SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 1911.03221 · 2019

Dreem Open Datasets: Multi-Scored Sleep Datasets to compare Human and Automated sleep staging

arXiv · PDF · Open in the Atlas

Code that ran

19 of 22 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (0 of them constructed an object rather than computing a result; 0 honoured the contract we drafted, 0 violated it and 16 have no contract verdict; 3 failures of our instrument; 3 no recorded run), measured 2026-11-22; repository links from the Papers with Code archive attributed under CC BY-SA, no affiliation; bodies inline only under a permissive licence as GitHub detects it. Not a reproduction of the paper's results.

"Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results. This record, dated, as JSON: /api/ran/1911.03221.json · badge: /api/ran/1911.03221.svg

Each repository's role is a label: a classifier's output recorded on the link, or, where no role is recorded, what the Papers with Code archive marked. The line under it names its basis. Neither says who wrote the code.

RepositoryRoleRan
Dreem-Organization/dreem-learning-open canonical (classifier label)
classifier basis: census
16 ran · 3 no recorded run
Dreem-Organization/dreem-learning-evaluation canonical (classifier label)
classifier basis: classifier_2025
3 ran · 0 no recorded run
FunctionStatusWhere it lives
affine Ran
ran on a synthesized input
Dreem-Organization/dreem-learning-open/dreem_learning_open/models/modulo_net/normalization.py
code served (permissive licence) · get_code("48b2a11244c8eec9")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned Tensor, ndim 2, dtype family float, shape (2, 4) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import numpy as np

torch.manual_seed(0)

# Create tiny tensors
x = torch.randn(2, 4)
gain = torch.randn(2, 4)
bias = 0

ARGS = {
    "x": x,
    "gain": gain,
    "bias": bias,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 2, dtype family float, shape (2, 4) — shape (2, 4) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
augment_data Ran
ran on a synthesized input
Dreem-Organization/dreem-learning-open/dreem_learning_open/datasets/data_augmentation.py
code served (permissive licence) · get_code("8ae689bec5854251")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned dict. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import numpy as np

# Set seed for reproducibility
np.random.seed(0)

# Create sample data with shape (batch, channels, height)
# Using distinct dimensions: batch=2, channels=4, height=3
sample_data = np.random.randn(2, 4, 3)

# Construct the data dict with a group
data = {
    'group1': sample_data.copy()
}

# Construct augmentation pipeline
augmentation_pipeline = [
    {
        'name': 'group1',
        'processing': [
            {
                'type': 'kill_channel',
                'args': {'p': 0.5}
            }
        ]
    }
]

ARGS = {
    'data': data,
    'augmentation_pipeline': augmentation_pipeline
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type dict — a dict output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a dict output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
build_consensus_hypnogram Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
Dreem-Organization/dreem-learning-evaluation/evaluation.py
code served (permissive licence) · get_code("918adff8add965aa")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1) and returned tuple, ndim 2, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import numpy as np

# Construct valid arguments for build_consensus_hypnogram
# The function expects ranked_hypnograms_consensus: a list of hypnograms
# Each hypnogram is a sequence of sleep stage labels (0-4, representing different stages)
# The function will add 1 to these values internally, so we use 0-4 range

# Create 3 ranked hypnograms, each with 5 epochs
# Each epoch has a sleep stage label from 0-4
hypnogram_1 = [0, 1, 2, 2, 3]  # Reference hypnogram (used for tie-breaking)
hypnogram_2 = [0, 1, 2, 3, 3]
hypnogram_3 = [0, 1, 1, 2, 3]

ranked_hypnograms_consensus = [hypnogram_1, hypnogram_2, hypnogram_3]

ARGS = {
    "ranked_hypnograms_consensus": ranked_hypnograms_consensus
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple, ndim 2, dtype family float — a tuple output has no array shape of its own; none is stated
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
clip Ran
ran on a synthesized input
Dreem-Organization/dreem-learning-open/dreem_learning_open/models/modulo_net/normalization.py
code served (permissive licence) · get_code("867d2c4a9b3665e3")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned Tensor, ndim 2, dtype family float, shape (2, 4) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch

torch.manual_seed(0)

