lessPython 0.2.0__tar.gz → 0.3.1__tar.gz

{lesspython-0.2.0/src/lessPython.egg-info → lesspython-0.3.1}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: lessPython
-Version: 0.2.0
+Version: 0.3.1
 Summary: Python port of R's lessR: analytic-view visualizations, statistical models, and data utilities, rendered with plotly
 Author-email: "David W. Gerbing" <gerbing@pdx.edu>
 License-Expression: GPL-2.0-or-later
@@ -113,7 +113,18 @@ of functions, grouped by category for navigation:
 
 The inferential functions print their analysis as in lessR and
 also return a results object whose numeric fields and plotly
-figures (in `.plots`) are available for programmatic use.
+figures (in `.plots`) are available for programmatic use. In a
+notebook, such as Jupyter or Colab, the figures also display
+when the analysis runs, even when the result is assigned, as
+`r = lp.Regression(...)`. From a script, show one with
+`r.plots["scatter"].show()`.
+
+Every function that draws also takes `pdf_file=` and `png_file=`
+(needs the png extra, `pip install "lessPython[png]"`, then
+`plotly_get_chrome -y` once). A name such as `png_file="reg"`
+names the file, or with several figures is the stem of each
+name (`reg_scatter_matrix.png`), as lessR's `pdf_file=`. On Colab
+the files also download to your computer.
 
 ## Development install
 

{lesspython-0.2.0 → lesspython-0.3.1}/README.md

@@ -76,7 +76,18 @@ of functions, grouped by category for navigation:
 
 The inferential functions print their analysis as in lessR and
 also return a results object whose numeric fields and plotly
-figures (in `.plots`) are available for programmatic use.
+figures (in `.plots`) are available for programmatic use. In a
+notebook, such as Jupyter or Colab, the figures also display
+when the analysis runs, even when the result is assigned, as
+`r = lp.Regression(...)`. From a script, show one with
+`r.plots["scatter"].show()`.
+
+Every function that draws also takes `pdf_file=` and `png_file=`
+(needs the png extra, `pip install "lessPython[png]"`, then
+`plotly_get_chrome -y` once). A name such as `png_file="reg"`
+names the file, or with several figures is the stem of each
+name (`reg_scatter_matrix.png`), as lessR's `pdf_file=`. On Colab
+the files also download to your computer.
 
 ## Development install
 

{lesspython-0.2.0 → lesspython-0.3.1}/pyproject.toml

@@ -12,7 +12,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "lessPython"
-version = "0.2.0"
+version = "0.3.1"
 description = "Python port of R's lessR: analytic-view visualizations, statistical models, and data utilities, rendered with plotly"
 authors = [{ name = "David W. Gerbing", email = "gerbing@pdx.edu" }]
 readme = "README.md"

{lesspython-0.2.0 → lesspython-0.3.1}/src/lessPy/ANOVA.py

@@ -22,10 +22,10 @@ import pandas as pd
 import plotly.graph_objects as go
 
 from .plotly_utils import (
-    BASE_COLORS, axis_cat, axis_format, axis_num, make_trans,
+    as_plotly_color, BASE_COLORS, axis_cat, axis_format, axis_num, make_trans,
     plot_border, plotly_style, to_hex, x_grid)
 from .Regression import _getdigits, _prntbl
-from .utils import fmt, get_column, get_option, pretty
+from .utils import fmt, get_column, get_option, pretty, file_output, show_plots
 
 
 class ANOVAResults:
@@ -380,18 +380,73 @@ def _oneway(fit, yv, gv, y_name, x_name, levels, n_keep, d,
         lines += _prntbl(tukey, d).split("\n")
 
     plots = {}
+    if graphics and tukey is not None:
+        plots["tukey"] = _tukey_plot(tukey, x_name, d)
     if graphics:
         plots["means"] = _anova_means_plot(
             yv, gv, y_name, x_name, levels, jitter_x, d)
 
