adviz - npm Package Compare versions

Comparing version 0.0.18 to 0.0.19

adviz/value_counts.py

+# AUTOGENERATED! DO NOT EDIT! File to edit: ../nbs/07_value_counts.ipynb.
+
+# %% auto 0
+__all__ = ['value_counts']
+
+# %% ../nbs/07_value_counts.ipynb 2
+import advertools as adv
+import pandas as pd
+from plotly.subplots import make_subplots
+
+# %% ../nbs/07_value_counts.ipynb 3
+def value_counts(
+    data,
+    dropna=False,
+    show_top=10,
+    sort_others=False,
+    title=None,
+    style=True,
+    width=900,
+    height=650,
+    colorscale='cividis',
+):
+    """
+    Count the values of `data` and return a table of counts (absolute, cumulative, percentage, and cumulative percentage).
+
+    Parameters
+    ----------
+    data : list, tuple, pandas.Series, pandas.DataFrame
+        A collection of items to count, using any of the above data structures.
+    dropna : bool
+        Wether or not to drop missing values.
+    show_top : int
+        How many top items to show. All remaining items, will be grouped into "Others:".
+    sort_others : bool
+        Whether or not to put "Others" in their sorted order. The default is to have this item at the bottom.
+    title : str
+        The title of the chart.
+    style : bool
+        Whether or not to style the resulting table with a heatmap.
+    width : int
+        The width in pixels of the resulting figure. Set this to None to make it use the widthe of its container.
+    height : int
+        The width in pixels of the resulting figure. Set this to None to make it use the widthe of its container.
+    colorscale : str
+        Which color scale to use for the heatmaps.
+
+    Returns
+    -------
+    plotly.graph_objects.Figure
+
+    """
+    if isinstance(data, (list, tuple)):
+        data = pd.Series(data)
+    # number of columns counted in cases it was a DataFrame:
+    num_dims = 1 if len(data.shape) == 1 else data.shape[-1]
+    val_counts = data.value_counts(dropna=dropna).reset_index()
+    if len(val_counts) > show_top:
+        others_df = pd.DataFrame(
+            [['Others:'] + ['' for i in range(len(val_counts.columns)-2)] + [val_counts[show_top:]['count'].sum()]],
+            columns=val_counts.columns)
+        val_counts = pd.concat([
+            val_counts[:show_top],
+            others_df
+        ])
+        if sort_others:
+            val_counts = val_counts.sort_values(by=['count'], ascending=False)
+
+    count_df = (val_counts
+                .assign(
+                    cum_count=lambda df: df['count'].cumsum(),
+                    perc=lambda df: df['count'].div(df['count'].sum()),
+                    cum_perc=lambda df: df['perc'].cumsum())
+                )
+    if not style:
+        return count_df
+
+    count_df.insert(0, 'rank', range(1, len(count_df)+1))
+    num_columns = count_df.shape[1]
+    count_df = count_df.rename(columns={'cum_count': 'cum. count', 'perc': '%', 'cum_perc': 'cum. %'})
+    subplot_titles = count_df.columns
+    fig = make_subplots(
+        cols=num_columns, rows=1,
+        column_widths=[0.05] + [0.65/(num_dims) for d in range(num_dims)] + [.15, .15, .15, .15],
+        horizontal_spacing=0,
+        subplot_titles=subplot_titles
+    )
+    for i in range(count_df.shape[1]):
+        tempdf = count_df.iloc[:, i].to_frame()
+        fig.add_heatmap(
+            z=[[100] for i in range(len(tempdf))] if i in range(num_dims+1) else tempdf,
+            col=i+1,
+            row=1,
+            hoverinfo='skip',
+            showscale=False,
+            name=subplot_titles[i],
+            colorscale='RdBu' if i in range(num_dims+1) else colorscale,
+            texttemplate="%{text}" if i in range(num_dims+1) else "%{text:,.0f}" if subplot_titles[i].endswith('count') else "%{text:.1%}",
+            textfont={'size': 16},
+            text=tempdf)
+    fig.layout.width = width
+    fig.layout.height = height
+    fig.update_xaxes(showticklabels=False, showgrid=False, zeroline=False)
+    fig.update_yaxes(showticklabels=False, showgrid=False, zeroline=False, autorange='reversed')
+    fig.layout.title = title
+    return fig

adviz.egg-info/PKG-INFO

 Metadata-Version: 2.1
 Name: adviz
-Version: 0.0.18
+Version: 0.0.19
 Summary: advertools visualizations
@@ -5,0 +5,0 @@ Home-page: https://github.com/eliasdabbas/adviz

adviz.egg-info/SOURCES.txt

@@ -15,2 +15,3 @@ LICENSE
 adviz/url_structure.py
+adviz/value_counts.py
 adviz/value_counts_plus.py
@@ -17,0 +18,0 @@ adviz.egg-info/PKG-INFO

adviz/__init__.py

@@ -1,4 +0,5 @@
-__version__ = "0.0.18"
+__version__ = "0.0.19"
 
 from .value_counts_plus import value_counts_plus
+from .value_counts import value_counts
 from .status_codes import status_codes
@@ -5,0 +6,0 @@ from .url_structure import url_structure

adviz/_modidx.py

@@ -17,3 +17,4 @@ # Autogenerated by nbdev
             'adviz.url_structure': {'adviz.url_structure.url_structure': ('url_structure.html#url_structure', 'adviz/url_structure.py')},
+            'adviz.value_counts': {'adviz.value_counts.value_counts': ('value_counts.html#value_counts', 'adviz/value_counts.py')},
             'adviz.value_counts_plus': { 'adviz.value_counts_plus.value_counts_plus': ( 'value_counts_plus.html#value_counts_plus',
                                                                                         'adviz/value_counts_plus.py')}}}

adviz/value_counts_plus.py

@@ -6,3 +6,3 @@ # AUTOGENERATED! DO NOT EDIT! File to edit: ../nbs/00_value_counts_plus.ipynb.
 
-# %% ../nbs/00_value_counts_plus.ipynb 3
+# %% ../nbs/00_value_counts_plus.ipynb 4
 import advertools as adv
@@ -12,3 +12,3 @@ import pandas as pd
 
-# %% ../nbs/00_value_counts_plus.ipynb 4
+# %% ../nbs/00_value_counts_plus.ipynb 5
 def value_counts_plus(
@@ -15,0 +15,0 @@ data,

PKG-INFO

 Metadata-Version: 2.1
 Name: adviz
-Version: 0.0.18
+Version: 0.0.19
 Summary: advertools visualizations
@@ -5,0 +5,0 @@ Home-page: https://github.com/eliasdabbas/adviz

settings.ini

 [DEFAULT]
 repo = adviz
 lib_name = adviz
-version = 0.0.18
+version = 0.0.19
 min_python = 3.7
@@ -6,0 +6,0 @@ license = apache2

Alert delta unavailable

Currently unable to show alert delta for PyPI packages.

Improved metrics

Total package byte prevSize

66397

6.63%

Number of package files

25

4.17%

Lines of code

1521

6.96%