# The clip function takes two parameters:
# - x: a tensor to clamp
# - value: the clamp bound (absolute value)

x = torch.randn(2, 4)
value = 0.5

ARGS = {
    "x": x,
    "value": value,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 2, dtype family float, shape (2, 4) — shape (2, 4) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
cohen_kappa_custom Ran
ran on a synthesized input
Dreem-Organization/dreem-learning-open/dreem_learning_open/utils/scores.py
code served (permissive licence) · get_code("7e53985f41c0c2e4")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned float, ndim 0, dtype family float, shape () as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import numpy as np
from sklearn.metrics import cohen_kappa_score, f1_score, accuracy_score

def cohen_kappa_custom(true, pred):
    labels = list(set([0, 1, 2, 3, 4]).intersection(set(true)))
    return cohen_kappa_score(true, pred, labels=labels)

# Create small sample data for true and pred
# Both should be 1D arrays/lists of class labels (0-4)
true = np.array([0, 1, 2, 3, 4, 0, 1, 2])
pred = np.array([0, 1, 2, 2, 4, 0, 1, 3])

ARGS = {
    "true": true,
    "pred": pred
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, sklearn — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type float, ndim 0, dtype family float, shape () — shape () as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
compute_memmap_hash Ran
ran on a synthesized input
Dreem-Organization/dreem-learning-open/dreem_learning_open/preprocessings/h5_to_memmap.py
code served (permissive licence) · get_code("856fb590eba3e24d")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned str. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import hashlib
import json

# The function compute_memmap_hash takes a memmap_description parameter
# and converts it to JSON, then computes a SHA1 hash of it.
# memmap_description should be something JSON-serializable (dict, list, etc.)

memmap_description = {
    "shape": [1024, 512],
    "dtype": "float32",
    "offset": 0,
    "filename": "data.bin",
    "mode": "r",
    "version": 1
}

ARGS = {
    "memmap_description": memmap_description
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type str — a str output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a str output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
compute_night_metrics Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
Dreem-Organization/dreem-learning-evaluation/evaluation.py
code served (permissive licence) · get_code("302691ee73a2e048")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1) and returned OrderedDict. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import numpy as np
from collections import OrderedDict

# Analyze the function: compute_night_metrics takes a hypnogram parameter
# A hypnogram is iterated over with enumerate, and each stage is compared to integers (0, 1, 2, 3, 4)
# It's a sequence of sleep stage labels

# Create a valid hypnogram - a list of integers representing sleep stages
# Stage 0 = Wake, 1 = N1, 2 = N2, 3 = N3, 4 = REM
# Need enough data to trigger the conditions in the function
# - Need at least 3 consecutive non-wake stages for 'sod' (sleep onset)
# - Need at least 10 consecutive non-wake stages for 'lps' (light period start)

hypnogram = [
    0, 0, 0,  # Initial wake
    1, 1, 1,  # N1 (3 consecutive non-wake, triggers sod at index 5)
    2, 2, 2, 2, 2, 2, 2,  # N2 (reaches 10 consecutive non-wake at index 12, triggers lps)
    3, 3, 3,  # N3
    4, 4, 4, 4,  # REM
    0, 0,  # WASO (wake after sleep onset)
    2, 2, 2,  # Back to N2
    0,  # Final wake
]

ARGS = {
    'hypnogram': hypnogram
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
numpy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type OrderedDict — a OrderedDict output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a OrderedDict output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
compute_soft_agreement Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
Dreem-Organization/dreem-learning-evaluation/evaluation.py
code served (permissive licence) · get_code("f92cfbe72d68e6b8")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1) and returned float64, ndim 0, dtype family float, shape () as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import numpy as np

# Construct valid arguments for compute_soft_agreement
# The function expects:
# - hypnogram: a sequence of sleep stage labels (integers -1 to 4, representing stages 0-5 after +1)
# - hypnograms_consensus: a list of such sequences from multiple raters

# Create a single hypnogram (ground truth or reference)
# Sleep stages are typically: -1 (unknown/artifact), 0 (wake), 1 (N1), 2 (N2), 3 (N3), 4 (REM)
# After +1 indexing in the function, these map to indices 0-5
hypnogram = np.array([0, 1, 2, 2, 3, 4, 1, 0], dtype=int)

