dsfb-rf 1.0.1

DSFB-RF Structural Semiotics Engine for RF Signal Monitoring - A Deterministic, Non-Intrusive Observer Layer for Typed Structural Interpretation of IQ Residual Streams in Electronic Warfare, Spectrum Monitoring, and Cognitive Radio
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

#!/usr/bin/env python3
"""
DeepSense-6G slice generator.

Primary path: ``extract_scenario23_slice`` reads a user-downloaded
Scenario-23 UAV-mmWave zip (``scenario23_dev_w_resources.zip`` from
deepsense6g.net) and emits a ≤ 2 MB HDF5 head slice with
``[time, beam]`` mmWave power + UAV telemetry.

Fallback path: ``generate`` — schema-preserving multimodal proxy (mmWave IQ
block + GPS trace + camera hash surrogate) emitted only when the local zip
is absent, loudly stamped ``[SYNTHETIC PROXY]``.
"""

from __future__ import annotations

import hashlib
import io
import json
import time
import zipfile
from pathlib import Path

import numpy as np

N_MMWAVE_SAMPLES = 65_536   # 0.5 MB cf32
N_GPS_SAMPLES = 1200
SCENARIO23_HEAD_N = 1000    # first 1000 UAV samples: ~0.26 MB mmWave + ~20 KB telemetry


def generate(
    out_bin: Path,
    out_meta: Path,
    rng: np.random.Generator,
    n_mmwave: int = N_MMWAVE_SAMPLES,
    n_gps: int = N_GPS_SAMPLES,
) -> None:
    n_mmwave = int(n_mmwave)
    n_gps = int(n_gps)

    # mmWave IQ block: Rayleigh + slow fading.
    t = np.arange(n_mmwave)
    h = (rng.standard_normal(n_mmwave) + 1j * rng.standard_normal(n_mmwave)).astype(np.complex64) / np.sqrt(2.0)
    slow = 0.5 + 0.5 * np.cos(2 * np.pi * t * 1e-4 + rng.uniform(0, 2 * np.pi))
    h = h * slow.astype(np.float32)
    cf32 = np.empty(n_mmwave * 2, dtype=np.float32)
    cf32[0::2] = h.real.astype(np.float32)
    cf32[1::2] = h.imag.astype(np.float32)

    # Fake deterministic camera hash (we never fabricate pixels).
    cam_hash = hashlib.sha256(cf32.tobytes()[:4096]).hexdigest()

    # NMEA-like GPS trace — linear walk around a fixed lat/lon.
    lat0, lon0 = 42.3399, -71.0869  # Northeastern-ish
    lat = lat0 + (np.arange(n_gps) * 1e-6 + rng.standard_normal(n_gps).cumsum() * 1e-7)
    lon = lon0 + (np.arange(n_gps) * 1e-6 + rng.standard_normal(n_gps).cumsum() * 1e-7)
    gps_rows = [f"{i*0.1:.1f},{lat[i]:.6f},{lon[i]:.6f}" for i in range(n_gps)]

    payload = cf32.tobytes() + b"\n---GPS---\n" + "\n".join(gps_rows).encode("utf-8")
    out_bin.write_bytes(payload)

    meta = {
        "dsfb_rf:provenance": "synthetic-proxy",
        "dsfb_rf:proxy_for": "DeepSense-6G scenario sample (deepsense6g.net)",
        "dsfb_rf:source_model": "Rayleigh mmWave fading + slow amplitude modulation + linear GPS walk + sha256 camera hash surrogate",
        "dsfb_rf:generated_at": time.strftime("%Y-%m-%dT%H:%M:%SZ", time.gmtime()),
        "modalities": {
            "mmwave_iq": {"datatype": "cf32", "n_samples": n_mmwave},
            "gps_trace": {"n_samples": n_gps, "format": "CSV: t_s,lat_deg,lon_deg"},
            "camera_hash": cam_hash,
        },
        "notice": "[SYNTHETIC PROXY] No real imagery fabricated; camera modality represented by a deterministic hash surrogate.",
    }
    out_meta.write_text(json.dumps(meta, indent=2) + "\n", encoding="utf-8")


def _read_scalar_txt(zf: zipfile.ZipFile, path: str) -> float:
    try:
        return float(zf.read(path).decode("ascii").strip().splitlines()[0])
    except (KeyError, ValueError):
        return float("nan")


def extract_scenario23_slice(
    zip_path: Path,
    out_h5: Path,
    out_meta: Path,
    n_head: int = SCENARIO23_HEAD_N,
) -> dict:
    """
    Extract a ≤ 2 MB head slice from a user-downloaded Scenario 23 zip.

    Schema of the emitted HDF5 file:
        /mmwave_power[time, beam]  float32   shape = (n_head, 64)
        /best_beam_index[time]     int16     shape = (n_head,)
        /altitude[time]            float32   shape = (n_head,)
        /speed[time]               float32   shape = (n_head,)
        /pitch[time]               float32   shape = (n_head,)
        /roll[time]                float32   shape = (n_head,)
        /distance[time]            float32   shape = (n_head,)
        /height[time]              float32   shape = (n_head,)

