added timing

fof/utility.py

@@ -3,6 +3,7 @@ import calendar
 
 import pymssql
 import sqlalchemy as sa
+import numpy as np
 import pandas as pd
 
 
@@ -63,7 +64,7 @@ def read_sql(query, source="mssql"):
     return pd.read_sql(query, con=engine)
 
 
-def nearest_report_date(date, freq="Q"):
+def nearest_report_date(date, freq="Q", method="forward"):
     if len(date) == 10:
         date = date.replace("-", "")
 
@@ -71,10 +72,15 @@ def nearest_report_date(date, freq="Q"):
     year = int(date[0:4])
     last_year = year
     
+    if method == "forward":
+        add = 1
+    else:
+        add = 0
+    
     if freq == "Q":
-        q = (mm // 3 + 1) * 3 - 3
+        q = (mm // 3 + add) * 3 - 3
     else:
-        q = (mm // 6 + 1) * 6 - 6
+        q = (mm // 6 + add) * 6 - 6
         
     if q <= 0:
         q += 12
@@ -217,8 +223,8 @@ def fd_assetportfolio(security_ids, report_dates_begin, report_dates_end=None):
         query = f"""
         SELECT SECODE as SECURITYID, REPORTDATE, BDRTO, CONVBDRTO, EQUITYINVERTO FROM TQ_FD_ASSETPORTFOLIO
         WHERE
-            REPORTDATE >= '{report_dates_begin}' AND
-            REPORTDATE <= '{report_dates_end}' AND
+            REPORTDATE >= '{_to_yyyymmdd(report_dates_begin)}' AND
+            REPORTDATE <= '{_to_yyyymmdd(report_dates_end)}' AND
             ISVALID = 1 AND
             SECODE in ({sec_id_strs})
         """
@@ -277,15 +283,21 @@ def fd_skdetail(security_ids, trade_dt):
     p_id_strs = _join_ids(security_ids)
 
     query = f"""
-    SELECT SECODE as pid, SKCODE as security_code, SKNAME, HOLDMKTCAP, HOLDAMT, NAVRTO FROM TQ_FD_SKDETAIL
+    SELECT SECODE as pid, PUBLISHDATE, SKCODE as security_code, SKNAME, HOLDMKTCAP, HOLDAMT, NAVRTO FROM TQ_FD_SKDETAIL
     WHERE
         ENDDATE = '{trade_dt}' AND
         ISVALID = 1 AND
         SECODE in ({p_id_strs})
     """
-    df = read_sql(query).sort_values("security_code")
-    total_value = df.HOLDMKTCAP.sum()
-    df["weight"] = df.HOLDMKTCAP / total_value
+    groups = read_sql(query).sort_values("security_code").groupby("pid")
+    dfs = []
+    for k, g in groups:
+        g = g[g.PUBLISHDATE == np.max(g.PUBLISHDATE)]
+        total_value = g.HOLDMKTCAP.sum()
+        g["weight"] = g.HOLDMKTCAP / total_value
+        dfs.append(g)
+    df = pd.concat(dfs)
+    del df["PUBLISHDATE"]
     return df.reset_index(drop=True)
 
 
@@ -530,3 +542,20 @@ def read_factors(security_ids, trade_dt, factors):
         flag = 1
     """
     return read_sql(query, source="mysql").sort_values("security_code").reset_index(drop=True)
+
+
+def index_rtn(security_ids, begin_date, end_date):
+    sec_id_strs = _join_ids(security_ids)
+    
+    query = f"""
+    SELECT security_code, min(trade_date) AS begin_date, max(trade_date) AS end_date, exp(sum(ln(1 + change_pct / 100.))) - 1.0 AS rtn FROM index_daily_price
+    WHERE
+        trade_date >= '{_to_yyyy_mm_dd(begin_date)}' AND
+        trade_date <= '{_to_yyyy_mm_dd(end_date)}' AND
+        security_code in ({sec_id_strs}) AND
+        flag = 1
+    GROUP BY
+        security_code
+    """
+    df = read_sql(query, source="mysql").sort_values("security_code").reset_index(drop=True)
+    return df