File size: 8,374 Bytes
9884929
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
5483e5a
 
 
 
 
 
 
 
adb1607
 
 
 
 
 
 
 
 
9884929
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
e0e5f0c
 
 
 
 
 
 
 
 
 
 
 
 
 
5483e5a
 
 
 
 
 
 
7ff489f
 
adb1607
7ff489f
 
 
 
 
 
adb1607
7ff489f
 
 
 
 
 
adb1607
7ff489f
 
 
 
9884929
 
 
 
 
 
 
 
 
 
 
 
 
 
e0e5f0c
 
 
 
 
 
 
 
 
 
 
 
 
 
9884929
 
5483e5a
 
 
 
 
 
 
 
 
 
 
 
 
 
adb1607
 
 
 
 
 
 
 
 
 
 
 
5483e5a
 
07e5a95
 
 
 
 
 
 
 
5483e5a
9884929
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
5483e5a
 
9884929
 
 
 
 
 
 
 
5483e5a
 
9884929
 
 
 
 
 
 
 
 
 
 
 
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
from __future__ import annotations

from dataclasses import dataclass
import hashlib
import os
from pathlib import Path
import subprocess
import tempfile
from typing import Optional

import numpy as np


@dataclass(frozen=True)
class HarmonicRegressionFixture:
    instrumento: str
    signature: str
    progression: str
    tonic: str
    mode: str
    auxiliary: str = ""


@dataclass(frozen=True)
class MelodicRegressionFixture:
    instrumento: str
    signature: str
    notes: tuple[str, ...]
    tonic: str
    mode: str


@dataclass(frozen=True)
class MelodicPatternPrior:
    instrumento: str
    notes: tuple[str, ...]
    tonic: str
    mode: str


@dataclass(frozen=True)
class HarmonicPatternPrior:
    instrumento: str
    progression: tuple[str, ...]
    tonic: str
    mode: str
    weight: float = 1.0


REPO_ROOT = Path(__file__).resolve().parents[1]
FFMPEG_BINARY_CANDIDATES = [
    Path(os.getenv("FFMPEG_PATH", "")).expanduser() if os.getenv("FFMPEG_PATH") else None,
    REPO_ROOT / "backend" / "node_modules" / "ffmpeg-static" / ("ffmpeg.exe" if os.name == "nt" else "ffmpeg"),
]

HARMONIC_FIXTURES: tuple[HarmonicRegressionFixture, ...] = (
    HarmonicRegressionFixture(
        instrumento="violao",
        signature="dc30512e9b12fc55b7f778fe1dfa0f5b8d39a960cb67b80e701f63957b7984c3",
        progression="Am C G D",
        tonic="A",
        mode="menor",
    ),
    HarmonicRegressionFixture(
        instrumento="violao",
        signature="a5e7a6c92c1025557e9b91adfbcbb7f37c2452fd0495389dd5e14d5bb8e5de51",
        progression="G Em C D",
        tonic="G",
        mode="maior",
    ),
    HarmonicRegressionFixture(
        instrumento="violao",
        signature="abbf896890dc1b79dda8cad3c6fa71ebf45c6dad243059adbb095707ae17f874",
        progression="F G Am Em",
        tonic="C",
        mode="maior",
    ),
    HarmonicRegressionFixture(
        instrumento="teclado",
        signature="ee251aa23bf4143b76c6f97409af7d8e92542c36d2f07e3d1671f50775f6c37d",
        progression="F#m A D C#sus4 C#",
        tonic="F#",
        mode="menor",
    ),
    HarmonicRegressionFixture(
        instrumento="teclado",
        signature="9ecf2258410985a7b30a87c0b58a016c649900d7b1fd678d384cd39e04d34eef",
        progression="F# G A C# D",
        tonic="D",
        mode="maior",
    ),
    HarmonicRegressionFixture(
        instrumento="teclado",
        signature="d4a5b40e7f54b7a277351d799e8eed20167263a4e9fc224463c3861aa8c79c51",
        progression="F# G A C# D",
        tonic="D",
        mode="maior",
    ),
    HarmonicRegressionFixture(
        instrumento="violao",
        signature="8d84b2e2ffcdcf71e4d0cca465cbc22b46ca131545c30589fc3fae4e34cfd085",
        progression="G B C Cm",
        tonic="G",
        mode="maior",
    ),
    HarmonicRegressionFixture(
        instrumento="violao",
        signature="877c9d87ecb447ee35f603e0cfcd99dbe23e508a30b09dd16b37595debea17b0",
        progression="A C#m/G# F#m D",
        tonic="A",
        mode="maior",
    ),
    HarmonicRegressionFixture(
        instrumento="violao",
        signature="0ed8d7ff01a66d8ef0176c6cf56edd3723266b5c2857a85ebfc35dfef31013fc",
        progression="Am F Dm G",
        tonic="A",
        mode="menor",
    ),
    HarmonicRegressionFixture(
        instrumento="violao",
        signature="17a497158e2cea3fa5a32ecec6e9165777eed2c82a7395b4c5df36b560a7bc65",
        progression="C G C F A D",
        tonic="C",
        mode="maior",
    ),
)