    Provenance stamped as HDF5 root attributes so a reviewer can inspect via
    ``h5dump -A``. Returns a small manifest dict for SLICE_MANIFEST.json.
    """
    import csv  # stdlib; optional dependency isolation

    try:
        import h5py  # type: ignore
    except ImportError as e:
        raise RuntimeError(
            "h5py is required for DeepSense-6G Scenario 23 real slice extraction; "
            "install via `pip install h5py`"
        ) from e

    zip_path = Path(zip_path)
    if not zip_path.is_file():
        raise FileNotFoundError(f"Scenario 23 zip not found: {zip_path}")

    out_h5 = Path(out_h5)
    out_meta = Path(out_meta)

    with zipfile.ZipFile(zip_path, "r") as zf:
        csv_bytes = zf.read("scenario23_dev/scenario23.csv")
        reader = csv.DictReader(io.StringIO(csv_bytes.decode("utf-8")))
        rows = []
        for i, row in enumerate(reader):
            if i >= n_head:
                break
            rows.append(row)
        if not rows:
            raise RuntimeError("scenario23.csv is empty")

        n = len(rows)
        mmwave = np.zeros((n, 64), dtype=np.float32)
        best_beam = np.zeros(n, dtype=np.int16)
        altitude = np.zeros(n, dtype=np.float32)
        speed = np.zeros(n, dtype=np.float32)
        pitch = np.zeros(n, dtype=np.float32)
        roll = np.zeros(n, dtype=np.float32)
        distance = np.zeros(n, dtype=np.float32)
        height = np.zeros(n, dtype=np.float32)

        for k, row in enumerate(rows):
            mm_path = "scenario23_dev/" + row["unit1_pwr_60ghz"].lstrip("./")
            try:
                mm = zf.read(mm_path).decode("ascii").strip().splitlines()
                vec = np.array([float(x) for x in mm], dtype=np.float32)
                if vec.size != 64:
                    raise ValueError(f"expected 64 beams, got {vec.size} in {mm_path}")
                mmwave[k] = vec
            except (KeyError, ValueError) as err:
                raise RuntimeError(f"mmWave read failed at row {k}: {err}") from err

            try:
                best_beam[k] = int(row["unit1_beam_index"])
            except (KeyError, ValueError):
                best_beam[k] = -1

            altitude[k] = _read_scalar_txt(zf, "scenario23_dev/" + row["unit2_altitude"].lstrip("./"))
            speed[k] = _read_scalar_txt(zf, "scenario23_dev/" + row["unit2_speed"].lstrip("./"))
            pitch[k] = _read_scalar_txt(zf, "scenario23_dev/" + row["unit2_pitch"].lstrip("./"))
            roll[k] = _read_scalar_txt(zf, "scenario23_dev/" + row["unit2_roll"].lstrip("./"))
            distance[k] = _read_scalar_txt(zf, "scenario23_dev/" + row["unit2_distance"].lstrip("./"))
            height[k] = _read_scalar_txt(zf, "scenario23_dev/" + row["unit2_height"].lstrip("./"))

    with h5py.File(out_h5, "w") as f:
        f.create_dataset("mmwave_power", data=mmwave, compression="gzip", compression_opts=4)
        f.create_dataset("best_beam_index", data=best_beam, compression="gzip", compression_opts=4)
        f.create_dataset("altitude", data=altitude, compression="gzip", compression_opts=4)
        f.create_dataset("speed", data=speed, compression="gzip", compression_opts=4)
        f.create_dataset("pitch", data=pitch, compression="gzip", compression_opts=4)
        f.create_dataset("roll", data=roll, compression="gzip", compression_opts=4)
        f.create_dataset("distance", data=distance, compression="gzip", compression_opts=4)
        f.create_dataset("height", data=height, compression="gzip", compression_opts=4)
        f.attrs["dsfb_rf:provenance"] = "real-local-zip"
        f.attrs["dsfb_rf:source"] = (
            "DeepSense-6G Scenario 23 UAV mmWave "
            "(deepsense6g.net/scenarios/scenario-23) — user-downloaded head slice"
        )
        f.attrs["dsfb_rf:parent_zip"] = zip_path.name
        f.attrs["dsfb_rf:parent_zip_sha256"] = hashlib.sha256(zip_path.read_bytes()[:64]).hexdigest() + "-first64B"
        f.attrs["dsfb_rf:head_n_samples"] = n
        f.attrs["dsfb_rf:extracted_at"] = time.strftime("%Y-%m-%dT%H:%M:%SZ", time.gmtime())
        f.attrs["dsfb_rf:schema"] = (
            "mmwave_power[time,beam] float32 (N,64); best_beam_index int16 (N,); "
            "telemetry (altitude, speed, pitch, roll, distance, height) float32 (N,)"
        )

    meta = {
        "dsfb_rf:provenance": "real-local-zip",
        "dsfb_rf:source": "DeepSense-6G Scenario 23 UAV mmWave (deepsense6g.net)",
        "dsfb_rf:parent_zip": zip_path.name,
        "dsfb_rf:head_n_samples": n,
        "dsfb_rf:schema": {
            "mmwave_power": "float32 (N,64) — 60 GHz beamformed power per beam",
            "best_beam_index": "int16 (N,) — ground-truth best-beam label",
            "altitude": "float32 (N,) — UAV altitude (meters)",
            "speed": "float32 (N,) — UAV speed (m/s)",
            "pitch": "float32 (N,) — UAV pitch (rad)",
            "roll": "float32 (N,) — UAV roll (rad)",
            "distance": "float32 (N,) — UAV-BS distance (m)",
            "height": "float32 (N,) — UAV height AGL (m)",
        },
        "dsfb_rf:extracted_at": time.strftime("%Y-%m-%dT%H:%M:%SZ", time.gmtime()),
        "notice": (
            "Real DeepSense-6G Scenario 23 (UAV mmWave) head slice; "
            "not a benchmark reproduction. Contextual residual-trace exhibit only."
        ),
    }
    out_meta.write_text(json.dumps(meta, indent=2) + "\n", encoding="utf-8")
    return meta