add plotting function Pt.6

Author: thierrymoudiki

ahead/Ridge2/Ridge2Regressor.py

@@ -1,6 +1,6 @@
 import numpy as np
 import rpy2.robjects.conversion as cv
-from rpy2.robjects import default_converter, FloatVector, numpy2ri, r
+from rpy2.robjects import default_converter, FloatVector, ListVector, numpy2ri, r
 from .. import config
 from ..utils import multivariate as mv
 from ..utils import unimultivariate as umv
@@ -139,7 +139,7 @@ def __init__(
         lambda_2 = 0.1,
         dropout = 0,
         type_pi = "gaussian",
-        block_length = 3, # can be NULL, but in R
+        block_length = 3, # can be NULL, but in R (use 0 in R instead of NULL for v0.7.0)
         B = 100,
         type_aggregation = "mean",
         centers = 2,
@@ -163,7 +163,7 @@ def __init__(
         self.block_length = block_length
         self.B = B
         self.type_aggregation = type_aggregation
-        self.centers = centers # can be NULL, but in R
+        self.centers = centers # can be NULL, but in R (use 0 in R instead of NULL for v0.7.0)
         self.type_clustering = type_clustering
         self.cl = cl
         self.date_formatting = date_formatting
@@ -224,10 +224,10 @@ def forecast(self, df, xreg = None):
                 lambda_2=self.lambda_2,
                 dropout=self.dropout,
                 type_pi=self.type_pi,
-                block_length=self.block_length, # can be NULL, but in R
+                block_length=self.block_length, # can be NULL, but in R (use 0 in R instead of NULL for v0.7.0)
                 B=self.B,
                 type_aggregation = self.type_aggregation,
-                centers = self.centers,  # can be NULL, but in R
+                centers = self.centers,  # can be NULL, but in R (use 0 in R instead of NULL for v0.7.0)
                 type_clustering = self.type_clustering,
                 cl=self.cl,
                 seed=self.seed,
@@ -242,14 +242,16 @@ def forecast(self, df, xreg = None):
 
             is_matrix_xreg = (len(self.xreg_.shape) > 1)
 
-            numpy2ri.activate()
+            numpy2ri.activate()            
+
+            xreg_ = r.matrix(FloatVector(self.xreg_.flatten()), 
+                        byrow = True, nrow = self.xreg_.shape[0], 
+                        ncol = self.xreg_.shape[1]) if is_matrix_xreg else r.matrix(FloatVector(self.xreg_.flatten()), 
+                        byrow=True, nrow=self.xreg_.shape[0], ncol=1)
 
             self.fcast_ = config.AHEAD_PACKAGE.ridge2f(
             y,
-            xreg = r.matrix(FloatVector(self.xreg_.flatten()), 
-                            byrow=True, nrow=self.xreg_.shape[0], 
-                            ncol=self.xreg_.shape[1]) if is_matrix_xreg else r.matrix(FloatVector(self.xreg_.flatten()), 
-                            byrow=True, nrow=self.xreg_.shape[0], ncol=1),  
+            xreg = xreg_,  
             h=self.h,
             level=self.level,
             lags=self.lags,
@@ -268,9 +270,7 @@ def forecast(self, df, xreg = None):
             type_clustering = self.type_clustering,
             cl=self.cl,
             seed=self.seed,
-        ) 
-            
-        print(f"self.fcast_.rx2['x']: {self.fcast_.rx2['x']}")
+        )                     
 
         # result -----
 

ahead/plot/plot.py

@@ -1,13 +1,7 @@
 from .. import config
 
 def plot(obj, selected_series = None):
-
-    print(f" \n obj.fcast_: {obj.fcast_} \n ")
-
-    if selected_series is not None:
-        try: 
-            config.PLOT_AHEAD(obj.fcast_, selected_series)
-        except:
-            pass
+    if selected_series is not None:        
+        return(config.PLOT_AHEAD(obj.fcast_, selected_series))
     else: 
-        config.PLOT_BASE(obj.fcast_)
+        return(config.PLOT_BASE(obj.fcast_))

ahead/utils/multivariate.py

@@ -5,7 +5,7 @@
 import rpy2.robjects as robjects
 import rpy2.robjects.packages as rpackages
 from rpy2.robjects.packages import importr
-from rpy2.robjects import FloatVector
+from rpy2.robjects import FloatVector, r
 from datetime import datetime
 from rpy2.robjects.vectors import StrVector
 from .unimultivariate import get_frequency
@@ -21,10 +21,14 @@ def compute_y_mts(df, df_frequency):
     input_series_tolist = input_series.tolist()
     xx = [item for sublist in input_series_tolist for item in sublist]
 
