Update plot_config to plotting_style and documentation (#239)

* update export tutorial to add explanation for standalone argument * minor fixes and remove cell output in notebooks * added contributing doc * fix bugs and uncomment some tests * remove raise warning * remove unnecessary import * split up rename test into two parts * fix setting warning, fix data_type bugs and add relevant tests * remove ordinal data type * add test for small dataframe resetting index * add loc and iloc tests * fix attribute access directly to dataframe * add small changes to code * added test for qcut and cut * add check if dtype is Interval * added qcut test * fix Record KeyError * add tests * take care of reset_index case * small edits * add data_model to column_group Clause * small edits for row_group * fixes to row group * basic matplotlib chart example * add config for start and cap for samples * finish sampling config and tests * black formatting * add documentation for sampling config * remove small added issues * minor changes to docs * migrate register default action to init * config class * oop png * implement heatmap flag and add tests * black formatting and documentation edits * all except hist * hist works * heatmap * only axis * added code * heatmap works * heatmap works * bar chart lim * everything works * add global var * ui refine * adjust comments * black reformat * add pd.io equalities for DataFrames * black reformat * matplotlib tests * reformat and req * add one default test * black format * removed .coverage; fixed typos; add nan import; add fig display statement * doc and font size * remove ipy * color and check code not yet * remove print * all outputs * update docs * color channels work * add titles * rm * all exports work * add tests * increase size * leg size * reformat * abbrev axis * remove bin * no warning * improve ui * init for runtime * merge * plot config matplotlib * add documentation * close fig to save mem * small change * mem * split tests * reformat * font * ui changes * make scatter narrow * matplotlib abbrev * matplotlib abbrev * change to plotting_style * update docs * revert * revert merged changes * minor change * fix doc * update doc * add fig * rem * minor fixes to docs Co-authored-by: Kunal Agarwal <kagarwal2@berkeley.edu> Co-authored-by: Doris Lee <dorisjunglinlee@gmail.com> Co-authored-by: Kunal Agarwal <32151899+westernguy2@users.noreply.github.com> Co-authored-by: Caitlyn Chen <caitlynachen@berkeley.edu> Co-authored-by: Ujjaini Mukhopadhyay <ujjaini@berkeley.edu>
lux-org · Jan 25, 2021 · 5e7d5c9 · 5e7d5c9
1 parent d4fbae3
commit 5e7d5c9
Show file tree

Hide file tree

Showing 9 changed files with 59 additions and 51 deletions.
diff --git a/doc/source/guide/FAQ.rst b/doc/source/guide/FAQ.rst
@@ -52,7 +52,7 @@ How do I change the plotting library used for visualization?
   
 I want to change the opacity of my chart, add title, change chart font size, etc. How do I modify chart settings?
 """""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""
-  To add custom plot settings to the recommendations, you can set the :code:`lux.config.plot_config` property. See `this tutorial <https://lux-api.readthedocs.io/en/latest/source/guide/style.html>`__ on how to configure chart properties. Lux currently only support chart modifications in Altair.
+  To add custom plot settings to the recommendations, you can set the :code:`lux.config.plotting_style` property. See `this tutorial <https://lux-api.readthedocs.io/en/latest/source/guide/style.html>`__ on how to configure chart properties. Lux supports chart modifications in Altair and Matplotlib.
 
 How do I change aggregation functions, binning, or axis channels to non-default values?
 """""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""

diff --git a/doc/source/guide/style.rst b/doc/source/guide/style.rst
@@ -4,7 +4,7 @@ Styling Custom Plot Settings
 
 .. note:: You can follow along this tutorial in a Jupyter notebook. [`Github <https://github.com/lux-org/lux-binder/blob/master/tutorial/4-chart-settings.ipynb>`_] [`Binder <https://mybinder.org/v2/gh/lux-org/lux-binder/master?urlpath=tree/tutorial/4-chart-settings.ipynb>`_]
 
-In the last tutorial, we saw how :code:`Vis` objects could be exported into visualization code for further editing. What if we want to change the chart settings for *all* the visualizations displayed in the widget. In Lux, we can change the chart settings and aesthetics by inputting global custom plot settings the :code:`plot_config`.
+In the last tutorial, we saw how :code:`Vis` objects could be exported into visualization code for further editing. What if we want to change the chart settings for *all* the visualizations displayed in the widget. In Lux, we can change the chart settings and aesthetics by inputting global custom plot settings the :code:`plotting_style`.
 
