cleanup plots

varats/varats/plots/interactions_change_distribution.py

@@ -34,6 +34,7 @@ def plot(self, view_mode: bool) -> None:
             "interactions_diff": [],
             "project": []
         })
+        plt.rcParams.update({"text.usetex": True, "font.family": "Helvetica"})
         for case_study in case_studys:
             cs_data = lines_per_interactions_squashed(case_study, True)
             cs_data["interactions_diff"] = cs_data.groupby(
@@ -49,20 +50,22 @@ def plot(self, view_mode: bool) -> None:
 
         df = data_sub.to_frame().reset_index()
         df["interactions_diff"] = df["interactions_diff"].apply(lambda x: x + 1)
+        df["project"] = df["project"].apply(lambda x: f"\\textsc{{{x}}}")
         axis = sns.violinplot(
             data=df,
             y="interactions_diff",
             x="project",
             bw=0.15,
             scale="width",
-            inner=None
+            inner=None,
+            cut=0,
         )
         axis.plot(range(len(case_studys)), [1 for _ in case_studys], "--k")
 
-        axis.set_ylabel("")
-        axis.yaxis.set_visible(False)
+        axis.set_ylabel("Change in $\\frac{interactions}{lines}$")
+        axis.set_xlabel("Project")
+        plt.gcf().set_size_inches(10, 5)
         plt.yscale("asinh")
-        axis.set_yticklabels([])
 
 
 class InteractionChangeAuthorDistribution(

varats/varats/plots/revision_impact.py

@@ -138,17 +138,39 @@ def __init__(self, plot_config: PlotConfig, **kwargs: tp.Any):
         super().__init__(plot_config, **kwargs)
 
     def plot(self, view_mode: bool) -> None:
-        case_studys: tp.List[CaseStudy] = self.plot_kwargs["case_study"]
         data = self.plot_kwargs["data"]
-        axis = sns.violinplot(
+        full_data = self.plot_kwargs["data"].copy()
+        full_data["project"] = "All"
+        data["project"] = data["project"].apply(lambda x: f"\\textsc{{{x}}}")
+        plt.rcParams.update({"text.usetex": True, "font.family": "Helvetica"})
+        grid = sns.JointGrid(
+            x="project",
+            y="impacted_commits",
+            data=data,
+        )
+
+        sns.violinplot(
+            ax=grid.ax_joint,
             data=data,
             y="impacted_commits",
             x="project",
             bw=0.15,
             scale="width",
-            inner=None
+            inner=None,
+            cut=0
         )
-        axis.plot(range(len(case_studys)), [0 for _ in case_studys], "--k")
+        sns.kdeplot(
+            ax=grid.ax_marg_y,
+            data=full_data,
+            y="impacted_commits",
+            bw=0.15,
+            cut=0,
+            fill=True,
+        )
+        grid.set_axis_labels("Projects", "Impact", fontsize=14)
+        plt.gcf().set_size_inches(10, 5)
+        plt.xticks(fontsize=12)
+        plt.yticks(fontsize=12)
 
 
 class RevisionImpactDistributionAll(
@@ -168,7 +190,8 @@ def __init__(self, plot_config: PlotConfig, **kwargs: tp.Any):
     def plot(self, view_mode: bool) -> None:
         data = self.plot_kwargs["data"]
         data.drop(columns=["project"], inplace=True)
-        sns.violinplot(data=data, y="impacted_commits", bw=0.15)
+        sns.violinplot(data=data, y="impacted_commits", bw=0.15, cut=0)
+        plt.ylabel("Impact", fontsize=10)
 
 
 class RevisionImpactScatterInteractions(
@@ -223,13 +246,17 @@ def plot(self, view_mode: bool) -> None:
                 )
             )
         data = apply_tukeys_fence(data, "line_change", 3)
-        multivariate_grid(
+        data["project"] = data["project"].apply(lambda x: f"\\textsc{{{x}}}")
+
+        plt.rcParams.update({"text.usetex": True, "font.family": "Helvetica"})
+        grid = multivariate_grid(
             data,
             "impacted_commits",
             "line_change",
             "project",
-            alpha=0.3,
         )
+        grid.set_axis_labels("Impact", "Changed Lines", fontsize=10)
+        plt.gcf().set_size_inches(10, 5)
         ymax = data["line_change"].max()
         plt.ylim(-0.001, ymax + 0.01)
 

