remove sample values & show most frequent

src/skrubview/_data/templates/base.css

@@ -113,6 +113,14 @@
     font-size: var(--skrubview-large);
 }
 
+{{ report_id_selector }}.skrubview-report .skrubview-box {
+    border: 1px solid #555555;
+    padding: var(--skrubview-tiny);
+    max-width: 40rem;
+    max-height: 10rem;
+    overflow: auto;
+}
+
 {{ report_id_selector }}.skrubview-report :is(.skrubview-ok, .skrubview-warning, .skrubview-critical) {
                                                  font-weight: bold;
                                              }

src/skrubview/_data/templates/column-summary.html

@@ -49,7 +49,17 @@ <h3 class="skrubview-margin-r-t">
         </div>
 
         {% for plot_name in column.plot_names %}
+        <div>
         <img class="pure-img" src="{{ column[plot_name] | svg_to_img_src | safe }}" alt={{ plot_name }} />
+        {% if plot_name == "value_counts_plot"  and column.dtype == "String" %}
+        <details>
+            <summary>Most frequent values</summary>
+            {% for value in column.value_counts %}
+            <pre class="skrubview-box">{{ value }}</pre>
+            {% endfor %}
+        </details>
+        {% endif %}
+        </div>
         {% endfor %}
 
     </div>

src/skrubview/_summarize.py

@@ -61,7 +61,6 @@ def _summarize_column(
         "value_is_constant": False,
     }
     _add_nulls_summary(summary, column, dataframe_summary=dataframe_summary)
-    _add_sample_values(summary, column)
     _add_value_counts(
         summary, column, dataframe_summary=dataframe_summary, with_plots=with_plots
     )
@@ -86,20 +85,6 @@ def _add_nulls_summary(summary, column, dataframe_summary):
         summary["nulls_level"] = "warning"
 
 
-def _add_sample_values(summary, column):
-    rng = np.random.default_rng(0)
-    non_missing = column.filter(~column.is_null())
-    n_non_missing = int(non_missing.len())
-    if n_non_missing == 0:
-        return
-    size = min(n_non_missing, 5)
-    sample_indices = sorted(rng.choice(range(n_non_missing), replace=False, size=size))
-    ns = column.__column_namespace__()
-    sample_indices = ns.column_from_sequence(sample_indices)
-    sample_values = np.asarray(non_missing.take(sample_indices).to_array()).tolist()
-    summary["sample_values"] = list(map(_utils.ellide_string, sample_values))
-
-
 def _add_value_counts(summary, column, *, dataframe_summary, with_plots):
     ns = column.__column_namespace__()
     dtype = _utils.get_dtype(column)

src/skrubview/_text.py

@@ -42,7 +42,7 @@ def _print_summary(summary, console):
 
 def _print_first_row(summary, console):
     console.print("First row:")
-    console.print(summary["first_row_dict"])
+    console.print({k: _utils.ellide_string(v) for (k, v) in summary["first_row_dict"].items()})
 
 def _print_constant_columns(summary, console):
     cols = [col for col in summary["columns"] if col.get("value_is_constant")]
@@ -72,10 +72,10 @@ def _print_column_summary(summary, console):
     text.append(f"[{color}]{summary['null_proportion']:0.2%}[/{color}]\n")
     if "n_unique" in summary:
         text.append(f"Unique values: {summary['n_unique']}\n")
-    if not summary["high_cardinality"]:
-        text.append(f"Value counts: {summary['value_counts']}\n")
-    elif summary["dtype"] == "String":
-        text.append(f"Sample values: {list(summary['sample_values'])}\n")
+    if "value_counts" in summary:
+        # TODO in theory ellide_string could create collisions
+        ellided = {_utils.ellide_string(k): v for (k, v) in summary["value_counts"].items()}
+        text.append(f"Most frequent value counts: {ellided}\n")
     if "mean" in summary:
         text.append(
             f"Mean: {summary['mean']:#0.3g} "