feat: add font_path arg for wordcloud (#1598)

using font_path argument in Wordcloud, if we don't use a font in Wordcloud, multi-language be broken Adjust lint for commit Co-authored-by: JoDongHyun <dong.jo@navercorp.com>
ydataai · Jun 17, 2024 · 0e8fedd · 0e8fedd
1 parent 9946037
commit 0e8fedd
Show file tree

Hide file tree

Showing 4 changed files with 13 additions and 2 deletions.
diff --git a/src/ydata_profiling/config.py b/src/ydata_profiling/config.py
@@ -173,6 +173,7 @@ class Plot(BaseModel):
     histogram: Histogram = Histogram()
     scatter_threshold: int = 1000
     cat_freq: CatFrequencyPlot = CatFrequencyPlot()
+    font_path: Optional[Union[Path, str]] = None
 
 
 class Theme(Enum):

diff --git a/src/ydata_profiling/config_default.yaml b/src/ydata_profiling/config_default.yaml
@@ -159,6 +159,8 @@ plot:
         # Maximum number of bins (when bins=0)
         max_bins: 250
 
+    font_path: null
+
 # The number of observations to show
 n_obs_unique: 5
 n_extreme_obs: 5

diff --git a/src/ydata_profiling/config_minimal.yaml b/src/ydata_profiling/config_minimal.yaml
@@ -159,6 +159,8 @@ plot:
         # Maximum number of bins (when bins=0)
         max_bins: 250
 
+    font_path: null
+
 # The number of observations to show
 n_obs_unique: 5
 n_extreme_obs: 5

diff --git a/src/ydata_profiling/visualisation/plot.py b/src/ydata_profiling/visualisation/plot.py
@@ -27,6 +27,7 @@ def format_fn(tick_val: int, tick_pos: Any) -> str:
 
 
 def _plot_word_cloud(
+    config: Settings,
     series: Union[pd.Series, List[pd.Series]],
     figsize: tuple = (6, 4),
 ) -> plt.Figure:
@@ -36,7 +37,12 @@ def _plot_word_cloud(
     for i, series_data in enumerate(series):
         word_dict = series_data.to_dict()
         wordcloud = WordCloud(
-            background_color="white", random_state=123, width=300, height=200, scale=2
+            font_path=config.plot.font_path,
+            background_color="white",
+            random_state=123,
+            width=300,
+            height=200,
+            scale=2,
         ).generate_from_frequencies(word_dict)
 
         ax = plot.add_subplot(1, len(series), i + 1)
@@ -124,7 +130,7 @@ def _plot_histogram(
 
 @manage_matplotlib_context()
 def plot_word_cloud(config: Settings, word_counts: pd.Series) -> str:
-    _plot_word_cloud(series=word_counts)
+    _plot_word_cloud(config=config, series=word_counts)
     return plot_360_n0sc0pe(config)