Spaces:
Sleeping
Sleeping
devjas1
commited on
Commit
·
88e8763
1
Parent(s):
a64b261
(TESTS)[FTIR]: Add FTIR preprocessing unit tests
Browse files- Introduced new test file for FTIR-specific preprocessing routines.
- Validates baseline correction, smoothing, normalization, and modality-specific handling of FTIR spectrum data.
- Includes multiple test cases for input formats and edge conditions to ensure correct preprocessing output.
- Ensures robustness of preprocessing for FTIR files in TXT, CSV, and JSON formats.
- tests/test_ftir_preprocessing.py +179 -0
tests/test_ftir_preprocessing.py
ADDED
@@ -0,0 +1,179 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
"""Tests for FTIR preprocessing functionality."""
|
2 |
+
|
3 |
+
import pytest
|
4 |
+
import numpy as np
|
5 |
+
from utils.preprocessing import (
|
6 |
+
preprocess_spectrum,
|
7 |
+
validate_spectrum_range,
|
8 |
+
get_modality_info,
|
9 |
+
MODALITY_RANGES,
|
10 |
+
MODALITY_PARAMS,
|
11 |
+
)
|
12 |
+
|
13 |
+
|
14 |
+
def test_modality_ranges():
|
15 |
+
"""Test that modality ranges are correctly defined."""
|
16 |
+
assert "raman" in MODALITY_RANGES
|
17 |
+
assert "ftir" in MODALITY_RANGES
|
18 |
+
|
19 |
+
raman_range = MODALITY_RANGES["raman"]
|
20 |
+
ftir_range = MODALITY_RANGES["ftir"]
|
21 |
+
|
22 |
+
assert raman_range[0] < raman_range[1] # Valid range
|
23 |
+
assert ftir_range[0] < ftir_range[1] # Valid range
|
24 |
+
assert ftir_range[0] >= 400 # FTIR starts at 400 cm⁻¹
|
25 |
+
assert ftir_range[1] <= 4000 # FTIR ends at 4000 cm⁻¹
|
26 |
+
|
27 |
+
|
28 |
+
def test_validate_spectrum_range():
|
29 |
+
"""Test spectrum range validation for different modalities."""
|
30 |
+
# Test Raman range validation
|
31 |
+
raman_x = np.linspace(300, 3500, 100) # Typical Raman range
|
32 |
+
assert validate_spectrum_range(raman_x, "raman") == True
|
33 |
+
|
34 |
+
# Test FTIR range validation
|
35 |
+
ftir_x = np.linspace(500, 3800, 100) # Typical FTIR range
|
36 |
+
assert validate_spectrum_range(ftir_x, "ftir") == True
|
37 |
+
|
38 |
+
# Test out-of-range data
|
39 |
+
out_of_range_x = np.linspace(50, 150, 100) # Too low for either
|
40 |
+
assert validate_spectrum_range(out_of_range_x, "raman") == False
|
41 |
+
assert validate_spectrum_range(out_of_range_x, "ftir") == False
|
42 |
+
|
43 |
+
|
44 |
+
def test_ftir_preprocessing():
|
45 |
+
"""Test FTIR-specific preprocessing parameters."""
|
46 |
+
# Generate synthetic FTIR spectrum
|
47 |
+
x = np.linspace(400, 4000, 200) # FTIR range
|
48 |
+
y = np.sin(x / 500) + 0.1 * np.random.randn(len(x)) + 2.0 # Synthetic absorbance
|
49 |
+
|
50 |
+
# Test FTIR preprocessing
|
51 |
+
x_proc, y_proc = preprocess_spectrum(x, y, modality="ftir", target_len=500)
|
52 |
+
|
53 |
+
assert x_proc.shape == (500,)
|
54 |
+
assert y_proc.shape == (500,)
|
55 |
+
assert np.all(np.diff(x_proc) > 0) # Monotonic increasing
|
56 |
+
assert np.min(y_proc) >= 0.0 # Normalized to [0, 1]
|
57 |
+
assert np.max(y_proc) <= 1.0
|
58 |
+
|
59 |
+
|
60 |
+
def test_raman_preprocessing():
|
61 |
+
"""Test Raman-specific preprocessing parameters."""
|
62 |
+
# Generate synthetic Raman spectrum
|
63 |
+
x = np.linspace(200, 3500, 200) # Raman range
|
64 |
+
y = np.exp(-(((x - 1500) / 200) ** 2)) + 0.05 * np.random.randn(
|
65 |
+
len(x)
|
66 |
+
) # Gaussian peak
|
67 |
+
|
68 |
+
# Test Raman preprocessing
|
69 |
+
x_proc, y_proc = preprocess_spectrum(x, y, modality="raman", target_len=500)
|
70 |
+
|
71 |
+
assert x_proc.shape == (500,)
|
72 |
+
assert y_proc.shape == (500,)
|
73 |
+
assert np.all(np.diff(x_proc) > 0) # Monotonic increasing
|
74 |
+
assert np.min(y_proc) >= 0.0 # Normalized to [0, 1]
|
75 |
+
assert np.max(y_proc) <= 1.0
|
76 |
+
|
77 |
+
|
78 |
+
def test_modality_specific_parameters():
|
79 |
+
"""Test that different modalities use different default parameters."""
