1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
use crate::TEMP_SUFFIX;
use crate::chart::Chart;
use crate::core::data::{ColumnVector, Dataset, get_quantile};
use crate::error::ChartonError;
use crate::mark::Mark;
use ahash::AHashMap;
impl<T: Mark> Chart<T> {
/// Performs high-performance statistical aggregation for Box Plots.
///
/// This version uses `unique_values()` to ensure that both X-axis categories
/// and Color dodge slots maintain a stable, appearance-based order.
pub(crate) fn transform_boxplot_data(mut self) -> Result<Self, ChartonError> {
let x_name = &self.encoding.x.as_ref().unwrap().field;
let y_name = &self.encoding.y.as_ref().unwrap().field;
// --- STEP 1: Capture raw columns and calculate global Y-axis boundaries ---
let x_col = self.data.column(x_name)?;
let y_col = self.data.column(y_name)?;
let row_count = self.data.height();
let mut global_min = f64::INFINITY;
let mut global_max = f64::NEG_INFINITY;
for i in 0..row_count {
if let Some(v) = y_col.get_f64(i) {
if v < global_min {
global_min = v;
}
if v > global_max {
global_max = v;
}
}
}
// --- STEP 2: Establish Deterministic Order for X and Color ---
let x_order = x_col.unique_values();
let mut color_field_name: Option<String> = None;
let mut color_order = Vec::new();
if let Some(color_enc) = &self.encoding.color {
let cf = color_enc.field.clone();
color_order = self.data.column(&cf)?.unique_values();
color_field_name = Some(cf);
}
let groups_count = if color_order.is_empty() {
1.0
} else {
color_order.len() as f64
};
// --- STEP 3: Grouping phase ---
let mut group_map: AHashMap<(String, Option<String>), Vec<usize>> = AHashMap::new();
for i in 0..row_count {
let x_val = x_col.get_str_or(i, "null");
let c_val = color_field_name
.as_ref()
.map(|f| self.data.get_str_or(f, i, "null"));
group_map.entry((x_val, c_val)).or_default().push(i);
}
// --- STEP 4: Cartesian Product & Statistical Computation ---
// We iterate through every possible X + Color combination to ensure
// gaps are preserved (Gap Filling).
let mut final_x = Vec::new();
let mut final_y = Vec::new();
let mut final_c = Vec::new();
let mut f_q1 = Vec::new();
let mut f_median = Vec::new();
let mut f_q3 = Vec::new();
let mut f_min = Vec::new();
let mut f_max = Vec::new();
let mut f_sub_idx = Vec::new();
let mut f_outliers = Vec::new();
for x_val in &x_order {
// Handle both grouped and non-grouped scenarios
let sub_tasks: Vec<(f64, Option<String>)> = if color_order.is_empty() {
vec![(0.0, None)]
} else {
color_order
.iter()
.enumerate()
.map(|(i, c)| (i as f64, Some(c.clone())))
.collect()
};
for (c_idx, c_val) in sub_tasks {
final_x.push(x_val.clone());
final_c.push(
c_val
.clone()
.unwrap_or_else(|| format!("{}_default", TEMP_SUFFIX)),
);
f_sub_idx.push(c_idx);
if let Some(indices) = group_map.get(&(x_val.clone(), c_val)) {
let mut vals: Vec<f64> =
indices.iter().filter_map(|&i| y_col.get_f64(i)).collect();
vals.sort_unstable_by(|a, b| {
a.partial_cmp(b).unwrap_or(std::cmp::Ordering::Equal)
});
if vals.is_empty() {
push_nan_row(
&mut final_y,
&mut f_q1,
&mut f_median,
&mut f_q3,
&mut f_min,
&mut f_max,
&mut f_outliers,
);
} else {
let q1 = get_quantile(&vals, 0.25);
let median = get_quantile(&vals, 0.50);
let q3 = get_quantile(&vals, 0.75);
let iqr = q3 - q1;
let lower_fence = q1 - 1.5 * iqr;
let upper_fence = q3 + 1.5 * iqr;
let mut outliers = Vec::new();
let mut whisker_min = q1;
let mut whisker_max = q3;
for &v in &vals {
if v < lower_fence || v > upper_fence {
outliers.push(v);
} else {
if v < whisker_min {
whisker_min = v;
}
if v > whisker_max {
whisker_max = v;
}
}
}
final_y.push(median);
f_q1.push(q1);
f_median.push(median);
f_q3.push(q3);
f_min.push(whisker_min);
f_max.push(whisker_max);
f_outliers.push(format!("{:?}", outliers));
}
} else {
// Force a placeholder row for missing data to maintain DODGE alignment
push_nan_row(
&mut final_y,
&mut f_q1,
&mut f_median,
&mut f_q3,
&mut f_min,
&mut f_max,
&mut f_outliers,
);
}
}
}
// --- STEP 5: Boundary Injection ---
// Append two extra rows with global min/max.