 Example #1 : Changing Color and Title of all charts
 ---------------------------------------------------
@@ -25,18 +25,18 @@ To change the plot configuration in Altair, we need to specify a function that t
 Let's say that we want to change all the graphical marks of the charts to green and add a custom title. We can define this `change_color_add_title` function, which configures the chart's mark as green and adds a custom title to the chart.
 
 .. code-block:: python
-	lux.config.plotting_backend = "altair" # or 'vegalite'
 
-	def change_color_add_title(chart):
-	    chart = chart.configure_mark(color="green") # change mark color to green
-	    chart.title = "Custom Title" # add title to chart
-	    return chart
+    lux.config.plotting_backend = "altair"
+    def change_color_add_title(chart):
+        chart = chart.configure_mark(color="green") # change mark color to green
+        chart.title = "Custom Title" # add title to chart
+        return chart
 
-We then set the global plot configuration of the dataframe by changing the :code:`plot_config` property. With the added plot_config, Lux runs this user-defined function after every `Vis` is rendered to a chart, allow the user-defined function to override any existing default chart settings.
+We then set the global plot configuration of the dataframe by changing the :code:`plotting_style` property. With the added plotting_style, Lux runs this user-defined function after every `Vis` is rendered to a chart, allow the user-defined function to override any existing default chart settings.
 
 .. code-block:: python
 	
-	lux.config.plot_config = change_color_add_title
+	lux.config.plotting_style = change_color_add_title
 
 We now see that the displayed visualizations adopt these new imported settings.
 
@@ -45,18 +45,30 @@ We now see that the displayed visualizations adopt these new imported settings.
   :align: center 
 
 Similarly, we can change the plot configurations for Matplotlib charts as well.
-The plot_config attribute for Matplotlib charts takes in both the figure and axis as parameters.
+The plotting_style attribute for Matplotlib charts takes in both the `fig` and `ax` as parameters.
+`fig` handles figure width and other plot size attributes. `ax` supports changing the chart title and other plot labels and configurations.
+
 .. code-block:: python
 	
-	lux.config.plotting_backend = "matplotlib" # or 'matplotlib_code'
-
-    def add_title(fig, ax):
-        ax.set_title("Test Title")
+    lux.config.plotting_backend = "matplotlib"
+    def change_width_add_title(fig, ax):
+        fig.set_figwidth(7)
+        ax.set_title("Custom Title")
         return fig, ax
 
 .. code-block:: python
 	
-	lux.config.plot_config = add_title
+	lux.config.plotting_style = change_width_add_title
+
+Moreover, we can set the color and other figure styles using the rcParams attribute of pyplot.
+
+.. code-block:: python
+
+	import matplotlib.pyplot as plt
+	import matplotlib
+
+	plt.rcParams['axes.prop_cycle'] = matplotlib.cycler(color='g')
+
 
 We now see that the displayed visualizations adopt these new imported settings.
 
@@ -129,7 +141,7 @@ We want to decrease the opacity of scatterplots, but keep the opacity for the ot
 
 .. code-block:: python
 	
-	lux.config.plot_config = changeOpacityScatterOnly
+	lux.config.plotting_style = changeOpacityScatterOnly
 	df
 
 .. image:: ../img/style-6.png

diff --git a/doc/source/reference/config.rst b/doc/source/reference/config.rst
@@ -26,8 +26,8 @@ In Lux, users can customize various global settings to configure the behavior of
             df # recommendations generated for the first time with config
 
 
-Change the default display of Lux
-~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+Change the default display 
+~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
 We can set the :code:`default_display` to change whether the Pandas table or Lux widget is displayed by default. In the following block, we set the default display to 'lux', therefore the Lux widget will display first.
 