varats/varats/plots/surviving_commits.py

@@ -248,6 +248,7 @@ def name(self) -> str:
     def plot(self, view_mode: bool) -> None:
         """Plot the evolution of a single commit."""
         _, axis = plt.subplots(1, 1)
+        plt.rcParams.update({"text.usetex": True, "font.family": "Helvetica"})
         case_study = self.plot_kwargs['case_study']
         lines: DataFrame = get_lines_per_commit_long(case_study, False)
 
@@ -278,19 +279,16 @@ def plot(self, view_mode: bool) -> None:
         ax = axis.twinx()
         x_axis = range(len(data))
         ax.scatter(x_axis, data['lines'], color="green")
-        ax.set_ylabel("Lines", color="g")
+        ax.set_ylabel("Lines", color="g", fontsize=14)
         axis.scatter(x_axis, data['interactions'], color="orange")
-        axis.set_ylabel("Interactions", color="orange")
+        axis.set_ylabel("Interactions", color="orange", fontsize=14)
+        axis.tick_params(axis="y", labelsize=14)
         ax.set_ylim(ymin=0)
         axis.set_ylim(ymin=0)
-        lines_legend = mpatches.Patch(color='green', label="Lines")
-        interactions_legend = mpatches.Patch(
-            color="orange", label='Interactions'
-        )
-        plt.legend(handles=[lines_legend, interactions_legend])
-        plt.ticklabel_format(axis='x', useOffset=False)
-        axis.set_xticklabels(data.index.map(cmap.short_time_id), rotation=90)
-        axis.tick_params(axis="x", labelrotation=90)
+        axis.set_xlabel("Revisions", fontsize=14)
+        axis.set_xticklabels([])
+        plt.xticks(fontsize=14)
+        plt.yticks(fontsize=14)
 
 
 class HeatMapPlot(Plot, plot_name=None):

varats/varats/tables/revision_impact.py

@@ -4,6 +4,7 @@
 import pandas as pd
 
 from varats.mapping.commit_map import CommitMap, get_commit_map
+from varats.paper.case_study import CaseStudy
 from varats.plots.revision_impact import impact_data
 from varats.project.project_util import get_primary_project_source
 from varats.table.table import Table
@@ -110,6 +111,16 @@ def tabulate(self, table_format: TableFormat, wrap_table: bool) -> str:
         )
 
 
+class RandomCommitsTable(Table, table_name="volatile_commits"):
+
+    def tabulate(self, table_format: TableFormat, wrap_table: bool) -> str:
+        case_studys: tp.List[CaseStudy] = self.table_kwargs["case_study"]
+        data = impact_data(case_studys)
+        df = data.sample(n=100)
+        df.to_csv("random-commits.csv")
+        return dataframe_to_table(df, table_format, wrap_table=wrap_table)
+
+
 class LineImpactDistribution(Table, table_name="line_impact_distribution"):
     """Table showing high impact revisions."""
 
@@ -241,3 +252,13 @@ def generate(self) -> tp.List['varats.table.table.Table']:
             ImpactDistribution(self.table_config, **self.table_kwargs),
             LineImpactDistribution(self.table_config, **self.table_kwargs)
         ]
+
+
+class RandomCommitsGenerator(
+    TableGenerator,
+    generator_name="random-commits",
+    options=[REQUIRE_MULTI_CASE_STUDY]
+):
+
+    def generate(self) -> tp.List['varats.table.table.Table']:
+        return [RandomCommitsTable(self.table_config, **self.table_kwargs)]

varats/varats/tables/surviving_commits.py

@@ -129,6 +129,7 @@ def commmit_data(commit: FullCommitHash,
                     commmit_data(row["base_hash"], row["interactions_diff"])
                 )
         df = pd.DataFrame(volatile_commits)
+        df.to_csv("volatile_commits.csv")
         return dataframe_to_table(df, table_format, wrap_table=wrap_table)