# Create multiple consensus hypnograms from different raters
# Each should be the same length as the reference hypnogram
hypnograms_consensus = [
    np.array([0, 1, 2, 2, 3, 4, 1, 0], dtype=int),
    np.array([0, 1, 1, 2, 3, 4, 1, 0], dtype=int),
    np.array([0, 0, 2, 2, 3, 4, 2, 0], dtype=int),
]

ARGS = {
    "hypnogram": hypnogram,
    "hypnograms_consensus": hypnograms_consensus,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type float64, ndim 0, dtype family float, shape () — shape () as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
cyclic_index_window Ran
ran on a synthesized input
Dreem-Organization/dreem-learning-open/dreem_learning_open/preprocessings/epoch_features_processing.py
code served (permissive licence) · get_code("248d2cceb9938a91")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned ndarray, ndim 2, dtype family float, shape (5, 5) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import numpy as np

# Create a small signal: 1200 samples at 40 Hz sampling frequency
# This gives us 30 windows with increment_duration=30
signal = np.random.randn(1200).astype(np.float32)

# Signal properties with sampling frequency
signal_properties = {
    'fs': 40  # 40 Hz sampling frequency
}

# Define arguments for cyclic_index_window
ARGS = {
    'signal': signal,
    'signal_properties': signal_properties,
    'increment_duration': 30,
    'cycles_length': [10, 30, 60, 90, 120],
    'padding_duration': 60
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 2, dtype family float, shape (5, 5) — shape (5, 5) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
f1_custom Ran
ran on a synthesized input
Dreem-Organization/dreem-learning-open/dreem_learning_open/utils/scores.py
code served (permissive licence) · get_code("a377531334afcb0a")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned float64, ndim 0, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import numpy as np

# Construct valid arguments for f1_custom
# The function takes true and pred as parameters, both should be array-like
# with integer class labels from 0 to 4

# Create true labels - array with 5 classes (0-4), with different occurrences
true = np.array([0, 0, 1, 1, 1, 2, 2, 3, 4, 4])

# Create predictions - some correct, some incorrect
pred = np.array([0, 0, 1, 1, 2, 2, 3, 3, 4, 0])

ARGS = {
    "true": true,
    "pred": pred
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type float64, ndim 0, dtype family float — no output shape is stated: observe_signature.py recorded float ndim 0, which disagrees with the fixture run
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
fft Ran
ran on a synthesized input
Dreem-Organization/dreem-learning-open/dreem_learning_open/preprocessings/epoch_features_processing.py
code served (permissive licence) · get_code("17f77ead9e021fe4")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned ndarray, ndim 2, dtype family float, shape (2, 151) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import numpy as np

# Create a small signal: 2 samples, each 5 seconds at 10 Hz
# increment_duration=30 seconds would need 300 samples per chunk
# So let's make signal with 600 samples total (2 chunks of 30 seconds each at 10 Hz)
signal_frequency = 10  # Hz
increment_duration = 30  # seconds
num_samples = increment_duration * signal_frequency * 2  # 600 samples for 2 increments
signal = np.random.randn(num_samples).astype(np.float32)

# Create signal_properties dict with sampling frequency
signal_properties = {
    'fs': signal_frequency
}

# Create ARGS dict with all parameters
ARGS = {
    'signal': signal,
    'signal_properties': signal_properties,
    'padding_duration': 0,
    'increment_duration': increment_duration
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 2, dtype family float, shape (2, 151) — shape (2, 151) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
filter Ran
ran on a synthesized input
Dreem-Organization/dreem-learning-open/dreem_learning_open/preprocessings/signal_processing.py
code served (permissive licence) · get_code("9ccdc341527e4e3d")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned tuple. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import numpy as np
from scipy.signal import iirfilter, lfilter, filtfilt