@@ -101,8 +101,8 @@ Lux currently only support Vega-Lite and matplotlib, and we plan to add support
   :width: 700
   :align: center
 
-Change the sampling parameters of Lux
-~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+Change the sampling parameters 
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
 To speed up the visualization processing, by default, Lux performs random sampling on datasets with more than 10000 rows. For datasets over 30000 rows, Lux will randomly sample 30000 rows from the dataset.
 
@@ -130,10 +130,11 @@ We can disable this feature and revert back to using a scatter plot by running t
 
     lux.config.heatmap = False
 
-Changing the plot styling
-~~~~~~~~~~~~~~~~~~~~~~~~~~
+Changing the plotting style
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
-Altair supports plot configurations to be applied on top of the generated graphs. To set a default plot configuration, first write a function that can take in a `chart` and returns a `chart`. For example:
+In Lux, we can change the chart settings and aesthetics by inputting global custom plot settings the :code:`plotting_style`.
+For charts rendered in Altair (default), we can change the plotting style by writing a function that takes a `AltairChart <https://altair-viz.github.io/user_guide/generated/toplevel/altair.Chart.html>`_ object as input and output. For example:
 
 .. code-block:: python
 
@@ -142,37 +143,29 @@ Altair supports plot configurations to be applied on top of the generated graphs
         chart.title = "Custom Title" # add title to chart
         return chart
 
-Then, set the `plot_config` to this function so that this function is applied to every plot generated.
+Then, set the `plotting_style` to this function so that this function is applied to every plot generated.
 
 .. code-block:: python
 
-    lux.config.plot_config = change_color_add_title
+    lux.config.plotting_style = change_color_add_title
 
 The above results in the following changes:
 
 .. image:: https://github.com/lux-org/lux-resources/blob/master/doc_img/style-2.png?raw=true
   :width: 600
   :align: center
 
-See `this page <https://lux-api.readthedocs.io/en/latest/source/guide/style.html>`__ for more details.
-
-Matplotlib also supports plot configurations to be applied on top of the generated graphs. To set a default plot configuration, first write a function that can take in a `fig` and 'ax' and returns a `fig` and 'ax. For example:
+Matplotlib also supports plot configurations to be applied on top of the generated graphs. To set a default plot configuration, first write a function that can take in a `fig` and `ax` and returns a `fig` and `ax`.
+`fig` handles figure width and other plot size attributes. `ax` supports changing the chart title and other plot labels and configurations. For example:
 
 .. code-block:: python
 
-    def add_title(fig, ax):
-        ax.set_title("Test Title")
+    def change_width_add_title(fig, ax):
+        fig.set_figwidth(7) # change figure width
+        ax.set_title("Custom Title") # add title to chart
         return fig, ax
 
-.. code-block:: python
-
-    lux.config.plot_config = add_title
-
-The above results in the following changes:
-
-.. image:: https://github.com/lux-org/lux-resources/blob/master/doc_img/style-7.png?raw=true
-  :width: 600
-  :align: center
+    lux.config.plotting_style = change_width_add_title
 
 See `this page <https://lux-api.readthedocs.io/en/latest/source/guide/style.html>`__ for more details.
 
@@ -207,5 +200,4 @@ If you would like to turn off the selection criteria completely and display ever
 
     lux.config.topk = False
 
-Beware that this may generate large numbers of visualizations (e.g., for 10 quantitative variables, this will generate 45 scatterplots in the Correlation action!)
-
+Beware that this may generate large numbers of visualizations (e.g., for 10 quantitative variables, this will generate 45 scatterplots in the Correlation action!)
diff --git a/doc/source/reference/gen/lux._config.config.Config.rst b/doc/source/reference/gen/lux._config.config.Config.rst
@@ -29,6 +29,8 @@ lux.\_config.config.Config
 
