Tighten exactness fixtures for analysis and reports

examples/readme_examples/example1_multitaper_and_spectrogram.py

@@ -7,7 +7,7 @@
 import numpy as np
 from scipy.signal import spectrogram
 
-from nstat.compat.matlab import SignalObj
+from nstat.SignalObj import SignalObj
 
 
 def _fallback_multitaper_psd(signal: np.ndarray, fs_hz: float) -> tuple[np.ndarray, np.ndarray]:
@@ -31,7 +31,7 @@ def main() -> None:
     time = np.arange(0.0, duration_s, dt, dtype=float)
 
     signal = np.sin(2.0 * np.pi * f0_hz * time)
-    sig_obj = SignalObj(time=time, data=signal, name="sine_signal", units="a.u.")
+    sig_obj = SignalObj(time=time, data=signal, name="sine_signal", yunits="a.u.")
 
     try:
         freq_hz, psd = sig_obj.MTMspectrum()

nstat/analysis.py

@@ -44,11 +44,22 @@ def _as_neuron_indices(trial: Trial, neuron_selector) -> list[int]:
     raise TypeError("neuron selector must be a MATLAB-style one-based index, name, or sequence of either")
 
 
-def _restore_trial_partition(trial: Trial, original_partition: np.ndarray) -> None:
+def _restore_trial_partition(trial: Trial, original_partition: np.ndarray, original_window: np.ndarray | None = None) -> None:
     trial.restoreToOriginal()
     if original_partition.size:
         trial.setTrialPartition(original_partition)
-        trial.setTrialTimesFor("training")
+        if original_window is None or original_window.size != 2:
+            trial.setTrialTimesFor("training")
+            return
+        training = original_partition[:2] if original_partition.size >= 2 else None
+        validation = original_partition[2:4] if original_partition.size >= 4 else None
+        if training is not None and training.size == 2 and np.allclose(original_window, training, rtol=0.0, atol=1e-12):
+            trial.setTrialTimesFor("training")
+        elif validation is not None and validation.size == 2 and np.allclose(original_window, validation, rtol=0.0, atol=1e-12):
+            trial.setTrialTimesFor("validation")
+        else:
+            trial.setMinTime(float(original_window[0]))
+            trial.setMaxTime(float(original_window[1]))
 
 
 def _time_rescaled_z(counts: np.ndarray, lam_per_bin: np.ndarray) -> np.ndarray:
@@ -154,7 +165,7 @@ def GLMFit(
         lambdaIndex: int,
         Algorithm: str = "GLM",
         *,
-        l2: float = 1e-6,
+        l2: float = 0.0,
         max_iter: int = 120,
     ):
         algorithm = str(Algorithm or "GLM").upper()
@@ -243,13 +254,14 @@ def run_analysis_for_neuron(
         config_collection: ConfigCollection,
         *,
         algorithm: str = "GLM",
-        l2: float = 1e-6,
+        l2: float = 0.0,
         max_iter: int = 120,
     ) -> FitResult:
         if neuron_index < 0:
             raise IndexError("neuron_index must be >= 0")
 
         original_partition = np.asarray(trial.getTrialPartition(), dtype=float).reshape(-1)
+        original_window = np.asarray([trial.minTime, trial.maxTime], dtype=float).reshape(-1)
         neuron_number = int(neuron_index) + 1
         labels: list[list[str]] = []
         lambda_parts: list[Covariate] = []
@@ -272,7 +284,10 @@ def run_analysis_for_neuron(
             spike_train.setName(str(neuron_number))
 
         for cfg_index in range(1, config_collection.numConfigs + 1):
-            _restore_trial_partition(trial, original_partition)
+            trial.restoreToOriginal()
+            if original_partition.size:
+                trial.setTrialPartition(original_partition)
+                trial.setTrialTimesFor("training")
             config_collection.setConfig(trial, cfg_index)
 
             current_labels = trial.getLabelsFromMask(neuron_number)
@@ -326,7 +341,7 @@ def run_analysis_for_neuron(
         for part in lambda_parts[1:]:
             merged_lambda = merged_lambda.merge(part)
 
-        _restore_trial_partition(trial, original_partition)
+        _restore_trial_partition(trial, original_partition, original_window)
         fit_result = FitResult(
             spike_train,
             labels,
@@ -357,7 +372,7 @@ def run_analysis_for_all_neurons(
         config_collection: ConfigCollection,
         *,
         algorithm: str = "GLM",
-        l2: float = 1e-6,
+        l2: float = 0.0,
         max_iter: int = 120,
     ) -> list[FitResult]:
         out: list[FitResult] = []
@@ -433,8 +448,15 @@ def computeFitResidual(nspikeObj, lambdaInput: Covariate, windowSize: float = 0.
         nCopy.setMinTime(lambdaInput.minTime)
         nCopy.setMaxTime(lambdaInput.maxTime)
 
-        sumSpikes = nCopy.getSigRep(windowSize)
+        # MATLAB's static Analysis.computeFitResidual ultimately operates on
+        # the resampled spike-count grid, even when a finer windowSize is
+        # requested. Preserve that canonical helper behavior here.
+        sumSpikes = nCopy.getSigRep(1.0 / float(nCopy.sampleRate), float(nCopy.minTime), float(nCopy.maxTime))
         windowTimes = np.linspace(float(nCopy.minTime), float(nCopy.maxTime), sumSpikes.time.size, dtype=float)
+        if np.isfinite(windowSize) and windowSize > 0:
+            origin = float(nCopy.minTime)
+            windowTimes = origin + np.round((windowTimes - origin) / float(windowSize)) * float(windowSize)
+            windowTimes = np.round(windowTimes, decimals=12)
         lambdaInt = lambdaInput.integral()
         lambdaIntVals = (
             lambdaInt.getValueAt(windowTimes[1:]).reshape(-1, lambdaInt.dimension)
@@ -465,8 +487,7 @@ def KSPlot(fitResults: FitResult, DTCorrection: int = 1, makePlot: int = 1):
 
     @staticmethod
     def plotFitResidual(fitResults: FitResult, windowSize: float = 0.01, makePlot: int = 1):
-        del windowSize
-        fitResults.computeFitResidual()
+        fitResults.computeFitResidual(window_size=windowSize)
         return fitResults.plotResidual() if makePlot else []
 
     @staticmethod

nstat/class_fidelity.py

@@ -8,6 +8,41 @@
 
 
 EXPECTED_RUNTIME_MEMBERS: dict[str, tuple[str, ...]] = {
+    "nstat.SignalObj": (
+        "shift",
+        "shiftMe",
+        "alignTime",
+        "power",
+        "sqrt",
+        "xcov",
+        "periodogram",
+        "MTMspectrum",
+        "spectrogram",
+        "plotVariability",
+        "plotAllVariability",
+        "plotPropsSet",
+        "areDataLabelsEmpty",
+        "isLabelPresent",
+        "convertNamesToIndices",
+        "clearPlotProps",
+    ),
+    "nstat.Trial": (
+        "findMinSampleRate",
+        "getAllLabels",
+        "getDesignMatrix",
+        "getNumHist",
+        "getEnsCovMatrix",
+        "getTrialPartition",
+        "plotCovariates",
+        "plotRaster",
+        "toStructure",
+        "fromStructure",
+    ),
+    "nstat.nstColl": (
+        "psthBars",
+        "estimateVarianceAcrossTrials",
+        "ssglm",
+    ),
     "nstat.Analysis": (
         "GLMFit",
         "RunAnalysisForNeuron",