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
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
//! Prediction **problem** shape — **structure only, does NOT predict**.
//!
//! This module represents the *shape* of a predictive-process-monitoring
//! problem: a prefix trace plus the kind of target being asked about
//! (next-activity, outcome, remaining-time, drift). It is a **problem
//! statement**, not a predictor.
//!
//! ## What this module **IS**
//!
//! - The structural vocabulary of prediction problems: [`crate::prediction::PredictionProblem`]
//! and the target witness markers [`crate::prediction::PrefixTrace`], [`crate::prediction::OutcomeLabel`],
//! [`crate::prediction::RemainingTime`], [`crate::prediction::NextActivity`], [`crate::prediction::DriftSignal`].
//! - A first-class [`crate::prediction::PredictionRefusal`] surface naming exactly why a problem
//! shape is inadmissible.
//!
//! ## What this module is **NOT**
//!
//! - **Not** a model, a feature encoder, a regressor, or a classifier. It states
//! and refuses *problem shapes*; it never *predicts* an answer.
//!
//! ## Graduation
//!
//! When you need to **train, encode, or run** a predictive model, graduate this
//! problem shape to the `wasm4pm` engine (via the `wasm4pm` feature). This
//! module only certifies that the *problem statement* is well-formed.
use PhantomData;
// ── Prediction horizon ───────────────────────────────────────────────────────
/// The look-ahead distance for a predictive-process-monitoring problem.
///
/// `PredictionHorizon` classifies *how far ahead* a prediction spans:
///
/// - `FullCase` — the prediction covers the entire remaining case (no fixed
/// bound). This is the default for outcome and remaining-time prediction.
/// - `Events(n)` — the prediction spans exactly `n` future events. Used for
/// next-activity or short-range sequence prediction.
/// - `TimeUnits(secs)` — the prediction spans a real-time window of `secs`
/// seconds ahead. Used for deadline and SLA compliance prediction.
///
/// ## What this is
///
/// A **shape** for the horizon concept: it names what is being asked, it does
/// not compute or enforce the horizon against a log. Graduate to `wasm4pm` for
/// horizon enforcement during prediction.
///
/// ## Usage
///
/// `PredictionProblem` stores the horizon as `Option<usize>` (event count) for
/// backward compatibility. `PredictionHorizon` is the richer named type for new
/// surfaces that need to distinguish time-based from event-based horizons.
///
/// ```
/// use wasm4pm_compat::prediction::PredictionHorizon;
/// assert!(matches!(PredictionHorizon::FullCase, PredictionHorizon::FullCase));
/// assert!(matches!(PredictionHorizon::Events(3), PredictionHorizon::Events(3)));
/// assert!(matches!(PredictionHorizon::TimeUnits(86400), PredictionHorizon::TimeUnits(_)));
/// ```
// ── Compliance kind ───────────────────────────────────────────────────────────
/// The sub-kind of a compliance-aware prediction target.
///
/// `ComplianceTarget` is a unit-struct phantom witness that identifies the
/// *target family* at the type level. `ComplianceKind` refines that by naming
/// the **operational context** in which compliance is evaluated: is it a live
/// monitoring check, a post-hoc audit, or a regulatory certification sweep?
///
/// ## What this is
///
/// A closed enum for runtime dispatch on compliance context. It travels
/// alongside `PredictionTarget::ComplianceConstraint` as a metadata tag, not
/// as a phantom type parameter.
///
/// ## What this is NOT
///
/// - Not a phantom type witness — use [`ComplianceTarget`] for that.
/// - Not a constraint definition — named rules are a `wasm4pm` concern.
/// - Not an enforcement mechanism — structure only.
///
/// ## Variants
///
/// | Variant | Meaning |
/// |---------|---------|
/// | `Monitoring` | Online / streaming compliance check during case execution. |
/// | `Audit` | Post-hoc audit of a completed or historical case. |
/// | `Certification` | Regulatory or standard-compliance sweep across a log. |
///
/// ```
/// use wasm4pm_compat::prediction::ComplianceKind;
/// let k = ComplianceKind::Audit;
/// assert_eq!(format!("{k}"), "audit");
/// ```
// ── Target witness markers ──────────────────────────────────────────────────
/// Witness: the problem's input is a **prefix trace** (a case observed so far).
;
/// Witness: the problem's target is a categorical **outcome label**.
;
/// Witness: the problem's target is a **remaining-time** regression value.
;
/// Witness: the problem's target is the **next activity** in the case.
;
/// Witness: the problem's target is a **drift signal** (a change-point claim).
;
/// Witness: the problem's target is a **risk score** (a threat / hazard
/// probability estimate).
;
/// Witness: the problem's prediction target is a **compliance constraint check**.
///
/// De Santis et al. (2026) introduce compliance-aware predictive process
/// monitoring (PPM) where the prediction target is not an outcome label but a
/// named compliance rule: "does this prefix comply with constraint C?". A
/// `PredictionProblem<ComplianceTarget>` encodes the shape of such a problem.
///
/// This witness is structurally distinct from [`OutcomeLabel`]: a compliance
/// target must name its constraint (see [`PredictionTarget::ComplianceConstraint`]).
/// Without it, a compliance-constrained prediction is indistinguishable from
/// a plain binary outcome problem.
///
/// Structure-only marker: the LTN training and inference routines graduate to
/// `wasm4pm`.
;
// ── Core shapes ─────────────────────────────────────────────────────────────
/// The kind of prediction target a problem asks about.
///
/// **Structure only**: records *what is being asked*, never *the answer*.
///
/// [`PredictionTarget::ComplianceConstraint`] is the target kind for
/// compliance-aware PPM (De Santis et al., 2026): the question is not "what is
/// the outcome?" but "does this prefix comply with named rule C?".
/// A complete prediction problem: the observed prefix and the target asked of
/// it, tagged with a target witness `T`.
///
/// The witness `T` (e.g. [`NextActivity`]) records the target family at the
/// type level. The top-level **shape** of a predictive monitoring problem; it
/// does **NOT** encode features, train a model, or emit a prediction. Graduate
/// to `wasm4pm` to actually predict.
///
/// `horizon` is the look-ahead distance (in events or time units) the
/// prediction spans. `None` means the prediction covers the full remaining
/// case.
/// First-class refusal law for prediction problem shapes.
///
/// Every variant names a **specific** structural law — never a bare
/// "InvalidInput".