orion-sdr 0.0.52

DSP/SDR block library targeting HF-to-EHF, satellites, and Python bindings. Roadmap inside.
Documentation
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
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
// Copyright (c) 2025-2026 G & R Associates LLC
// SPDX-License-Identifier: MIT OR Apache-2.0

// src/python/ofdm.rs — PyO3 bindings for OFDM: config, TX, RX, sync.
//
// `CarrierGrid`/`FftBlock`/`GridMap` are not exposed individually, matching
// how the per-order symbol mappers aren't exposed today: `PyOfdmMod` and
// `PyOfdmDemod` are the two main entry points, plus a free-function
// `ofdm_sync()` mirroring `ft8_sync`'s pattern.

use num_complex::Complex32;
use numpy::{IntoPyArray, PyArray1, PyReadonlyArray1};
use pyo3::exceptions::PyValueError;
use pyo3::prelude::*;
use pyo3::types::{PyDict, PyList};

use crate::core::Block;
use crate::demodulate::{EqualizerMethod, OfdmDecider, OfdmEqualizer};
use crate::fec::{
    ConvCode, CrcKind, DecodeRule, HeaderFormat, InnerFec, InterleaverKind, LdpcCode, OuterFec,
    PunctureRate, ScramblerKind, ScramblerPos, SeedMode,
};
use crate::modulate::{ConstellationOrder, OfdmConfig, OfdmMod};
use crate::multicarrier::{CarrierGrid, CarrierPlan, CyclicPrefixRemove, FftBlock, GridExtract};
use crate::sync::{OfdmPreamble, generate_ofdm_preamble, ofdm_sync as ofdm_sync_fn};

type SoftDemodulateResult<'py> = (Bound<'py, PyArray1<Complex32>>, Bound<'py, PyArray1<u8>>);

fn parse_constellation(s: &str) -> PyResult<ConstellationOrder> {
    match s {
        "bpsk" => Ok(ConstellationOrder::Bpsk),
        "qpsk" => Ok(ConstellationOrder::Qpsk),
        "qam16" => Ok(ConstellationOrder::Qam16),
        "qam64" => Ok(ConstellationOrder::Qam64),
        "qam256" => Ok(ConstellationOrder::Qam256),
        other => Err(PyValueError::new_err(format!(
            "OfdmConfig: unknown constellation {:?} (expected one of: bpsk, qpsk, qam16, qam64, qam256)",
            other
        ))),
    }
}

// ── OfdmConfig ────────────────────────────────────────────────────────────────

/// OFDM waveform configuration: carrier plan (FFT size, cyclic-prefix
/// length, data/pilot carrier layout) plus RF/constellation parameters.
#[pyclass(name = "OfdmConfig", eq, skip_from_py_object)]
#[derive(Clone, PartialEq)]
pub struct PyOfdmConfig(pub(crate) OfdmConfig);

#[pymethods]
impl PyOfdmConfig {
    /// `pilot_carrier_indices`/`pilot_carrier_values` are parallel arrays
    /// (same length): `pilot_carrier_indices[i]` carries the known symbol
    /// `pilot_carrier_values[i]`. Pass empty arrays for no pilots.
    #[new]
    #[pyo3(signature = (n_fft, cp_len, data_carriers, pilot_carrier_indices, pilot_carrier_values, fs, rf_hz, gain, constellation))]
    #[allow(clippy::too_many_arguments)] // mirrors the Python-facing signature
    fn new<'py>(
        n_fft: usize,
        cp_len: usize,
        data_carriers: PyReadonlyArray1<'py, i32>,
        pilot_carrier_indices: PyReadonlyArray1<'py, i32>,
        pilot_carrier_values: PyReadonlyArray1<'py, Complex32>,
        fs: f32,
        rf_hz: f32,
        gain: f32,
        constellation: &str,
    ) -> PyResult<Self> {
        let pilot_indices = pilot_carrier_indices.as_slice()?;
        let pilot_values = pilot_carrier_values.as_slice()?;
        if pilot_indices.len() != pilot_values.len() {
            return Err(PyValueError::new_err(format!(
                "OfdmConfig: pilot_carrier_indices ({}) and pilot_carrier_values ({}) must have the same length",
                pilot_indices.len(),
                pilot_values.len()
            )));
        }
        let order = parse_constellation(constellation)?;
        let pilots: Vec<(i32, Complex32)> = pilot_indices
            .iter()
            .copied()
            .zip(pilot_values.iter().copied())
            .collect();
        let plan = CarrierPlan::new(n_fft, cp_len)
            .with_data_carriers(data_carriers.as_slice()?.iter().copied())
            .with_pilot_carriers(pilots);
        plan.validate()
            .map_err(|e| PyValueError::new_err(e.to_string()))?;
        Ok(Self(OfdmConfig::new(plan, fs, rf_hz, gain, order)))
    }

