Update app.py

Month, Week, Day deltas for stocks

app.py

@@ -129,11 +129,11 @@ def get_stock_news(search_symbol):
         rv += f'Link: [URL]({item["link"]})\n\n'
 
     if have_symbol:
-        (plot_df, ymax, pct_delta) = stock_week_df(search_symbol)
+        (plot_df, ymax, deltas) = stock_week_df(search_symbol)
     else:
-        (plot_df, ymax, pct_delta) = (pd.DataFrame(), 0.0, 0.0)
+        (plot_df, ymax, deltas) = (pd.DataFrame(), 0.0, (0.0, 0.0, 0.0))
 
-    return (rv, plot_df, ymax, pct_delta)
+    return (rv, plot_df, ymax, deltas)
 
 def stock_history_df(num_weeks):
     values = []
@@ -177,18 +177,30 @@ def stock_history_df(num_weeks):
 #     # ax.set_xlim(dates[0], dates[-1:])
 #     return fig
 
-def lms_fit_trend(dates, values):
-    # days = []
-    # fit_data = []
-    days = list(range(0, len(dates)))
-    fit = np.polyfit(days, values, 1)
-    delta = len(dates) * float(fit[0])
-    avg = float(fit[1]) + 0.5 * delta
-    pct_delta = 100 * delta / avg
-    # for day in days:
-    #     fit_data.append(float(fit[0]) * day + float(fit[1]) )
-    # return fit_data
-    return pct_delta
+# def lms_fit_trend(dates, values):
+#     # days = []
+#     # fit_data = []
+#     days = list(range(0, len(dates)))
+#     fit = np.polyfit(days, values, 1)
+#     delta = len(dates) * float(fit[0])
+#     avg = float(fit[1]) + 0.5 * delta
+#     pct_delta = 100 * delta / avg
+#     # for day in days:
+#     #     fit_data.append(float(fit[0]) * day + float(fit[1]) )
+#     # return fit_data
+#     return pct_delta
+
+def stock_deltas(values):
+    num = len(values)
+    month_end_avg = float(np.average(np.array(values[-3:])))
+    month_start_avg = float(np.average(np.array(values[0:4])))
+    week_start_avg = float(np.average(np.array(values[-7:-4])))
+    week_end_avg = float(np.average(np.array(values[-2:])))
+    month_delta = 100 * (month_end_avg - month_start_avg)/month_start_avg
+    week_delta = 100 * (week_end_avg - week_start_avg)/week_start_avg
+    daily_delta = 100 * ((float(values[num-1])/float(values[num-2])) - 1.0)
+    # avg = np.average(npa)
+    return (month_delta, week_delta, daily_delta)
 
 def stock_week_df(symbol):
     try:
@@ -219,16 +231,17 @@ def stock_week_df(symbol):
         for row in df.index:
             dates.append(row.strftime('%Y-%m-%d'))
         # fit_data = lms_fit(dates, values)
-        pct_delta = lms_fit_trend(dates, values)
+        # pct_delta = lms_fit_trend(dates, values)
+        deltas = stock_deltas(values)
         data = {
             "date": dates,
             "value" : values,
             # "fit" : fit_data
             }
         # fig = make_mp_figure(dates, values, fit_data, ymax)
-        return (pd.DataFrame(data), ymax, pct_delta)
+        return (pd.DataFrame(data), ymax, deltas)
     except:
-        return (pd.DataFrame(), ymax, 0.0)
+        return (pd.DataFrame(), ymax, (0.0, 0.0, 0.0))
 
 def get_stock_report(verbose = True, offset = 0):
     try:
@@ -613,9 +626,9 @@ def chat(prompt, user_window, pwd_window, past, response, gptModel, uploaded_ima
             elif num >= 3:
                 if args[1] == 'news':
                     symbol = ' '.join(args[2:])
-                    (response, plot_df, ymax, pct_delta) = get_stock_news(symbol)
+                    (response, plot_df, ymax, (dm, dw, dd)) = get_stock_news(symbol)
                     ymax *= 1.1
-                    mdtxt = md(f'News for {symbol}:\nTrend = {int(pct_delta)}%\n\n' + response)
+                    mdtxt = md(f'News for {symbol}:\nTrends: Month = {dm:.1f}%, Week = {dw:.1f}%, Day = {dd:.1f}%\n\n' + response)
                     if plot_df.empty:
                         return [past, mdtxt, None, gptModel, uploaded_image_file, plot]
                     else: