bug fix

cleanData.py

@@ -0,0 +1,10 @@
+for i in stockIds:
+    path = get_event_path(i,'finance_report')
+    try:
+        if os.path.exists(path):
+            df = pd.read_csv(path)
+            df = index_df_with_time(df,index='reportDate')
+    except Exception as e:
+        print(path)
+        if os.path.exists(path):
+            os.remove(path)

fooltrader/api/finance.py

@@ -236,7 +236,7 @@ def get_balance_sheet_items(security_item, start_date=None, report_period=None,
             if start_date:
                 if pd.Timestamp(reportDate[idx]) < pd.Timestamp(start_date):
                     continue
-
+            print(reportDate[idx])
             reportEventDate = get_report_event_date(security_item, report_date=reportDate[idx])
 
             if report_period and not is_same_date(report_period, reportDate[idx]):

fooltrader/spiders/stock_finance_report_event_spider.py

@@ -50,6 +50,9 @@ def yield_request(self, item):
     def report_period_from_title(title, period_type, report_event_date):
         try:
             year = re.match('.*(\d{4}).*', title).group(1)
+            report_event_year = pd.Timestamp(report_event_date).date().year
+            if int(year)<int(report_event_year)-2 or int(year)>int(report_event_year):
+                raise Exception('wrong report year')
         except Exception as e:
             year = pd.Timestamp(report_event_date).date().year
             if period_type == 'ndbg':