added plot_is_static, colors, timeseries notebook

wise_pizza/explain.py

@@ -490,11 +490,16 @@ def explain_timeseries(
  assert almost_equals(fitted_sizes, sf2.weights)
 
  out = SlicerPair(sf1, sf2)
- out.plot = lambda width=600, height=1200, average_name=None, use_fitted_weights=False: plot_ts_pair(
+ out.relevant_cluster_names1 = sf1.relevant_cluster_names
+ out.relevant_cluster_names2 = sf2.relevant_cluster_names
+
+ out.plot = lambda plot_is_static=False, width=600, height=1200, return_fig=False, average_name=None, use_fitted_weights=False: plot_ts_pair(
  out.s1,
  out.s2,
+ plot_is_static=plot_is_static,
  width=width,
  height=height,
+ return_fig=return_fig,
  average_name=average_name,
  use_fitted_weights=use_fitted_weights,
  )
@@ -557,7 +562,7 @@ def _explain_timeseries(
  for c in chosen_cols:
  pre_basis[c + "_a"] = pre_basis["Slope"] - pre_basis[c]
 
- print("yay!")
+ # print("yay!")
 
  df, avg_df = add_average_over_time(
  df,
@@ -632,10 +637,10 @@ def _explain_timeseries(
  # sf.reg.intercept_ += average
  sf.plot = lambda plot_is_static=False, width=1200, height=2000, return_fig=False, average_name=None: plot_time(
  sf,
- # plot_is_static=plot_is_static,
+ plot_is_static=plot_is_static,
  width=width,
  height=height,
- # return_fig=return_fig,
+ return_fig=return_fig,
  average_name=average_name,
  )
  sf.task = "time"

wise_pizza/plotting_time.py

@@ -4,6 +4,9 @@
 
 import numpy as np
 import pandas as pd
+from IPython.display import Image, display
+import plotly.io as pio
+from plotly.io import to_image
 from plotly import graph_objects as go
 from plotly.subplots import make_subplots
 
@@ -25,6 +28,8 @@ def plot_time(
  width: int = 1000,
  height: int = 1000,
  average_name: Optional[str] = None,
+ plot_is_static: bool = False,
+ return_fig: bool = False,
 ):
  plot_data = preprocess_for_ts_plot(sf, average_name)
  num_rows = len(plot_data.nonflat_segments) + 1
@@ -48,7 +53,14 @@ def plot_time(
  width=width,
  height=height,
  )
- fig.show()
+ if plot_is_static:
+ image_bytes = to_image(fig, format="png", scale=2)
+ return Image(image_bytes, height=height, width=width)
+ else:
+ if return_fig:
+ return fig
+ else:
+ fig.show()
 
 
 def plot_ts_pair(
@@ -57,6 +69,8 @@ def plot_ts_pair(
  width,
  height,
  average_name: str = None,
+ plot_is_static: bool = False,
+ return_fig: bool = False,
  use_fitted_weights: bool = False,
 ):
  # if use_fitted_weights:
@@ -102,7 +116,15 @@ def plot_ts_pair(
  width=width,
  height=height,
  )
- fig.show()
+
+ if plot_is_static:
+ image_bytes = to_image(fig, format="png", scale=2)
+ return Image(image_bytes, height=height, width=width)
+ else:
+ if return_fig:
+ return fig
+ else:
+ fig.show()
 
 
 def plot_single_ts(
@@ -246,13 +268,19 @@ def preprocess_for_ts_plot(
  global_time_label = ""
 
  seg_names = ["All" + global_time_label] + [
- str(drop_time(s["segment"])) for s in nonflat_segments
+ drop_time(s["segment"]) for s in nonflat_segments
  ]
  sub_titles = [
  [
- f"{sf.size_name} for {s} ",
- f"{average_name} for {s}",
- f"{sf.total_name} for {s}",
+ f"{sf.size_name} for <br>" + f"{s}",
+ f"{average_name} for <br>" + f"{s}",
+ f"{sf.total_name} for <br>" + f"{s}",
+ ]
+ if s != "All" and s != global_time_label
+ else [
+ f"{sf.size_name} for " + "<br>".join([key for key in s]),
+ f"{average_name} for " + "<br>".join([key for key in s]),
+ f"{sf.total_name} for " + "<br>".join([key for key in s]),
  ]
  for s in seg_names
  ]
@@ -306,7 +334,7 @@ def simple_ts_plot(
  x=time,
  y=totals * mult,
  name=f"Actuals",
- marker=dict(color="orange"),
+ marker=dict(color="#ffc091"),
  showlegend=showlegend and col == col_nums[0],
  ),
  row=row_num,
@@ -319,7 +347,7 @@ def simple_ts_plot(
  y=reg_totals * mult,
  mode="lines",
  name=f"Regression",
- line=dict(color="blue"),
+ line=dict(color="#a0e1e1"),
  showlegend=showlegend and col == col_nums[0],
  ),
  row=row_num,
@@ -332,7 +360,7 @@ def simple_ts_plot(
  y=reg_seg * mult,
  mode="lines",
  name=f"Segment's reg contribution (Right axis)",
- line=dict(color="teal"),
+ line=dict(color="#9fe870"),
  showlegend=showlegend and col == col_nums[0],
  ),
  row=row_num,
@@ -345,7 +373,7 @@ def simple_ts_plot(
  x=time,
  y=leftover_totals if col == 1 else leftover_avgs,
  name=f"Leftover actuals",
- marker=dict(color="purple"),
+ marker=dict(color="#ff685f"),
  showlegend=showlegend and col == col_nums[0],
  ),
  row=row_num,

wise_pizza/slicer_facades.py

@@ -145,6 +145,8 @@ def __init__(
  # Now let's make sure single-segment impacts add up to total impact
  self.segment_mult = 1.0
 
+ self.relevant_cluster_names = sf.relevant_cluster_names
+
  # Try to make individual segment impacts add up to total regression post-transform
  # Didn't really make much difference
  # sum_marginals = 0