-    return stats.ts(
-        base.matrix(FloatVector(xx), byrow=True, nrow=len(input_series_tolist)),
-        frequency=get_frequency(df_frequency)
-    )
+    ts = r.matrix(FloatVector(xx), 
+                  byrow = True, 
+                  nrow = len(input_series_tolist), 
+                  ncol = df.shape[1])        
+    
+    ts.colnames = StrVector(df.columns.tolist())
+
+    return stats.ts(ts, frequency=get_frequency(df_frequency))
 
 
 def format_multivariate_forecast(

ahead/utils/univariate.py

@@ -17,11 +17,15 @@ def compute_y_ts(df, df_frequency):
 
     input_series = df.to_numpy()
 
-    return stats.ts(
+    ts = stats.ts(
         FloatVector(input_series.flatten()),
         frequency=get_frequency(df_frequency),
     )
 
+    ts.colnames = StrVector(df.columns.tolist())
+
+    return ts
+
 
 def format_univariate_forecast(date_formatting, output_dates, horizon, fcast):
 

examples/ridge2regressor.py

@@ -102,8 +102,15 @@
 d4 = Ridge2Regressor(h = h, date_formatting = "original", 
 type_pi="bootstrap", B=3, seed=1)
 
-xreg  = np.asarray(range(1, df.shape[0] + 1), 
-                   dtype='float64')
+# Data frame containing the time series 
+dataset = {
+ 'date' : ['2001-01-01', '2001-02-01', '2001-03-01', '2001-04-01', '2001-05-01'],
+ 'series1' : [34, 30, 35.6, 33.3, 38.1],    
+ 'series2' : [4, 5.5, 5.6, 6.3, 5.1],
+ 'series3' : [100, 100.5, 100.6, 100.2, 100.1]}
+df = pd.DataFrame(dataset).set_index('date')
+
+xreg = np.asarray(range(1, df.shape[0] + 1), dtype='float64')
 
 print(f"\n xreg: {xreg} \n")
 
@@ -134,38 +141,53 @@
 print(d4.lower_)
 print(d4.upper_)
 print("\n")
-plot(d4, selected_series="y1")
+plot(d4, selected_series="series1")
 print("\n")
 
-print("Example 5 -----")
+# print("Example 5 -----")
 
-d5 = Ridge2Regressor(h = h, date_formatting = "original", 
-type_pi="blockbootstrap", B=5)
+# d5 = Ridge2Regressor(h = h, date_formatting = "original", 
+# type_pi="bootstrap", B=3, seed=1)
 
-xreg  = np.column_stack((range(df.shape[0]), 
-                        [0.2, 0.5, 0.4, 0.3, 0.1]))
+# # Data frame containing the time series 
+# dataset = {
+#  'date' : ['2001-01-01', '2001-02-01', '2001-03-01', '2001-04-01', '2001-05-01'],
+#  'series1' : [34, 30, 35.6, 33.3, 38.1],    
+#  'series2' : [4, 5.5, 5.6, 6.3, 5.1],
+#  'series3' : [100, 100.5, 100.6, 100.2, 100.1]}
+# df = pd.DataFrame(dataset).set_index('date')
 
-start = time()
-d5.forecast(df, xreg = xreg)
-print(f"Elapsed: {time()-start} \n")
+# xreg = np.asarray(range(1, df.shape[0] + 1), dtype='float64')
 
-print("\n mean ---------- \n")
-print(d5.fcast_.rx2['mean']) # R output
-print(d5.mean_) # Python output
+# print(f"\n xreg: {xreg} \n")
 
-print("\n lower ---------- \n")
-print(d5.fcast_.rx2['lower']) # R output
-print(d5.lower_) # Python output
+# start = time()
+# d5.forecast(df, xreg = xreg)
+# print(f"Elapsed: {time()-start} \n")
 
-print("\n upper ---------- \n")
-print(d5.fcast_.rx2['upper']) # R output
-print(d5.upper_) # Python output
+# print(d5.fcast_.rx2['mean'])
+# print(d5.averages_[1])
+# print(np.asarray(d5.fcast_.rx2['mean']))
 
-print("\n d5.averages_ ---------- \n")
-print(d5.averages_)
+# print(d5.fcast_.rx2['sims'][0])
+# res = np.asarray(d5.fcast_.rx2['sims'][1])
+# print(res)
+# print(res.shape)
 
-print("\n d5.ranges_ ---------- \n")
-print(d5.ranges_)
+# print("\n result_dfs_: \n")
+# print(d5.result_dfs_)
+
+# print("\n sims_: \n")
+# print(d5.sims_)
+
+# print("\n output_dates_: \n")
+# print(d5.output_dates_)
+
+# print("\n mean, lower, upper as numpy arrays: \n")
+# print(d5.mean_)
+# print(d5.lower_)
+# print(d5.upper_)
+# print("\n")
+# plot(d5, selected_series="series1")
+# print("\n")
 
-print("\n")
-plot(d5, selected_series="y1")