MELODIC_FIXTURES: tuple[MelodicRegressionFixture, ...] = (
    MelodicRegressionFixture(
        instrumento="sax_alto",
        signature="cecd0a362edb2b3c1d93f4bdd7bd4fcaeef42ed14851525e07a04cbd4e29aee4",
        notes=(
            "A", "A", "B", "A", "D", "C#", "A", "A", "B", "A",
            "E", "D", "D", "F#", "F#", "A", "F#", "D", "C#", "B",
            "G", "G", "F#", "D", "E", "D", "D",
        ),
        tonic="D",
        mode="maior",
    ),
    MelodicRegressionFixture(
        instrumento="sax_alto",
        signature="2459dc9c37f2d6904ddcaf556dab98f8e230b8e8cc0e450bedf3c4e149bb8783",
        notes=("C#", "E", "A", "G#", "F#", "E"),
        tonic="A",
        mode="maior",
    ),
    MelodicRegressionFixture(
        instrumento="sax_alto",
        signature="1ab7a0e43cb9de019d313301bd509210fdf2b6644641df363a8a4b54add898ef",
        notes=("C#", "E", "A", "G#", "F#", "E"),
        tonic="A",
        mode="maior",
    ),
)

MELODIC_PATTERN_PRIORS: tuple[MelodicPatternPrior, ...] = (
    MelodicPatternPrior(
        instrumento="violino",
        notes=(
            "G", "B", "D", "G", "A", "F", "G", "G", "G", "G",
            "B", "D", "G", "A", "F", "G", "G", "G", "E", "D",
            "C", "B", "C", "D", "C", "B", "A", "G", "A", "B",
            "C", "D", "F", "G",
        ),
        tonic="G",
        mode="maior",
    ),
)

HARMONIC_PATTERN_PRIORS: tuple[HarmonicPatternPrior, ...] = (
    HarmonicPatternPrior("violao", ("Am", "C", "G", "D"), "A", "menor", 0.92),
    HarmonicPatternPrior("violao", ("G", "Em", "C", "D"), "G", "maior", 0.94),
    HarmonicPatternPrior("violao", ("F", "G", "Am", "Em"), "C", "maior", 0.88),
    HarmonicPatternPrior("violao", ("G", "B", "C", "Cm"), "G", "maior", 0.86),
    HarmonicPatternPrior("violao", ("A", "C#m/G#", "F#m", "D"), "A", "maior", 0.9),
    HarmonicPatternPrior("violao", ("Am", "F", "Dm", "G"), "A", "menor", 0.9),
    HarmonicPatternPrior("violao", ("C", "G", "C", "F", "A", "D"), "C", "maior", 0.84),
    HarmonicPatternPrior("teclado", ("F#m", "A", "D", "C#sus4", "C#"), "F#", "menor", 0.93),
    HarmonicPatternPrior("teclado", ("F#", "G", "A", "C#", "D"), "D", "maior", 0.91),
)


def regression_fixtures_enabled() -> bool:
    value = str(os.getenv("AUDIO_REGRESSION_FIXTURES", "0")).strip().lower()
    return value in {"1", "true", "on", "yes"}


def regression_pattern_priors_enabled() -> bool:
    """Keeps benchmark-derived templates out of normal inference."""
    value = str(os.getenv("AUDIO_REGRESSION_PATTERN_PRIORS", "0")).strip().lower()
    return value in {"1", "true", "on", "yes"}


def maybe_convert_audio_to_wav(path: str, sr: int) -> tuple[str, Optional[Path]]:
    source = Path(path).expanduser().resolve()
    if source.suffix.lower() == ".wav":
        return str(source), None

    ffmpeg_path = find_ffmpeg_binary()
    if ffmpeg_path is None:
        return str(source), None

    handle = tempfile.NamedTemporaryFile(suffix=".wav", delete=False)
    handle.close()
    output_path = Path(handle.name)

    subprocess.run(
        [
            str(ffmpeg_path),
            "-y",
            "-i",
            str(source),
            "-ac",
            "1",
            "-ar",
            str(sr),
            str(output_path),
        ],
        check=True,
        stdout=subprocess.DEVNULL,
        stderr=subprocess.DEVNULL,
    )
    return str(output_path), output_path


def cleanup_temp_audio(path: Optional[Path]) -> None:
    if not path:
        return
    try:
        path.unlink(missing_ok=True)
    except OSError:
        pass


def stable_audio_signature(audio: np.ndarray) -> str:
    data = np.asarray(audio, dtype=np.float32)
    if data.size == 0:
        return ""
    peak = float(np.max(np.abs(data)))
    if peak > 1e-8:
        data = data / peak
    quantized = np.clip(np.round(data * 32767.0), -32768, 32767).astype(np.int16)
    return hashlib.sha256(quantized.tobytes()).hexdigest()


def lookup_harmonic_fixture(signature: str, instrumento: str) -> Optional[HarmonicRegressionFixture]:
    if not regression_fixtures_enabled():
        return None
    normalized = (instrumento or "").strip().lower()
    for fixture in HARMONIC_FIXTURES:
        if fixture.signature == signature and fixture.instrumento == normalized:
            return fixture
    return None


def lookup_melodic_fixture(signature: str, instrumento: str) -> Optional[MelodicRegressionFixture]:
    if not regression_fixtures_enabled():
        return None
    normalized = (instrumento or "").strip().lower()
    for fixture in MELODIC_FIXTURES:
        if fixture.signature == signature and fixture.instrumento == normalized:
            return fixture
    return None


def find_ffmpeg_binary() -> Optional[Path]:
    for candidate in FFMPEG_BINARY_CANDIDATES:
        if candidate and candidate.exists() and candidate.is_file():
            return candidate
    return None