       ~Config.default_display
       ~Config.heatmap
+      ~Config.interestingness_fallback
+      ~Config.pandas_fallback
       ~Config.plotting_backend
       ~Config.sampling
       ~Config.sampling_cap

diff --git a/doc/source/reference/gen/lux.core.frame.LuxDataFrame.rst b/doc/source/reference/gen/lux.core.frame.LuxDataFrame.rst
@@ -176,6 +176,7 @@ lux.core.frame.LuxDataFrame
       ~LuxDataFrame.sem
       ~LuxDataFrame.set_SQL_table
       ~LuxDataFrame.set_axis
+      ~LuxDataFrame.set_data_type
       ~LuxDataFrame.set_index
       ~LuxDataFrame.set_intent
       ~LuxDataFrame.set_intent_as_vis
@@ -246,6 +247,7 @@ lux.core.frame.LuxDataFrame
       ~LuxDataFrame.axes
       ~LuxDataFrame.columns
       ~LuxDataFrame.current_vis
+      ~LuxDataFrame.data_type
       ~LuxDataFrame.dtypes
       ~LuxDataFrame.empty
       ~LuxDataFrame.exported

diff --git a/lux/_config/config.py b/lux/_config/config.py
@@ -17,7 +17,7 @@ class Config:
 
     def __init__(self):
         self._default_display = "pandas"
-        self.plot_config = None
+        self.plotting_style = None
         self.SQLconnection = ""
         self.executor = None
         # holds registered option metadata

diff --git a/lux/core/series.py b/lux/core/series.py
@@ -32,7 +32,7 @@ class LuxSeries(pd.Series):
         "_rec_info",
         "_pandas_only",
         "_min_max",
-        "plot_config",
+        "plotting_style",
         "_current_vis",
         "_widget",
         "_recommendation",

diff --git a/lux/vislib/altair/AltairRenderer.py b/lux/vislib/altair/AltairRenderer.py
@@ -86,10 +86,10 @@ def create_vis(self, vis, standalone=True):
             chart = None
 
         if chart:
-            if lux.config.plot_config and (
+            if lux.config.plotting_style and (
                 lux.config.plotting_backend == "vegalite" or lux.config.plotting_backend == "altair"
             ):
-                chart.chart = lux.config.plot_config(chart.chart)
+                chart.chart = lux.config.plotting_style(chart.chart)
             if self.output_type == "VegaLite":
                 chart_dict = chart.chart.to_dict()
                 # this is a bit of a work around because altair must take a pandas dataframe and we can only generate a luxDataFrame
@@ -101,9 +101,9 @@ def create_vis(self, vis, standalone=True):
             elif self.output_type == "Altair":
                 import inspect
 
-                if lux.config.plot_config:
+                if lux.config.plotting_style:
                     chart.code += "\n".join(
-                        inspect.getsource(lux.config.plot_config).split("\n    ")[1:-1]
+                        inspect.getsource(lux.config.plotting_style).split("\n    ")[1:-1]
                     )
                 chart.code += "\nchart"
                 chart.code = chart.code.replace("\n\t\t", "\n")

diff --git a/lux/vislib/matplotlib/MatplotlibRenderer.py b/lux/vislib/matplotlib/MatplotlibRenderer.py
@@ -86,11 +86,11 @@ def create_vis(self, vis, standalone=True):
             return chart
         if chart:
             plt.tight_layout()
-            if lux.config.plot_config and (
+            if lux.config.plotting_style and (
                 lux.config.plotting_backend == "matplotlib"
                 or lux.config.plotting_backend == "matplotlib_code"
             ):
-                chart.fig, chart.ax = lux.config.plot_config(chart.fig, chart.ax)
+                chart.ax = lux.config.plotting_style(chart.fig, chart.ax)
             plt.tight_layout()
             tmpfile = BytesIO()
             chart.fig.savefig(tmpfile, format="png")
@@ -100,11 +100,11 @@ def create_vis(self, vis, standalone=True):
             if self.output_type == "matplotlib":
                 return {"config": chart.chart, "vislib": "matplotlib"}
             if self.output_type == "matplotlib_code":
-                if lux.config.plot_config:
+                if lux.config.plotting_style:
                     import inspect
 
                     chart.code += "\n".join(
-                        inspect.getsource(lux.config.plot_config).split("\n    ")[1:-1]
+                        inspect.getsource(lux.config.plotting_style).split("\n    ")[1:-1]
                     )
                 chart.code += "\nfig"
                 chart.code = chart.code.replace("\n\t\t", "\n")