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
//! Applied decoding and component mappings retained by processed datasets.
use super::{CanonicalDatasetDigests, SourceFile, SourceId};
/// Applied decoding, separate from subsequent numerical processing.
/// Only a reader can attach this immutable record to a dataset.
#[derive(Clone, Debug, PartialEq)]
pub struct ProcessedReadRecord {
transform: ProcessedReadTransform,
components: Vec<SourceId>,
component_indices: Vec<Vec<usize>>,
output: crate::provenance::CanonicalDatasetDigests,
}
/// Format-specific numeric decoding parameters actually used by a reader.
#[non_exhaustive]
#[derive(Clone, Debug, PartialEq)]
pub enum ProcessedReadTransform {
/// Experimental two-dimensional JEOL section decoding, with disk-axis ranges.
JeolDelta2D {
/// True for binary64 source values.
float64: bool,
/// Numeric source byte order.
big_endian: bool,
/// Original point counts in x,y disk order.
disk_points: [usize; 2],
/// First retained x,y points.
crop_start: [usize; 2],
/// Retained x,y point counts.
crop_points: [usize; 2],
/// JEOL section-reordering tile edge.
submatrix_edge: usize,
/// Applied direct imaginary multiplier.
imaginary_multiplier: i8,
},
/// Binary Bruker scalar decoding and intensity multiplication by 2^NC_PROC.
Bruker {
/// Applied power-of-two intensity exponent.
nc_proc: i32,
/// True for IEEE binary64, false for signed int32.
float64: bool,
/// True for big-endian scalars.
big_endian: bool,
},
/// JCAMP ASDF coordinate-checkpoint and intensity scaling.
JcampDx {
/// Factor applied to encoded X checkpoints, not FIRSTX/LASTX.
x_factor: f64,
/// Factor applied to decoded Y values.
y_factor: f64,
},
/// Supported one-dimensional JEOL section decoding and valid-window extraction.
/// Records which library rule was applied; this is not vendor certification.
JeolDelta {
/// True for binary64, false for binary32 source values.
float64: bool,
/// Byte order of numeric sample sections.
big_endian: bool,
/// Logical point count in each original section.
disk_points: usize,
/// First retained point in the original section.
crop_start: usize,
/// Number of retained points.
crop_points: usize,
/// Multiplier applied to the stored imaginary section.
imaginary_multiplier: i8,
/// Edge used by the existing section-reordering rule.
submatrix_edge: usize,
/// Decimal unit multiplier applied to header or explicit-list coordinates.
coordinate_scale: f64,
/// Whether coordinates came from an explicit source axis list.
explicit_coordinates: bool,
},
}
impl ProcessedReadRecord {
pub(crate) fn validate_recorded(
&self,
descriptor: &crate::processed::ProcessedDescriptor,
digests: CanonicalDatasetDigests,
sources: &[SourceFile],
) -> Result<(), crate::internal::ModelError> {
use crate::internal::ModelError;
let counts = descriptor.component_counts();
if self.output != digests
|| self.components.len() != self.component_indices.len()
|| self.components.len() != counts.iter().product::<usize>()
{
return Err(ModelError::Structure);
}
for (i, (id, coordinate)) in self
.components
.iter()
.zip(&self.component_indices)
.enumerate()
{
if !sources.iter().any(|source| source.id() == Some(*id))
|| coordinate.len() != counts.len()
|| coordinate.iter().zip(&counts).any(|(v, n)| v >= n)
|| self.component_indices[..i].contains(coordinate)
{
return Err(ModelError::Structure);
}
}
let valid = match self.transform {
ProcessedReadTransform::JeolDelta2D {
disk_points,
crop_start,
crop_points,
submatrix_edge,
imaginary_multiplier,
..
} => {
counts.len() == 2
&& submatrix_edge > 0
&& matches!(imaginary_multiplier, -1 | 1)
&& (0..2).all(|a| {
crop_points[a] > 0
&& crop_points[a] == descriptor.axes()[1 - a].points()
&& crop_start[a]
.checked_add(crop_points[a])
.is_some_and(|end| end <= disk_points[a])
})
}
ProcessedReadTransform::Bruker { nc_proc, .. } => {
let scale = 2f64.powi(nc_proc);
scale.is_finite() && scale != 0.0
}
ProcessedReadTransform::JcampDx { x_factor, y_factor } => {
x_factor.is_finite() && y_factor.is_finite() && x_factor != 0.0 && y_factor != 0.0
}
ProcessedReadTransform::JeolDelta {
disk_points,
crop_start,
crop_points,
imaginary_multiplier,
submatrix_edge,
coordinate_scale,
..