    #[getter]
    fn bits_per_ofdm_symbol(&self) -> usize {
        self.0.bits_per_ofdm_symbol()
    }

    #[getter]
    fn samples_per_ofdm_symbol(&self) -> usize {
        self.0.samples_per_ofdm_symbol()
    }

    // ── COFDM frame-layer configuration (builder-style setters) ──
    //
    // Each returns a new config with the field set, so Python can chain:
    //   cfg = (OfdmConfig(...).with_inner_fec("ldpc", "n512r12")
    //                         .with_outer_fec("bch", 8)
    //                         .with_payload_crc("crc32"))

    /// Sets the outer FEC. `kind` is `"none"`, `"bch"`, or `"reed_solomon"`.
    /// For `"bch"`, `a` is `t` (errors per codeword). For `"reed_solomon"`,
    /// `a` is `n` (codeword bytes) and `b` is `n_parity` (`= 2t`).
    #[pyo3(signature = (kind, a = 0, b = 0))]
    fn with_outer_fec(&self, kind: &str, a: usize, b: usize) -> PyResult<Self> {
        let outer = match kind {
            "none" => OuterFec::None,
            "bch" => OuterFec::Bch { t: a },
            "reed_solomon" | "rs" => OuterFec::ReedSolomon { n: a, n_parity: b },
            other => {
                return Err(PyValueError::new_err(format!(
                    "with_outer_fec: unknown kind {other:?} (expected none|bch|reed_solomon)"
                )));
            }
        };
        let mut cfg = self.0.clone();
        cfg.outer_fec = outer;
        Ok(Self(cfg))
    }

    /// Sets the inner FEC. `kind` is `"none"`, `"ldpc"`, `"convolutional"`
    /// (K=5), or `"convolutional_k7"` (DVB-T's K=7 code). For `"ldpc"`, `code`
    /// is `"n512r12"`, `"n576r23"`, or `"n512r34"`. For the convolutional
    /// kinds, `code` is a puncture rate `"1/2"`, `"2/3"`, `"3/4"`, `"5/6"`, or
    /// `"7/8"`.
    #[pyo3(signature = (kind, code = ""))]
    fn with_inner_fec(&self, kind: &str, code: &str) -> PyResult<Self> {
        let inner = match kind {
            "none" => InnerFec::None,
            "ldpc" => InnerFec::Ldpc(parse_ldpc_code(code)?),
            "convolutional" | "conv" => InnerFec::Convolutional {
                rate: parse_puncture_rate(code)?,
                code: ConvCode::K5,
            },
            "convolutional_k7" | "conv_k7" | "dvb_t" => InnerFec::Convolutional {
                rate: parse_puncture_rate(code)?,
                code: ConvCode::DvbK7,
            },
            other => {
                return Err(PyValueError::new_err(format!(
                    "with_inner_fec: unknown kind {other:?} (expected none|ldpc|convolutional|convolutional_k7)"
                )));
            }
        };
        let mut cfg = self.0.clone();
        cfg.inner_fec = inner;
        Ok(Self(cfg))
    }