-    return ANOVAResults(
+    return show_plots(ANOVAResults(
         formula=formula, design="oneway", n_obs=n_obs,
         n_keep=n_keep, digits_d=d, response=y_name,
         factors=[x_name], descriptive=desc, grand_mean=grand,
         anova=tbl,
         effects={"R_squared": rsq, "R_sq_adjusted": rsq_adj,
                  "omega_squared": omsq, "cohen_f": cohen_f},
-        tukey=tukey, plots=plots)
+        tukey=tukey, plots=plots))
+
+
+def _tukey_plot(tukey, x_name, d):
+    """Each pairwise difference of means with its 95% family-wise
+    interval, the first comparison at the top, and a dashed
+    reference line at 0. ~ .ANOVAz1 plot(TukeyHSD())"""
+    style = plotly_style()
+    k = len(tukey)
+    ypos = list(range(k, 0, -1))
+    col = as_plotly_color(get_option("pt_color", "#324E5C"))
+    fig = go.Figure()
+    for yp, (lbl, r) in zip(ypos, tukey.iterrows()):
+        fig.add_trace(go.Scatter(
+            x=[r["lwr"], r["upr"]], y=[yp, yp], mode="lines",
+            line=dict(color=col, width=2), hoverinfo="skip",
+            showlegend=False))
+    fig.add_trace(go.Scatter(
+        x=tukey["diff"], y=ypos, mode="markers",
+        marker=dict(symbol="circle", size=9, color=col),
+        customdata=np.column_stack([tukey["lwr"], tukey["upr"],
+                                    tukey["p adj"]]),
+        text=list(tukey.index),
+        hovertemplate=("%{text}<br>diff: %{x:." + str(d) + "f}"
+                       "<br>95% CI: %{customdata[0]:." + str(d)
+                       + "f} to %{customdata[1]:." + str(d) + "f}"
+                       "<br>p adj: %{customdata[2]:.3f}"
+                       "<extra></extra>"),
+        showlegend=False))
+    axT1 = pretty(min(0.0, float(tukey["lwr"].min())),
+                  max(0.0, float(tukey["upr"].max())))
+    ax_x = axis_num(f"Differences in mean levels of {x_name}", axT1,
+                    axis_format(axT1, d))
+    ax_x.update(showgrid=True,
+                gridcolor=as_plotly_color(style["grid_col"]),
+                gridwidth=1, griddash="dot",
+                range=[axT1[0], axT1[-1]], zeroline=False)
+    ax_y = axis_cat("")
+    ax_y.update(tickmode="array", tickvals=ypos,
+                ticktext=list(tukey.index), range=[0.5, k + 0.5])
+    zero = dict(type="line", xref="x", yref="paper", x0=0, x1=0,
+                y0=0, y1=1, layer="below",
+                line=dict(color=as_plotly_color("gray50"), width=1,
+                          dash="dash"))
+    fig.update_layout(
+        xaxis=ax_x, yaxis=ax_y, shapes=[zero] + list(plot_border()),
+        template=None,
+        plot_bgcolor=as_plotly_color(style["panel_fill"]),
+        paper_bgcolor=as_plotly_color(style["window_fill"]),
+        title=dict(text="95% family-wise confidence level",
+                   x=0.5, xanchor="center",
+                   font=dict(size=round(
+                       16 * get_option("main_size", 1)))))
+    return fig
 