} => {
crop_points > 0
&& crop_start
.checked_add(crop_points)
.is_some_and(|end| end <= disk_points)
&& matches!(imaginary_multiplier, -1 | 1)
&& submatrix_edge > 0
&& coordinate_scale.is_finite()
&& coordinate_scale > 0.0
}
};
if !valid {
return Err(ModelError::Structure);
}
Ok(())
}
/// Version of the decoding and component-interleaving rule.
pub fn algorithm_version(&self) -> &'static str {
match self.transform {
ProcessedReadTransform::JeolDelta2D { .. } => "jeol.processed-2d.v1",
ProcessedReadTransform::Bruker { .. } if self.component_indices[0].len() == 1 => {
"bruker.processed-1d.v1"
}
ProcessedReadTransform::Bruker { .. } => "bruker.processed-2d.v1",
ProcessedReadTransform::JcampDx { .. } => "jcamp.xydata-asdf.v1",
ProcessedReadTransform::JeolDelta { .. } => "jeol.processed-1d.v1",
}
}
/// Typed format-specific decoding parameters.
pub fn transform(&self) -> &ProcessedReadTransform {
&self.transform
}
/// Source planes in reader order; see component_indices for tensor mapping.
pub fn components(&self) -> &[SourceId] {
&self.components
}
/// Per-plane tensor component coordinates, in the same order as components.
pub fn component_indices(&self) -> &[Vec<usize>] {
&self.component_indices
}
/// Canonical output identity produced by this reader execution.
pub fn output_digests(&self) -> crate::provenance::CanonicalDatasetDigests {
self.output
}
pub(crate) fn bruker(
nc_proc: i32,
float64: bool,
big_endian: bool,
components: Vec<SourceId>,
component_indices: Vec<Vec<usize>>,
output: crate::provenance::CanonicalDatasetDigests,
) -> Self {
Self {
transform: ProcessedReadTransform::Bruker {
nc_proc,
float64,
big_endian,
},
components,
component_indices,
output,
}
}
pub(crate) fn jcamp(
source: SourceId,
x_factor: f64,
y_factor: f64,
output: crate::provenance::CanonicalDatasetDigests,
) -> Self {
Self {
transform: ProcessedReadTransform::JcampDx { x_factor, y_factor },
components: vec![source],
component_indices: vec![vec![0]],
output,
}
}
pub(crate) fn jeol(
source: SourceId,
transform: ProcessedReadTransform,
component_count: usize,
output: crate::provenance::CanonicalDatasetDigests,
) -> Self {
Self {
transform,
components: vec![source; component_count],
component_indices: (0..component_count)
.map(|component| vec![component])
.collect(),
output,
}
}
pub(crate) fn jeol_2d(
source: SourceId,
transform: ProcessedReadTransform,
counts: &[usize],
output: CanonicalDatasetDigests,
) -> Self {
Self {
transform,
components: vec![source; counts.iter().product()],
component_indices: (0..counts[0])
.flat_map(|a| (0..counts[1]).map(move |b| vec![a, b]))
.collect(),
output,
}
}
}
// Crate-private model decomposition; no wire tags or encoding policy.
impl ProcessedReadRecord {
#[allow(clippy::type_complexity)]
pub(crate) fn model_parts(
&self,
) -> (
&ProcessedReadTransform,
&Vec<SourceId>,
&Vec<Vec<usize>>,
&crate::provenance::CanonicalDatasetDigests,
) {
(
&self.transform,
&self.components,
&self.component_indices,
&self.output,
)
}
#[allow(clippy::type_complexity)]
pub(crate) fn from_model_parts(
parts: (
ProcessedReadTransform,
Vec<SourceId>,
Vec<Vec<usize>>,
crate::provenance::CanonicalDatasetDigests,
),
) -> Result<Self, crate::internal::ModelError> {
let (transform, components, component_indices, output) = parts;
let value = Self {
transform,
components,
component_indices,
output,
};
if value.components.is_empty()
|| value.components.len() != value.component_indices.len()
|| value.component_indices.iter().any(Vec::is_empty)
{
return Err(crate::internal::ModelError::Structure);
}
Ok(value)
}
}