    /// Selects the receiver's LDPC check-node decode rule:
    /// `"sum_product"` (default, exact), `"min_sum"`, or `"scaled_min_sum"`.
    /// `scale` applies only to `"scaled_min_sum"` (≈0.75 recovers most of the
    /// coding gain). Min-sum trades ≲0.3 dB for ~2× decode throughput.
    #[pyo3(signature = (kind, scale = 0.75))]
    fn with_ldpc_decode_rule(&self, kind: &str, scale: f32) -> PyResult<Self> {
        let rule = match kind {
            "sum_product" | "sum-product" => DecodeRule::SumProduct,
            "min_sum" | "min-sum" => DecodeRule::MinSum,
            "scaled_min_sum" | "scaled-min-sum" => DecodeRule::ScaledMinSum(scale),
            other => {
                return Err(PyValueError::new_err(format!(
                    "with_ldpc_decode_rule: unknown kind {other:?} \
                     (expected sum_product|min_sum|scaled_min_sum)"
                )));
            }
        };
        let mut cfg = self.0.clone();
        cfg.ldpc_decode_rule = rule;
        Ok(Self(cfg))
    }

    /// Sets a rectangular block interleaver on the given stage
    /// (`"inner"` or `"outer"`). `rows`/`cols` = 0 disables it.
    #[pyo3(signature = (stage, rows, cols))]
    fn with_interleaver(&self, stage: &str, rows: usize, cols: usize) -> PyResult<Self> {
        let il = if rows == 0 || cols == 0 {
            InterleaverKind::None
        } else {
            InterleaverKind::Block { rows, cols }
        };
        self.set_interleaver(stage, il)
    }

    /// Sets a DVB-T Forney convolutional interleaver on the given stage
    /// (`"inner"` or `"outer"`; DVB-T uses it on `"outer"`). `branches` (`I`) and
    /// `depth` (`M`) default to DVB-T's 12/17. `branches` = 0 disables it.
    #[pyo3(signature = (stage, branches = 12, depth = 17))]
    fn with_conv_interleaver(&self, stage: &str, branches: usize, depth: usize) -> PyResult<Self> {
        let il = if branches == 0 || depth == 0 {
            InterleaverKind::None
        } else {
            InterleaverKind::Convolutional { branches, depth }
        };
        self.set_interleaver(stage, il)
    }

    /// Sets the payload CRC: `"none"`, `"crc16"`, or `"crc32"`.
    fn with_payload_crc(&self, kind: &str) -> PyResult<Self> {
        let mut cfg = self.0.clone();
        cfg.payload_crc = parse_crc(kind)?;
        Ok(Self(cfg))
    }

    /// Sets the header CRC: `"none"`, `"crc16"`, or `"crc32"`.
    fn with_header_crc(&self, kind: &str) -> PyResult<Self> {
        let mut cfg = self.0.clone();
        cfg.header_crc = parse_crc(kind)?;
        Ok(Self(cfg))
    }

    /// Sets the header format: `"orion_sdr"` (default), `"none"`, or `"dvb_tps"`
    /// (DVB-T TPS signalling; decode via the dedicated DVB-T frame path).
    fn with_header_format(&self, kind: &str) -> PyResult<Self> {
        let hf = match kind {
            "orion_sdr" | "orionsdr" => HeaderFormat::OrionSdr,
            "none" | "no_header" => HeaderFormat::NoHeader,
            "dvb_tps" | "dvbtps" => HeaderFormat::DvbTps,
            other => {
                return Err(PyValueError::new_err(format!(
                    "with_header_format: unknown format {other:?} (expected orion_sdr|none|dvb_tps)"
                )));
            }
        };
        let mut cfg = self.0.clone();
        cfg.header_format = hf;
        Ok(Self(cfg))
    }

    /// Sets an additive scrambler. `poly`/`width` define the LFSR; `seed` is a
    /// fixed seed, or pass `per_frame_random = True` for a per-frame seed
    /// carried in the header. `position` is `"before_outer"` (default) or
    /// `"after_inner"`. Pass `poly = 0` to disable scrambling.
    #[pyo3(signature = (poly, width, seed = 1, per_frame_random = false, position = "before_outer"))]
    fn with_scrambler(
        &self,
        poly: u32,
        width: u8,
        seed: u32,
        per_frame_random: bool,
        position: &str,
    ) -> PyResult<Self> {
        let scrambler = if poly == 0 {
            ScramblerKind::None
        } else {
            let seed_mode = if per_frame_random {
                SeedMode::PerFrameRandom
            } else {
                SeedMode::Fixed(seed)
            };
            ScramblerKind::Additive {
                poly,
                width,
                seed: seed_mode,
            }
        };
        let pos = match position {
            "before_outer" => ScramblerPos::BeforeOuterFec,
            "after_inner" => ScramblerPos::AfterInnerFec,
            other => {
                return Err(PyValueError::new_err(format!(
                    "with_scrambler: unknown position {other:?} (expected before_outer|after_inner)"
                )));
            }
        };
        let mut cfg = self.0.clone();
        cfg.scrambler = scrambler;
        cfg.scrambler_pos = pos;
        Ok(Self(cfg))
    }