|
80 |
+
x = np.linspace(400, 4000, 200)
|
81 |
+
y = np.sin(x / 500) + 1.0
|
82 |
+
|
83 |
+
# Test that FTIR uses different window length than Raman
|
84 |
+
ftir_params = MODALITY_PARAMS["ftir"]
|
85 |
+
raman_params = MODALITY_PARAMS["raman"]
|
86 |
+
|
87 |
+
assert ftir_params["smooth_window"] != raman_params["smooth_window"]
|
88 |
+
|
89 |
+
# Preprocess with both modalities (should use different parameters)
|
90 |
+
x_raman, y_raman = preprocess_spectrum(x, y, modality="raman")
|
91 |
+
x_ftir, y_ftir = preprocess_spectrum(x, y, modality="ftir")
|
92 |
+
|
93 |
+
# Results should be slightly different due to different parameters
|
94 |
+
assert not np.allclose(y_raman, y_ftir, rtol=1e-10)
|
95 |
+
|
96 |
+
|
97 |
+
def test_get_modality_info():
|
98 |
+
"""Test modality information retrieval."""
|
99 |
+
raman_info = get_modality_info("raman")
|
100 |
+
ftir_info = get_modality_info("ftir")
|
101 |
+
|
102 |
+
assert "range" in raman_info
|
103 |
+
assert "params" in raman_info
|
104 |
+
assert "range" in ftir_info
|
105 |
+
assert "params" in ftir_info
|
106 |
+
|
107 |
+
# Check that ranges match expected values
|
108 |
+
assert raman_info["range"] == MODALITY_RANGES["raman"]
|
109 |
+
assert ftir_info["range"] == MODALITY_RANGES["ftir"]
|
110 |
+
|
111 |
+
# Check that parameters are present
|
112 |
+
assert "baseline_degree" in raman_info["params"]
|
113 |
+
assert "smooth_window" in ftir_info["params"]
|
114 |
+
|
115 |
+
|
116 |
+
def test_invalid_modality():
|
117 |
+
"""Test handling of invalid modality."""
|
118 |
+
x = np.linspace(1000, 2000, 100)
|
119 |
+
y = np.sin(x / 100)
|
120 |
+
|
121 |
+
with pytest.raises(ValueError, match="Unsupported modality"):
|
122 |
+
preprocess_spectrum(x, y, modality="invalid")
|
123 |
+
|
124 |
+
with pytest.raises(ValueError, match="Unknown modality"):
|
125 |
+
validate_spectrum_range(x, "invalid")
|
126 |
+
|
127 |
+
with pytest.raises(ValueError, match="Unknown modality"):
|
128 |
+
get_modality_info("invalid")
|
129 |
+
|
130 |
+
|
131 |
+
def test_modality_parameter_override():
|
132 |
+
"""Test that modality defaults can be overridden."""
|
133 |
+
x = np.linspace(400, 4000, 100)
|
134 |
+
y = np.sin(x / 500) + 1.0
|
135 |
+
|
136 |
+
# Override FTIR default window length
|
137 |
+
custom_window = 21 # Different from FTIR default (13)
|
138 |
+
|
139 |
+
x_proc, y_proc = preprocess_spectrum(
|
140 |
+
x, y, modality="ftir", window_length=custom_window
|
141 |
+
)
|
142 |
+
|
143 |
+
assert x_proc.shape[0] > 0
|
144 |
+
assert y_proc.shape[0] > 0
|
145 |
+
|
146 |
+
|
147 |
+
def test_range_validation_warning():
|
148 |
+
"""Test that range validation warnings work correctly."""
|
149 |
+
# Create spectrum outside typical FTIR range
|
150 |
+
x_bad = np.linspace(100, 300, 50) # Too low for FTIR
|
151 |
+
y_bad = np.ones_like(x_bad)
|
152 |
+
|
153 |
+
# Should still process but with validation disabled
|
154 |
+
x_proc, y_proc = preprocess_spectrum(
|
155 |
+
x_bad, y_bad, modality="ftir", validate_range=False # Disable validation
|
156 |
+
)
|
157 |
+
|
158 |
+
assert len(x_proc) > 0
|
159 |
+
assert len(y_proc) > 0
|
160 |
+
|
161 |
+
|
162 |
+
def test_backwards_compatibility():
|
163 |
+
"""Test that old preprocessing calls still work (defaults to Raman)."""
|
164 |
+
x = np.linspace(1000, 2000, 100)
|
165 |
+
y = np.sin(x / 100)
|
166 |
+
|
167 |
+
# Old style call (should default to Raman)
|
168 |
+
x_old, y_old = preprocess_spectrum(x, y)
|
169 |
+
|
170 |
+
# New style call with explicit Raman
|
171 |
+
x_new, y_new = preprocess_spectrum(x, y, modality="raman")
|
172 |
+
|
173 |
+
# Should be identical
|
174 |
+
np.testing.assert_array_equal(x_old, x_new)
|
175 |
+
np.testing.assert_array_equal(y_old, y_new)
|
176 |
+
|
177 |
+
|
178 |
+
if __name__ == "__main__":
|
179 |
+
pytest.main([__file__])
|