Update app.py

app.py

@@ -1,43 +1,34 @@
-import numpy as np 
-import pandas as pd 
+import os
+
+import numpy as np
+import pandas as pd
 
 from datetime import datetime, date
 
 import yfinance as yf
 
-
-
-import os
-
 from dotenv import load_dotenv
 
 import holoviews as hv
 import hvplot.pandas  # noqa
-
-
-import hvplot.pandas  # Provides hvplot method for DataFrames
-import holoviews as hv
-from holoviews import opts
 import panel as pn
-
 from holoviews.plotting.links import RangeToolLink
 
-pn.extension("bokeh", template="bootstrap")
-
+pn.extension("bokeh", template="fast")
 
-# from datasets import load_dataset
-from utils import load_hf_dataset  # ,upload_to_hf_dataset, download_from_hf_dataset
+from datasets import load_dataset
+from utils import load_hf_dataset, upload_to_hf_dataset, download_from_hf_dataset
 
 import stumpy
 # import matplotlib.pyplot as plt
 
-
-
+### Preparting the data
 # Load environment variables from .env file
 load_dotenv()
 
 # Get the Hugging Face API token from the environment; either set in .env file or in the environment directly in GitHub
 HF_TOKEN = os.getenv("HF_TOKEN")
+
 # Get the name of the GuruFocus dataset for TradingView, Finviz, MarketBeat and TipRanks to read from
 dataset_name_TradingView_input = os.getenv("dataset_name_TradingView_input")
 
@@ -49,15 +40,15 @@ def get_tradingview(current_datetime):
     return load_hf_dataset("america.csv", HF_TOKEN, dataset_name_TradingView_input)
 
 
-
 # Reading gurufocus,finviz from crawling pipelines with GitHub Action
 current_datetime = datetime.now().strftime("%Y-%m-%d")
 
 DF = get_tradingview(current_datetime)
-ticker_list = list(DF.query(' `Market Capitalization`>10e9').Ticker)
+ticker_list = list(DF.query(" `Market Capitalization`>10e9").Ticker)
 
-ticker1 ="CRM"
-ticker2= "MSFT"
+# Widgets
+ticker1 = "CRM"
+ticker2 = "MSFT"
 
 ticker1_widget = pn.widgets.AutocompleteInput(
     name="Ticker A",
@@ -76,77 +67,91 @@ ticker2_widget = pn.widgets.AutocompleteInput(
 )
 
 m_widget = pn.widgets.IntSlider(
-    name='Window Size (m)',
-    value=200,
-    start=50,
-    end=400,
-    step=10
+    name="Window Size (m)", value=250, start=5, end=400, step=10
 )
 
+similarity_rank_widget = pn.widgets.IntSlider(
+    name="Similarity Rank", value=0, start=0, end=50, step=1
+)
 
 # Create a DatePicker widget with a minimum date of 2000-01-01
 date_start = pn.widgets.DatePicker(
-    name ="Start Date",
-    description='Select a Date',
-    start= date(2000, 1, 1)
+    name="Start Date", description="Select a Date", start=date(2000, 1, 1)
 )
 
 date_end = pn.widgets.DatePicker(
-    name ="End Date",# value=datetime(2000, 1, 1),
-    description='Select a Date',
-    end= date.today() #date(2023, 9, 1)
+    name="End Date",  # value=datetime(2000, 1, 1),
+    description="Select a Date",
+    end=date.today(),  # date(2023, 9, 1)
 )
 
-date_start.value = date(2000,1,1)
+date_start.value = date(2000, 1, 1)
 date_end.value = date.today()
 
-def get_plot(ticker1,ticker2, m_widget, date_start , date_end):
-
-    DF1 = yf.Ticker(ticker1).history(start=date_start , end=date_end); DF1 = DF1.resample('5D').mean(); DF1['Date'] = DF1.index
-    DF2 = yf.Ticker(ticker2).history(start=date_start , end=date_end); DF2 = DF2.resample('5D').mean(); DF2['Date'] = DF2.index
-
-    m = 200
-    m = m_widget
-    varcol = 'Close'
-    mp = stumpy.stump(T_A = DF1[varcol],
-                            m = m,
-                            T_B = DF2[varcol],
-                            ignore_trivial = False)
-
-    ticker1_motif_index = mp[:, 0].argmin()
+@pn.cache()
+def get_DF(ticker1, date_start, date_end):
+    DF = yf.Ticker(ticker1).history(start=date_start, end=date_end)
+    return DF
+
+def get_plot(ticker1, ticker2, m_widget, similarity_rank_widget, date_start, date_end):
+    DF1 = get_DF(ticker1, date_start, date_end)
+    DF1 = DF1.resample("5D").mean()
+    DF1["Date"] = DF1.index
+    DF2 = get_DF(ticker2, date_start, date_end)
+    DF2 = DF2.resample("5D").mean()
+    DF2["Date"] = DF2.index
+
+    m = m_widget  # m = 200
+    varcol = "Close"
+    mp = stumpy.stump(T_A=DF1[varcol], m=m, T_B=DF2[varcol], ignore_trivial=True)
+
+    # ticker1_motif_index = mp[:, 0].argmin()
+    ticker1_motif_index = np.argpartition(mp[:, 0],similarity_rank_widget)[similarity_rank_widget]
     print(f'The motif is located at index {ticker1_motif_index} of "{ticker1}"')
 
     ticker2_motif_index = mp[ticker1_motif_index, 1]
     print(f'The motif is located at index {ticker2_motif_index} of "{ticker2}"')
 