    /// Selects DVB-T energy dispersal (the exact standard PRBS whitener,
    /// byte-domain, before the outer FEC) as the scrambler.
    fn with_dvb_t_scrambler(&self) -> Self {
        let mut cfg = self.0.clone();
        cfg.scrambler = ScramblerKind::DvbTEnergyDispersal;
        cfg.scrambler_pos = ScramblerPos::BeforeOuterFec;
        Self(cfg)
    }

    /// Validates the frame-layer configuration, raising `ValueError` on an
    /// inconsistent combination.
    fn validate_frame(&self) -> PyResult<()> {
        self.0
            .validate()
            .map_err(|e| PyValueError::new_err(e.to_string()))
    }
}

impl PyOfdmConfig {
    /// Clones the wrapped `OfdmConfig` for the frame-layer bindings.
    pub(crate) fn inner_config(&self) -> OfdmConfig {
        self.0.clone()
    }

    /// Installs `il` on the named stage (`"inner"`/`"outer"`), shared by the
    /// block and convolutional interleaver bindings.
    fn set_interleaver(&self, stage: &str, il: InterleaverKind) -> PyResult<Self> {
        let mut cfg = self.0.clone();
        match stage {
            "inner" => cfg.inner_interleaver = il,
            "outer" => cfg.outer_interleaver = il,
            other => {
                return Err(PyValueError::new_err(format!(
                    "unknown interleaver stage {other:?} (expected inner|outer)"
                )));
            }
        }
        Ok(Self(cfg))
    }
}

fn parse_ldpc_code(s: &str) -> PyResult<LdpcCode> {
    match s {
        "n512r12" => Ok(LdpcCode::N512R12),
        "n576r23" => Ok(LdpcCode::N576R23),
        "n512r34" => Ok(LdpcCode::N512R34),
        other => Err(PyValueError::new_err(format!(
            "unknown LDPC code {other:?} (expected n512r12|n576r23|n512r34)"
        ))),
    }
}

fn parse_puncture_rate(s: &str) -> PyResult<PunctureRate> {
    match s {
        "1/2" => Ok(PunctureRate::R1_2),
        "2/3" => Ok(PunctureRate::R2_3),
        "3/4" => Ok(PunctureRate::R3_4),
        "5/6" => Ok(PunctureRate::R5_6),
        "7/8" => Ok(PunctureRate::R7_8),
        other => Err(PyValueError::new_err(format!(
            "unknown puncture rate {other:?} (expected 1/2|2/3|3/4|5/6|7/8)"
        ))),
    }
}

fn parse_crc(s: &str) -> PyResult<CrcKind> {
    match s {
        "none" => Ok(CrcKind::None),
        "crc16" => Ok(CrcKind::Crc16),
        "crc32" => Ok(CrcKind::Crc32),
        other => Err(PyValueError::new_err(format!(
            "unknown CRC {other:?} (expected none|crc16|crc32)"
        ))),
    }
}

// ── OfdmMod ───────────────────────────────────────────────────────────────────

/// OFDM transmitter: fused mapper + resource-grid mapping + IFFT + cyclic
/// prefix + optional RF upconversion.
///
/// Input: uint8 array of bits (LSB of each byte); consumed
/// `bits_per_ofdm_symbol` at a time, zero-padding a final partial symbol.
/// Output: complex64 IQ array, `samples_per_ofdm_symbol` samples per symbol.
#[pyclass(name = "OfdmMod")]
pub struct PyOfdmMod(OfdmMod);

