"]
|
||
break
|
||
print(f"[ai-service] 警告：响应被截断 (finish_reason=length)")
|
||
break
|
||
|
||
if finish_reason == "tool_calls":
|
||
for tool_call in message.get("tool_calls", []):
|
||
func_name = tool_call["function"]["name"]
|
||
func_args = json.loads(tool_call["function"]["arguments"])
|
||
tools_used.append(func_name)
|
||
|
||
result = await execute_tool(func_name, func_args)
|
||
messages.append({
|
||
"role": "tool",
|
||
"tool_call_id": tool_call["id"],
|
||
"content": result
|
||
})
|
||
|
||
summary = _extract_summary(final_content)
|
||
report_id = _save_report(trade_date, final_content, summary, tools_used, total_tokens)
|
||
|
||
return {"id": report_id, "tokens_used": total_tokens, "tools_used": tools_used, "truncated": was_truncated}
|
||
|
||
|
||
def _extract_summary(content: str) -> str:
|
||
"""提取报告头部的"今日定调"引用行作为摘要；缺失时退回正文截断"""
|
||
match = _TONE_LINE_RE.search(content or "")
|
||
if match:
|
||
text = match.group(1).strip()
|
||
if text:
|
||
return text[:200]
|
||
return (content or "")[:200].replace("\n", " ")
|
||
|
||
|
||
def _num(v) -> str:
|
||
"""快照数值安全转字符串"""
|
||
if isinstance(v, float):
|
||
return f"{v:g}"
|
||
return str(v if v is not None else "-")
|
||
|
||
|
||
def _fmt_amount(v) -> str:
|
||
"""成交额（元）→ 万亿/亿 可读格式"""
|
||
try:
|
||
v = float(v)
|
||
except (TypeError, ValueError):
|
||
return "-"
|
||
if v >= 1e12:
|
||
return f"{v / 1e12:.2f}万亿"
|
||
if v >= 1e8:
|
||
return f"{v / 1e8:.0f}亿"
|
||
return f"{v:.0f}元"
|
||
|
||
|
||
def _fmt_index(idx: dict) -> str:
|
||
if not idx:
|
||
return "-"
|
||
sign = "+" if idx.get("changePct", 0) >= 0 else ""
|
||
return f"{idx.get('price', '-')}（{sign}{idx.get('changePct', 0)}%）"
|
||
|
||
|
||
async def _capture_market_snapshot(trade_date: str) -> bool:
|
||
"""采集当日盘面快照入库（指数+涨跌统计），供次日分析做环比"""
|
||
try:
|
||
from routes.market_dashboard import _build_dashboard
|
||
data = await _build_dashboard()
|
||
payload = json.dumps(
|
||
{"indices": data.get("indices", []), "marketStats": data.get("marketStats", {})},
|
||
ensure_ascii=False,
|
||
default=str,
|
||
)
|
||
conn = get_connection()
|
||
try:
|
||
conn.execute(
|
||
"""INSERT INTO daily_market_stats (trade_date, payload) VALUES (?, ?)