// This ensures the Y-axis scale covers all data including outliers across all groups.
if global_min.is_finite() {
inject_boundary_row(
global_min,
&mut final_x,
&mut final_y,
&mut final_c,
&mut f_q1,
&mut f_median,
&mut f_q3,
&mut f_min,
&mut f_max,
&mut f_sub_idx,
&mut f_outliers,
);
inject_boundary_row(
global_max,
&mut final_x,
&mut final_y,
&mut final_c,
&mut f_q1,
&mut f_median,
&mut f_q3,
&mut f_min,
&mut f_max,
&mut f_sub_idx,
&mut f_outliers,
);
}
// --- STEP 6: Final Dataset Assembly ---
let mut new_ds = Dataset::new();
let result_len = final_x.len();
new_ds.add_column(
x_name,
ColumnVector::String {
data: final_x,
validity: None,
},
)?;
new_ds.add_column(y_name, ColumnVector::F64 { data: final_y })?;
new_ds.add_column(
format!("{}_q1", TEMP_SUFFIX),
ColumnVector::F64 { data: f_q1 },
)?;
new_ds.add_column(
format!("{}_median", TEMP_SUFFIX),
ColumnVector::F64 { data: f_median },
)?;
new_ds.add_column(
format!("{}_q3", TEMP_SUFFIX),
ColumnVector::F64 { data: f_q3 },
)?;
new_ds.add_column(
format!("{}_min", TEMP_SUFFIX),
ColumnVector::F64 { data: f_min },
)?;
new_ds.add_column(
format!("{}_max", TEMP_SUFFIX),
ColumnVector::F64 { data: f_max },
)?;
new_ds.add_column(
format!("{}_sub_idx", TEMP_SUFFIX),
ColumnVector::F64 { data: f_sub_idx },
)?;
new_ds.add_column(
format!("{}_groups_count", TEMP_SUFFIX),
ColumnVector::F64 {
data: vec![groups_count; result_len],
},
)?;
new_ds.add_column(
format!("{}_outliers", TEMP_SUFFIX),
ColumnVector::String {
data: f_outliers,
validity: None,
},
)?;
if let Some(ref f) = color_field_name {
new_ds.add_column(
f,
ColumnVector::String {
data: final_c,
validity: None,
},
)?;
}
self.data = new_ds;
Ok(self)
}
}
// Helper to push NaN rows for gaps
fn push_nan_row(
y: &mut Vec<f64>,
q1: &mut Vec<f64>,
med: &mut Vec<f64>,
q3: &mut Vec<f64>,
min: &mut Vec<f64>,
max: &mut Vec<f64>,
out: &mut Vec<String>,
) {
y.push(f64::NAN);
q1.push(f64::NAN);
med.push(f64::NAN);
q3.push(f64::NAN);
min.push(f64::NAN);
max.push(f64::NAN);
out.push("[]".to_string());
}
/// Helper to inject invisible boundary points to ensure the Y-axis scale
/// covers the absolute global range (including outliers).
#[allow(clippy::too_many_arguments)]
fn inject_boundary_row(
val: f64,
x: &mut Vec<String>,
y: &mut Vec<f64>,
c: &mut Vec<String>,
q1: &mut Vec<f64>,
med: &mut Vec<f64>,
q3: &mut Vec<f64>,
min: &mut Vec<f64>,
max: &mut Vec<f64>,
s_idx: &mut Vec<f64>,
out: &mut Vec<String>,
) {
// STRATEGY: "Cloaking" (data exists, but is invisible)
// We use a unique boundary name that the renderer will ignore.
// This forced value in the 'y' column ensures the Scale accommodates
// the furthest outliers without drawing any visual box marks.
x.push(format!("{}_boundary", TEMP_SUFFIX));
y.push(val);
c.push(format!("{}_default", TEMP_SUFFIX));
q1.push(f64::NAN);
med.push(f64::NAN);
q3.push(f64::NAN);
min.push(f64::NAN);
max.push(f64::NAN);
s_idx.push(0.0);
out.push("[]".to_string());
}