#[pymethods]
impl PyOfdmMod {
    #[new]
    fn new(cfg: &PyOfdmConfig) -> Self {
        Self(OfdmMod::new(&cfg.0))
    }

    fn modulate<'py>(
        &mut self,
        py: Python<'py>,
        bits: PyReadonlyArray1<'py, u8>,
    ) -> PyResult<Bound<'py, PyArray1<Complex32>>> {
        let input = bits.as_slice()?;
        let iq = self.0.modulate(input);
        Ok(iq.into_pyarray(py))
    }
}

// ── OfdmDemod ─────────────────────────────────────────────────────────────────

/// OFDM receiver: fused cyclic-prefix removal + FFT + channel equalization
/// + resource-grid extraction + hard-decision decoding.
///
/// `equalizer` selects the channel-estimation method: `"training_symbol"`
/// (the default — one estimate per packet, held constant; call
/// `estimate_channel()` once with a demodulated training symbol before
/// `demodulate()`) or `"pilot_interp"` (re-estimated every symbol from
/// in-band pilots, no separate estimation call needed).
///
/// Input: complex64 IQ array, `samples_per_ofdm_symbol` samples per symbol.
/// Output: uint8 array of bits, `bits_per_ofdm_symbol` per symbol.
#[pyclass(name = "OfdmDemod")]
pub struct PyOfdmDemod {
    cfg: OfdmConfig,
    cp_remove: CyclicPrefixRemove,
    fft: FftBlock,
    equalizer: OfdmEqualizer,
    grid_extract: GridExtract,
    decider: OfdmDecider,
    n_fft: usize,
    samples_per_symbol: usize,
    num_data_carriers: usize,
    bits_per_symbol: usize,
}

#[pymethods]
impl PyOfdmDemod {
    #[new]
    #[pyo3(signature = (cfg, equalizer = "training_symbol"))]
    fn new(cfg: &PyOfdmConfig, equalizer: &str) -> PyResult<Self> {
        let method = match equalizer {
            "training_symbol" => EqualizerMethod::TrainingSymbolHold,
            "pilot_interp" => EqualizerMethod::PerSymbolPilotInterp,
            other => {
                return Err(PyValueError::new_err(format!(
                    "OfdmDemod: unknown equalizer {:?} (expected 'training_symbol' or 'pilot_interp')",
                    other
                )));
            }
        };
        let grid = CarrierGrid::from_plan(&cfg.0.carrier_plan);
        let n_fft = cfg.0.carrier_plan.n_fft();
        let cp_len = cfg.0.carrier_plan.cp_len();
        Ok(Self {
            cp_remove: CyclicPrefixRemove::new(n_fft, cp_len),
            fft: FftBlock::new(n_fft),
            equalizer: OfdmEqualizer::new(&cfg.0, method),
            grid_extract: GridExtract::new(grid.clone()),
            decider: OfdmDecider::new(&cfg.0),
            n_fft,
            samples_per_symbol: cfg.0.samples_per_ofdm_symbol(),
            num_data_carriers: grid.num_data_carriers(),
            bits_per_symbol: cfg.0.bits_per_ofdm_symbol(),
            cfg: cfg.0.clone(),
        })
    }

    /// Estimates and holds the channel from one already-demodulated training
    /// symbol's raw IQ (`samples_per_ofdm_symbol` samples, CP included). Only
    /// meaningful for the `"training_symbol"` equalizer; a no-op under
    /// `"pilot_interp"`.
    fn estimate_channel<'py>(
        &mut self,
        training_iq: PyReadonlyArray1<'py, Complex32>,
    ) -> PyResult<()> {
        let input = training_iq.as_slice()?;
        if input.len() < self.samples_per_symbol {
            return Err(PyValueError::new_err(format!(
                "OfdmDemod.estimate_channel: input too short ({} < {})",
                input.len(),
                self.samples_per_symbol
            )));
        }
        let mut time = vec![Complex32::default(); self.n_fft];
        self.cp_remove
            .process(&input[..self.samples_per_symbol], &mut time);
        let mut freq = vec![Complex32::default(); self.n_fft];
        self.fft.process(&time, &mut freq);
        self.equalizer.estimate_from_training_symbol(&freq);
        Ok(())
    }

    fn demodulate<'py>(
        &mut self,
        py: Python<'py>,
        iq: PyReadonlyArray1<'py, Complex32>,
    ) -> PyResult<Bound<'py, PyArray1<u8>>> {
        let (_soft, bits) = self.demodulate_inner(iq.as_slice()?)?;
        Ok(bits.into_pyarray(py))
    }