 
 def _anova_means_plot(yv, gv, y_name, x_name, levels, jitter_x,
@@ -410,18 +465,18 @@ def _anova_means_plot(yv, gv, y_name, x_name, levels, jitter_x,
         x=xn, y=yv, mode="markers",
         marker=dict(symbol="circle", size=6,
                     color=make_trans(pt, 0.85), opacity=1,
-                    line=dict(color=to_hex(pt), width=1)),
+                    line=dict(color=as_plotly_color(pt), width=1)),
         hoverinfo="y", showlegend=False))
     means = [yv[gv == lv].mean() for lv in levels]
     for i, mval in enumerate(means):
         fig.add_shape(type="line", xref="x", yref="y",
                       x0=-0.5, x1=len(levels) - 0.5,
                       y0=mval, y1=mval, layer="below",
-                      line=dict(color=to_hex("gray70"), width=1))
+                      line=dict(color=as_plotly_color("gray70"), width=1))
     fig.add_trace(go.Scatter(
         x=list(range(len(levels))), y=means, mode="markers",
         marker=dict(symbol="diamond", size=13,
-                    color=to_hex(get_option("fit_color",
+                    color=as_plotly_color(get_option("fit_color",
                                             "#5C4032"))),
         hoverinfo="x+y", showlegend=False))
     axT2 = pretty(float(yv.min()), float(yv.max()))
@@ -431,13 +486,13 @@ def _anova_means_plot(yv, gv, y_name, x_name, levels, jitter_x,
                 ticktext=levels, range=[-0.5, len(levels) - 0.5])
     ax_y = axis_num(y_name, axT2, axis_format(axT2, d))
     ax_y.update(showgrid=True,
-                gridcolor=to_hex(style["grid_col"]),
+                gridcolor=as_plotly_color(style["grid_col"]),
                 gridwidth=1, griddash="dot")
     fig.update_layout(
         xaxis=ax_x, yaxis=ax_y, shapes=fig.layout.shapes
         + tuple(plot_border()), template=None,
-        plot_bgcolor=to_hex(style["panel_fill"]),
-        paper_bgcolor=to_hex(style["window_fill"]),
+        plot_bgcolor=as_plotly_color(style["panel_fill"]),
+        paper_bgcolor=as_plotly_color(style["window_fill"]),
         title=dict(text="Scatterplot with Cell Means",
                    x=0.5, xanchor="center",
                    font=dict(size=round(
@@ -611,12 +666,12 @@ def _twoway(fit, yv, fvals, y_name, facs, flevs, n_keep, d,
                 x1v, np.asarray(fit.fittedvalues), x2v,
                 "Fitted", f1, f2, l1, l2, d, "Fitted Values")
 
-    return ANOVAResults(
+    return show_plots(ANOVAResults(
         formula=formula, design=design, n_obs=n_obs,
         n_keep=n_keep, digits_d=d, response=y_name,
         factors=[f1, f2], cell_n=cell_n,
         marginal_means={f1: m1, f2: m2}, grand_mean=grand,
-        anova=t, effects=eff, tukey=tukey, plots=plots)
+        anova=t, effects=eff, tukey=tukey, plots=plots))
 
 
 def _cat_line_fig(title, x_name, y_name, levels_x, series, d,
@@ -644,13 +699,13 @@ def _cat_line_fig(title, x_name, y_name, levels_x, series, d,
                 range=[-0.4, len(levels_x) - 0.6])
     ax_y = axis_num(y_name, axT2, axis_format(axT2, d))
     ax_y.update(showgrid=True,
-                gridcolor=to_hex(style["grid_col"]),
+                gridcolor=as_plotly_color(style["grid_col"]),
                 gridwidth=1, griddash="dot")
     fig.update_layout(
         xaxis=ax_x, yaxis=ax_y, shapes=plot_border(),
         template=None, legend=dict(title=dict(text=legend_title)),
-        plot_bgcolor=to_hex(style["panel_fill"]),
-        paper_bgcolor=to_hex(style["window_fill"]),
+        plot_bgcolor=as_plotly_color(style["panel_fill"]),
+        paper_bgcolor=as_plotly_color(style["window_fill"]),
         title=dict(text=title, x=0.5, xanchor="center",
                    font=dict(size=round(
                        16 * get_option("main_size", 1)))))
@@ -678,3 +733,7 @@ def _blocks_plot(x1v, yvals, x2v, y_name, f1, f2, l1, l2, d,
         sub = dp[dp["b"] == b].set_index("a")["y"]
         series.append((b, [sub.get(a, np.nan) for a in l1]))
     return _cat_line_fig(title, f1, y_name, l1, series, d, f2)
+
+
+# pdf_file= and png_file= write the figures, as lessR's pdf_file=
+ANOVA = file_output(ANOVA)