-
-    plt1 = hv.Curve( DF1.iloc[ticker1_motif_index : ticker1_motif_index + m][varcol].values, label=f'{ticker1}')
-    plt2 = hv.Curve(DF2.iloc[ticker2_motif_index:ticker2_motif_index+m][varcol].values, label=f'{ticker2}')
-
-
-    plt3_1 = hv.Curve(DF1[varcol]) 
-    # plt3_2= hv.Curve(DF1.iloc[ticker1_motif_index : ticker1_motif_index + m ,-1].values , DF1.iloc[ticker1_motif_index : ticker1_motif_index + m][varcol].values)
-
-    plt4_1 = hv.Curve(DF2[varcol])
-    # plt4_2=hv.Curve(DF2.iloc[ticker2_motif_index:ticker2_motif_index+m ,-1].values , DF2.iloc[ticker2_motif_index:ticker2_motif_index+m][varcol].values)
+    plt1 = hv.Curve(
+        DF1.iloc[ticker1_motif_index : ticker1_motif_index + m][varcol].values,
+        label=f"{ticker1}",
+    )
+    plt2 = hv.Curve(
+        DF2.iloc[ticker2_motif_index : ticker2_motif_index + m][varcol].values,
+        label=f"{ticker2}",
+    )
 
     # Plot for DF1
-    plot1 = DF1.hvplot.line(x='Date', y=varcol, title=ticker1).opts(width=500, height=400)
-    motif1 = DF1.iloc[ticker1_motif_index: ticker1_motif_index + m]
-    motif1_plot = motif1.hvplot.line( y=varcol, color='red')
+    plot1 = DF1.hvplot.line(x="Date", y=varcol, title=ticker1).opts(
+        width=500,
+        height=400,
+        show_grid=True,
+        ylim=(DF1[varcol].min(), DF1[varcol].max()),
+    )
+    motif1 = DF1.iloc[ticker1_motif_index : ticker1_motif_index + m]
+    motif1_plot = motif1.hvplot.line(y=varcol, color="red")
     combined_plot1 = plot1 * motif1_plot  # Overlay the motif on the main plot
 
-    # # Plot for DF2
-    plot2 = DF2.hvplot.line(x='Date', y=varcol, title=ticker2).opts(width=500, height=400)
-    motif2 = DF2.iloc[ticker2_motif_index: ticker2_motif_index + m]
-    motif2_plot = motif2.hvplot.line( y=varcol, color='red')
+    # Plot for DF2
+    plot2 = DF2.hvplot.line(x="Date", y=varcol, title=ticker2).opts(
+        width=500,
+        height=400,
+        show_grid=True,
+        ylim=(DF1[varcol].min(), DF1[varcol].max()),
+    )
+    motif2 = DF2.iloc[ticker2_motif_index : ticker2_motif_index + m]
+    motif2_plot = motif2.hvplot.line(y=varcol, color="red")
     combined_plot2 = plot2 * motif2_plot  # Overlay the motif on the main plot
 
-
-    # return pn.Row(plt1*plt2 , pn.Column(plt3_1.opts(width=800, height=400) ,plt4_1.opts(width=800, height=400) ))   
-    return pn.Row(plt1*plt2 , pn.Column(combined_plot1.opts(width=800, height=400) ,combined_plot2.opts(width=800, height=400) ))   
-
-
-pn.Row( pn.Column(ticker1_widget, ticker2_widget, m_widget, date_start , date_end),
-    pn.bind(get_plot, ticker1_widget, ticker2_widget , m_widget, date_start , date_end)).servable(title="Find Similarity between two stocks")
-
+    return pn.Row(
+        plt1 * plt2.opts(width=500, height=400),
+        pn.Column(
+            combined_plot1.opts(width=800, height=400),
+            combined_plot2.opts(width=800, height=400),
+        ),
+    )
+
+pn.Row(
+    pn.Column(ticker1_widget, ticker2_widget, m_widget, similarity_rank_widget, date_start, date_end),
+    pn.bind(get_plot, ticker1_widget, ticker2_widget, m_widget, similarity_rank_widget, date_start, date_end),
+).servable(title="Find Similarity in Stock Price Patterns")