    /// Like `demodulate()`, but also returns the pre-decision soft symbols
    /// (post-equalization, post-grid-extract), for callers that want to
    /// build an [`PyOfdmRxFrame`] via `build_ofdm_rx_frame`.
    fn demodulate_soft<'py>(
        &mut self,
        py: Python<'py>,
        iq: PyReadonlyArray1<'py, Complex32>,
    ) -> PyResult<SoftDemodulateResult<'py>> {
        let (soft, bits) = self.demodulate_inner(iq.as_slice()?)?;
        Ok((soft.into_pyarray(py), bits.into_pyarray(py)))
    }
}

impl PyOfdmDemod {
    fn demodulate_inner(&mut self, input: &[Complex32]) -> PyResult<(Vec<Complex32>, Vec<u8>)> {
        if input.len() < self.samples_per_symbol {
            return Err(PyValueError::new_err(format!(
                "OfdmDemod.demodulate: input too short ({} < {})",
                input.len(),
                self.samples_per_symbol
            )));
        }

        let mut time = vec![Complex32::default(); self.n_fft];
        self.cp_remove
            .process(&input[..self.samples_per_symbol], &mut time);
        let mut freq = vec![Complex32::default(); self.n_fft];
        self.fft.process(&time, &mut freq);
        let mut equalized = vec![Complex32::default(); self.n_fft];
        self.equalizer.process(&freq, &mut equalized);
        let mut soft = vec![Complex32::default(); self.num_data_carriers];
        self.grid_extract.process(&equalized, &mut soft);
        let mut bits = vec![0u8; self.bits_per_symbol];
        self.decider.process(&soft, &mut bits);

        let _ = &self.cfg;
        Ok((soft, bits))
    }
}

// ── OfdmRxFrame ───────────────────────────────────────────────────────────────

/// Per-packet OFDM RX diagnostics. Fields that require acquisition or
/// equalization stay `None` until the caller has actually run those stages.
#[pyclass(name = "OfdmRxFrame")]
pub struct PyOfdmRxFrame {
    bits: Vec<u8>,
    num_symbols: usize,
    evm_db: Option<f32>,
    cfo_hz: Option<f32>,
    timing_offset_samples: Option<i32>,
    channel_mse: Option<f32>,
}

#[pymethods]
impl PyOfdmRxFrame {
    #[getter]
    fn bits<'py>(&self, py: Python<'py>) -> Bound<'py, PyArray1<u8>> {
        self.bits.clone().into_pyarray(py)
    }
    #[getter]
    fn num_symbols(&self) -> usize {
        self.num_symbols
    }
    #[getter]
    fn evm_db(&self) -> Option<f32> {
        self.evm_db
    }
    #[getter]
    fn cfo_hz(&self) -> Option<f32> {
        self.cfo_hz
    }
    #[getter]
    fn timing_offset_samples(&self) -> Option<i32> {
        self.timing_offset_samples
    }
    #[getter]
    fn channel_mse(&self) -> Option<f32> {
        self.channel_mse
    }
}

/// Builds an [`PyOfdmRxFrame`] from demodulated soft symbols and their
/// corresponding hard-decided bits, mirroring
/// `demodulate::ofdm::build_ofdm_rx_frame`.
#[pyfunction]
#[pyo3(name = "build_ofdm_rx_frame")]
fn py_build_ofdm_rx_frame<'py>(
    cfg: &PyOfdmConfig,
    soft_symbols: PyReadonlyArray1<'py, Complex32>,
    bits: PyReadonlyArray1<'py, u8>,
) -> PyResult<PyOfdmRxFrame> {
    let soft = soft_symbols.as_slice()?;
    let bits_vec = bits.as_slice()?.to_vec();
    let frame = crate::demodulate::ofdm::build_ofdm_rx_frame(&cfg.0, soft, bits_vec);
    Ok(PyOfdmRxFrame {
        bits: frame.bits,
        num_symbols: frame.num_symbols,
        evm_db: frame.evm_db,
        cfo_hz: frame.cfo_hz,
        timing_offset_samples: frame.timing_offset_samples,
        channel_mse: frame.channel_mse,
    })
}