|
||
ON CONFLICT(trade_date) DO UPDATE SET payload = excluded.payload""",
|
||
(trade_date, payload),
|
||
)
|
||
conn.commit()
|
||
finally:
|
||
conn.close()
|
||
return True
|
||
except Exception:
|
||
print(f"[ai-service] 市场快照采集失败 {trade_date}：")
|
||
traceback.print_exc()
|
||
return False
|
||
|
||
|
||
def _get_prev_snapshot_section(trade_date: str) -> str:
|
||
"""读取上一交易日盘面快照，格式化为 prompt 中的环比数据段"""
|
||
conn = get_connection()
|
||
try:
|
||
row = conn.execute(
|
||
"SELECT trade_date, payload FROM daily_market_stats WHERE trade_date < ? ORDER BY trade_date DESC LIMIT 1",
|
||
(trade_date,),
|
||
).fetchone()
|
||
if not row:
|
||
return ""
|
||
try:
|
||
snap = json.loads(row["payload"])
|
||
except (TypeError, ValueError):
|
||
return ""
|
||
stats = snap.get("marketStats") or {}
|
||
indices = {i.get("name"): i for i in (snap.get("indices") or []) if isinstance(i, dict)}
|
||
idx_line = "、".join(
|
||
f"{name} {_fmt_index(indices.get(name))}"
|
||
for name in ("上证指数", "深证成指", "创业板指", "科创50")
|
||
)
|
||
return f"""以下是前一交易日（{row["trade_date"]}）的盘面数据快照，报告中的量能与情绪数字必须给出与它的环比对比：
|
||
|
||
- 两市成交额：{_fmt_amount(stats.get("totalTurnover"))}
|
||
- 上涨/下跌/平盘：{_num(stats.get("upCount"))}/{_num(stats.get("downCount"))}/{_num(stats.get("flatCount"))}，涨停 {_num(stats.get("limitUp"))} 家、跌停 {_num(stats.get("limitDown"))} 家、炸板 {_num(stats.get("limitBreak"))} 家（炸板率 {_num(stats.get("breakRate"))}%）
|
||
- 强势/弱势股：{_num(stats.get("strongCount"))}/{_num(stats.get("weakCount"))}，市场宽度 {_num(stats.get("marketBreadth"))}%
|
||
- 市场温度：{_num(stats.get("temperature"))} 分，竞价信号：{stats.get("auctionSignal") or "-"}
|
||
- 指数收盘：{idx_line}
|
||
|
||
---
|
||
"""
|
||
finally:
|
||
conn.close()
|
||
|
||
|
||
def _save_report(trade_date: str, content: str, summary: str, tools_used: list, tokens_used: int) -> int:
|
||
"""保存报告到数据库（同日重生成：覆盖内容、generation_count+1、tokens 记当次消耗）"""
|
||
conn = get_connection()
|
||
try:
|
||
conn.execute(
|
||
"""INSERT INTO ai_reports
|
||
(trade_date, report_type, title, content, summary, tools_used, model, tokens_used, updated_at)
|
||
VALUES (?, 'daily', ?, ?, ?, ?, ?, ?, datetime('now','localtime'))
|
||
ON CONFLICT(trade_date, report_type) DO UPDATE SET
|
||
title = excluded.title,
|
||
content = excluded.content,
|
||
summary = excluded.summary,
|
||
tools_used = excluded.tools_used,
|
||
model = excluded.model,
|
||
tokens_used = excluded.tokens_used,
|
||
updated_at = excluded.updated_at,
|
||
generation_count = ai_reports.generation_count + 1""",
|
||
(
|
||
trade_date,
|
||
f"{trade_date} A股收盘分析",
|
||
content,
|
||
summary,
|
||
json.dumps(tools_used),
|
||
AI_MODEL,
|
||
tokens_used,
|
||
),
|
||
)
|
||
conn.commit()
|
||
row = conn.execute(
|
||
"SELECT id FROM ai_reports WHERE trade_date = ? AND report_type = 'daily'",
|
||
(trade_date,),
|
||
).fetchone()
|
||
return row["id"]
|
||
finally:
|
||
conn.close()
|
||
|
||
|
||
def _get_prev_report_section(trade_date: str) -> str:
|
||
"""获取前一个交易日的报告，用于上下文参考"""
|
||
conn = get_connection()
|
||
try:
|
||
row = conn.execute(
|
||
"SELECT trade_date, content FROM ai_reports WHERE trade_date < ? AND report_type = 'daily' ORDER BY trade_date DESC LIMIT 1",
|
||
(trade_date,)
|
||
).fetchone()
|
||
if not row:
|
||
return ""
|
||
prev_date = row["trade_date"]
|
||
prev_content = row["content"] or ""
|
||
return f"""以下是前一个交易日（{prev_date}）的分析报告，请参考其中的分析逻辑和关注方向，结合今日数据进行对比分析：
|
||
|
||
{prev_content}
|
||
|
||
---
|
||
"""
|
||
finally:
|
||
conn.close()
|