# Create a small synthetic signal
# The function expects a 1D numpy array and processes it with axis=0
# signal_properties dict needs an 'fs' key for sampling frequency
np.random.seed(0)
signal = np.random.randn(100)  # Small 1D array of 100 samples

signal_properties = {
    'fs': 250.0  # Sampling frequency in Hz
}

ARGS = {
    'signal': signal,
    'signal_properties': signal_properties
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, scipy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple — a tuple output has no array shape of its own; none is stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
filter_base Ran
ran on a synthesized input
Dreem-Organization/dreem-learning-open/dreem_learning_open/preprocessings/signal_processing.py
code served (permissive licence) · get_code("02e50ac0d452d271")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned ndarray, ndim 1, dtype family float, shape (500,) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import numpy as np
from scipy.signal import iirfilter, lfilter, filtfilt

np.random.seed(0)

# Create a simple 1D signal with some samples
signal = np.random.randn(500)

# Define ARGS with all parameters for filter_base
ARGS = {
    'signal': signal,
    'axis': -1,
    'fs': 250.0
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, scipy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 1, dtype family float, shape (500,) — shape (500,) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
iir_bandpass_filter Ran
ran on a synthesized input
Dreem-Organization/dreem-learning-open/dreem_learning_open/preprocessings/signal_processing.py
code served (permissive licence) · get_code("282ac4f0e343a8bf")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned ndarray, ndim 1, dtype family float, shape (1000,) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import numpy as np

# Create a simple 1D signal for bandpass filtering
np.random.seed(0)
signal = np.random.randn(1000)  # 1D array with 1000 samples

ARGS = {
    'signal': signal,
    'fs': 250.0,
    'order': 4,
    'frequency_band': [0.4, 18],
    'filter_type': 'butter',
    'axis': -1,
    'forward_backward': False,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, scipy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 1, dtype family float, shape (1000,) — shape (1000,) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
index_window Ran
ran on a synthesized input
Dreem-Organization/dreem-learning-open/dreem_learning_open/preprocessings/epoch_features_processing.py
code served (permissive licence) · get_code("be0b88d3dbb9fec8")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned ndarray, ndim 2, dtype family float, shape (0, 1) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import numpy as np

# Create a small signal (audio-like data)
# Using 1000 samples at 100 Hz for 10 seconds
signal = np.random.randn(1000)

# Create signal_properties dict with sampling frequency
signal_properties = {
    'fs': 100  # 100 Hz sampling frequency
}

# Define the ARGS dictionary with all parameters
ARGS = {
    'signal': signal,
    'signal_properties': signal_properties,
    'increment_duration': 30,
    'padding_duration': None
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 2, dtype family float, shape (0, 1) — shape (0, 1) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
kill_channel Ran
ran on a synthesized input
Dreem-Organization/dreem-learning-open/dreem_learning_open/datasets/data_augmentation.py
code served (permissive licence) · get_code("b7851f924aab2ec7")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned ndarray, ndim 3, dtype family float, shape (2, 4, 3) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import numpy as np

# Set seed for reproducibility
np.random.seed(42)

# Create a small 3D numpy array with shape (batch=2, channels=4, spatial=3)
x = np.random.randn(2, 4, 3).astype(np.float32)

# Define probability parameter
p = 0.5

# Construct ARGS dict with all parameters
ARGS = {
    'x': x,
    'p': p
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 3, dtype family float, shape (2, 4, 3) — shape (2, 4, 3) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
mask_non_scored_values Ran
ran on a synthesized input
Dreem-Organization/dreem-learning-open/dreem_learning_open/utils/scores.py
code served (permissive licence) · get_code("efd2486d369320eb")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned function. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import numpy as np

def simple_metric(true, pred):
    """A simple metric function that compares true and pred values"""
    return np.mean(true == pred)

def mask_non_scored_values(func):
    """ We do not considere unlabeled data to estimate metrics"""

    def lol(true, pred, *args, **kwargs):
        true = np.array(true)
        pred = np.array(pred)
        mask = np.array([y in [0, 1, 2, 3, 4] for y in true])
        return func(true[mask], pred[mask], *args, **kwargs)
    return lol

ARGS = {
    "func": simple_metric
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type function — a function output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a function output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
memmap_hash Ran
ran on a synthesized input
Dreem-Organization/dreem-learning-open/dreem_learning_open/logger/logger.py
code served (permissive licence) · get_code("548eb2a6d247f18e")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned str. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import hashlib
import json

def memmap_hash(memmap_description):
    return hashlib.sha1(json.dumps(memmap_description).encode()).hexdigest()[:10]

# The function takes memmap_description and converts it to JSON,
# then hashes it. memmap_description should be JSON-serializable.
# A typical memmap description would be a dict with metadata about memory-mapped arrays.