// ── ofdm_sync ─────────────────────────────────────────────────────────────────

/// Searches an OFDM IQ buffer for a repeated-segment preamble match.
///
/// Returns a list of dicts, sorted by descending score:
///   {"start_sample": int, "cfo_hz": float, "integer_cfo_bins": int, "score": float}
///
/// `integer_cfo_bins` is only meaningful (nonzero) if `with_training_symbol`
/// was set when generating the preamble; total CFO is
/// `cfo_hz + integer_cfo_bins * (fs / n_fft)`.
#[pyfunction]
#[pyo3(signature = (iq, fs, num_repeats, repeat_len, search_start, search_end, training_n_fft = None, training_cp_len = None))]
#[allow(clippy::too_many_arguments)] // mirrors the Python-facing signature
fn ofdm_sync<'py>(
    py: Python<'py>,
    iq: PyReadonlyArray1<'py, Complex32>,
    fs: f32,
    num_repeats: usize,
    repeat_len: usize,
    search_start: usize,
    search_end: usize,
    training_n_fft: Option<usize>,
    training_cp_len: Option<usize>,
) -> PyResult<Bound<'py, PyList>> {
    let input = iq.as_slice()?;
    let mut preamble = OfdmPreamble::new(num_repeats, repeat_len);
    if let (Some(n_fft), Some(cp_len)) = (training_n_fft, training_cp_len) {
        preamble = preamble.with_training_symbol(n_fft, cp_len);
    }

    let results = ofdm_sync_fn(input, fs, &preamble, search_start, search_end);

    let list = PyList::empty(py);
    for r in results {
        let d = PyDict::new(py);
        d.set_item("start_sample", r.start_sample)?;
        d.set_item("cfo_hz", r.cfo_hz)?;
        d.set_item("integer_cfo_bins", r.integer_cfo_bins)?;
        d.set_item("score", r.score)?;
        list.append(d)?;
    }
    Ok(list)
}

/// Generates a repeated-segment preamble (plus training symbol, if
/// `training_n_fft`/`training_cp_len` are given) for prepending before OFDM
/// data symbols.
#[pyfunction]
#[pyo3(name = "generate_ofdm_preamble")]
#[pyo3(signature = (cfg, num_repeats, repeat_len, training_n_fft = None, training_cp_len = None))]
fn generate_ofdm_preamble_py<'py>(
    py: Python<'py>,
    cfg: &PyOfdmConfig,
    num_repeats: usize,
    repeat_len: usize,
    training_n_fft: Option<usize>,
    training_cp_len: Option<usize>,
) -> Bound<'py, PyArray1<Complex32>> {
    let mut preamble = OfdmPreamble::new(num_repeats, repeat_len);
    if let (Some(n_fft), Some(cp_len)) = (training_n_fft, training_cp_len) {
        preamble = preamble.with_training_symbol(n_fft, cp_len);
    }
    let iq = generate_ofdm_preamble(&preamble, &cfg.0);
    iq.into_pyarray(py)
}

pub(crate) fn register(m: &Bound<'_, PyModule>) -> PyResult<()> {
    m.add_class::<PyOfdmConfig>()?;
    m.add_class::<PyOfdmMod>()?;
    m.add_class::<PyOfdmDemod>()?;
    m.add_class::<PyOfdmRxFrame>()?;
    m.add_function(wrap_pyfunction!(py_build_ofdm_rx_frame, m)?)?;
    m.add_function(wrap_pyfunction!(ofdm_sync, m)?)?;
    m.add_function(wrap_pyfunction!(generate_ofdm_preamble_py, m)?)?;
    Ok(())
}