Test ai

Author: lsbardel

pyproject.toml

@@ -22,6 +22,8 @@ Documentation = "https://quantmind.github.io/quantflow/"
 
 [project.optional-dependencies]
 ai = [
+    "asciichartpy>=1.5.25",
+    "ccy[holidays]>=1.7.1",
     "google-genai>=1.61.0",
     "mcp>=1.26.0",
     "openai>=2.16.0",

quantflow/ai/tools/base.py

@@ -1,11 +1,7 @@
 from dataclasses import dataclass, field
-from io import StringIO
 from pathlib import Path
 
-import pandas as pd
-from ccy.cli.console import df_to_rich
 from mcp.server.fastmcp.exceptions import ToolError
-from rich.console import Console
 
 from quantflow.data.fmp import FMP
 from quantflow.data.fred import Fred
@@ -35,10 +31,3 @@ def fred(self) -> Fred:
                 " Please add it using the vault_add tool."
             )
         return Fred(key=key)
-
-    def rich(self, df: pd.DataFrame) -> str:
-        table = df_to_rich(df)
-        buf = StringIO()
-        console = Console(file=buf, no_color=True)
-        console.print(table)
-        return buf.getvalue()

quantflow/ai/tools/crypto.py

@@ -37,7 +37,7 @@ async def crypto_historical_volatility(currency: str) -> str:
             df = await client.get_volatility(currency)
         if df.empty:
             return f"No volatility data for {currency}"
-        return f"Historical volatility for {currency}:\n{df.to_string(index=False)}"
+        return df.to_csv(index=False)
 
     @mcp.tool()
     async def crypto_term_structure(currency: str) -> str:
@@ -52,7 +52,7 @@ async def crypto_term_structure(currency: str) -> str:
             loader = await client.volatility_surface_loader(currency)
         vs: VolSurface = loader.surface()
         ts = vs.term_structure().round({"ttm": 4})
-        return f"Term structure for {currency}:\n{ts.to_string(index=False)}"
+        return ts.to_csv(index=False)
 
     @mcp.tool()
     async def crypto_implied_volatility(currency: str, maturity_index: int = -1) -> str:
@@ -71,7 +71,7 @@ async def crypto_implied_volatility(currency: str, maturity_index: int = -1) ->
         vs.bs(index=index)
         df = vs.options_df(index=index)
         df["implied_vol"] = df["implied_vol"].map("{:.2%}".format)
-        return f"Implied volatility for {currency}:\n{df.to_string(index=False)}"
+        return df.to_csv(index=False)
 
     @mcp.tool()
     async def crypto_prices(symbol: str, frequency: str = "") -> str:
@@ -87,4 +87,4 @@ async def crypto_prices(symbol: str, frequency: str = "") -> str:
         if df.empty:
             return f"No price data for {symbol}"
         df = df[["date", "open", "high", "low", "close", "volume"]].sort_values("date")
-        return f"Prices for {symbol}:\n{df.tail(50).to_string(index=False)}"
+        return df.tail(50).to_csv(index=False)

quantflow/ai/tools/fred.py

@@ -26,7 +26,7 @@ async def fred_subcategories(category_id: str | None = None) -> str:
         import pandas as pd
 
         df = pd.DataFrame(cats, columns=["id", "name"])
-        return f"FRED categories:\n{df.to_string(index=False)}"
+        return df.to_csv(index=False)
 
     @mcp.tool()
     async def fred_series(category_id: str) -> str:
@@ -55,7 +55,7 @@ async def fred_series(category_id: str) -> str:
                 "observation_end",
             ],
         ).sort_values("popularity", ascending=False)
-        return f"FRED series for category {category_id}:\n{df.to_string(index=False)}"
+        return df.to_csv(index=False)
 
     @mcp.tool()
     async def fred_data(
@@ -79,4 +79,4 @@ async def fred_data(
                     sort_order="desc",
                 )
             )
-        return f"FRED data for {series_id}:\n{df.to_string(index=False)}"
+        return df.to_csv(index=False)

quantflow/ai/tools/stocks.py

@@ -1,8 +1,12 @@
 """Stocks tools for the quantflow MCP server."""
 
 from datetime import timedelta
+from typing import cast
 
 import pandas as pd
+from ccy import period as to_period
+from ccy.tradingcentres import prevbizday
+from fluid.utils.data import compact_dict
 from mcp.server.fastmcp import FastMCP
 
 from quantflow.utils.dates import utcnow
@@ -17,7 +21,7 @@ async def stock_indices() -> str:
         """List available stock market indices."""
         async with tool.fmp() as client:
             data = await client.indices()
-        return tool.rich(pd.DataFrame(data))
+        return pd.DataFrame(data).to_csv(index=False)
 
     @mcp.tool()
     async def stock_search(query: str) -> str:
@@ -30,7 +34,7 @@ async def stock_search(query: str) -> str:
             data = await client.search(query)
 
         df = pd.DataFrame(data, columns=["symbol", "name", "currency", "stockExchange"])
-        return f"Search results for '{query}':\n{df.to_string(index=False)}"
+        return df.to_csv(index=False)
 
     @mcp.tool()
     async def stock_profile(symbol: str) -> str:
@@ -62,7 +66,7 @@ async def stock_prices(symbol: str, frequency: str = "") -> str:
         if df.empty:
             return f"No price data for {symbol}"
         df = df[["date", "open", "high", "low", "close", "volume"]].sort_values("date")
-        return f"Prices for {symbol}:\n{df.tail(50).to_string(index=False)}"
+        return df.to_csv(index=False)
 
     @mcp.tool()
     async def sector_performance(period: str = "1d") -> str:
@@ -71,10 +75,6 @@ async def sector_performance(period: str = "1d") -> str:
         Args:
             period: Time period - 1d, 1w, 1m, 3m, 6m, 1y (default: 1d)
         """
-        from ccy import period as to_period
-        from ccy.tradingcentres import prevbizday
-        from fluid.utils.data import compact_dict
-
         async with tool.fmp() as client:
             to_date = utcnow().date()
             if period != "1d":
@@ -89,16 +89,11 @@ async def sector_performance(period: str = "1d") -> str:
             pe = await client.sector_pe(
                 params=compact_dict(date=prevbizday(to_date, 0).isoformat())
             )
-
-        from typing import cast
-
-        import pandas as pd
-
         spd = cast(dict, sp)
         pes = {k["sector"]: round(float(k["pe"]), 3) for k in pe if k["sector"] in spd}
         rows = [
             {"sector": k, "performance": float(v), "pe": pes.get(k, float("nan"))}
             for k, v in spd.items()
         ]
         df = pd.DataFrame(rows).sort_values("performance", ascending=False)
-        return f"Sector performance ({period}):\n{df.to_string(index=False)}"
+        return df.to_csv(index=False)