ARGS = {
    "memmap_description": {
        "filename": "data.bin",
        "dtype": "float32",
        "shape": [1000, 512],
        "offset": 0,
        "mode": "r"
    }
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type str — a str output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a str output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
processs_group Ran
ran on a synthesized input
Dreem-Organization/dreem-learning-open/dreem_learning_open/preprocessings/h5_to_memmap.py
code served (permissive licence) · get_code("87ffbd79b27cf280")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned tuple. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import json
import numpy as np
from unittest.mock import MagicMock

# Create a mock h5 object with required attributes and methods
h5 = MagicMock()

# Create record description
record_description = [
    {'path': 'signal1', 'fs': 100},
    {'path': 'signal2', 'fs': 100},
]
h5.attrs = {'description': json.dumps(record_description)}

# Create mock datasets for signals
signal1_data = np.random.randn(1000).astype(np.float32)
signal2_data = np.random.randn(1000).astype(np.float32)

h5.__getitem__ = MagicMock(side_effect=lambda key: MagicMock(
    __getitem__=lambda self, idx: signal1_data if key == 'signal1' else signal2_data
))

# Create group description
group_description = {
    'signals': ['signal1', 'signal2'],
    'processings': []
}

# Create signal_processings global dict (required by the function)
signal_processings = {}

ARGS = {
    'h5': h5,
    'group_description': group_description,
    'truncate_to_second': 2
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple — a tuple output has no array shape of its own; none is stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
compute_epoch_features Not yet run
unverified (harvested, no recorded run)
Dreem-Organization/dreem-learning-open/dreem_learning_open/preprocessings/h5_to_memmap.py
code served (permissive licence) · get_code("1161728228a99af7")
How it ran

no invocation recorded.

inference_on_dataset Not yet run
unverified (harvested, no recorded run)
Dreem-Organization/dreem-learning-open/dreem_learning_open/logger/logger.py
code served (permissive licence) · get_code("c7091685d495bb47")
How it ran

no invocation recorded.

torch_spectrogram Not yet run
unverified (harvested, no recorded run)
Dreem-Organization/dreem-learning-open/dreem_learning_open/models/modulo_net/normalization.py
code served (permissive licence) · get_code("a31dce1f7ac6d818")
How it ran

no invocation recorded.

Also present in other papers' code

Of 22 samples lifted from this paper's linked repositories, 22 carry a function identity and 0 do not. Of the 22 distinct functions with an identity, 0 are also present in other papers whose code we hold (0 with at least one paper dated earlier, 0 the same month, 0 later) and 22 are only here among the 52,977 papers whose code Syntology holds (as of 2026-09-26).

Only here means: among the 52,977 papers whose code Syntology holds (51,887 with lifted samples, 1,090 through their assemblies only), as of 2026-09-26, no other paper's lifted samples or current assembly bricks contain an identical function body after normalisation.

Presence, not direction: an identical function body in two papers' repositories does not say who wrote it, which came from which, or that either paper's method rests on it. Matching is exact after normalisation, so a renamed or edited function is not matched. Dates are when papers appeared, not when code was written. Only .py files the harvester lifted are covered; notebooks are not read. A new-style arXiv id YYMM.NNNN dates a paper to the month 20YY-MM; otherwise Paper.year dates it to the year; otherwise it is undated. Dates are when the paper appeared, not when the code was written. Read 2026-11-22.

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

Sleep stage classification constitutes an important element of sleep disorder diagnosis. It relies on the visual inspection of polysomnography records by trained sleep technologists. Automated approaches have been designed to alleviate this resource-intensive task. However, such approaches are usually compared to a single human scorer annotation despite an inter-rater agreement of about 85 % only. The present study introduces two publicly-available datasets, DOD-H including 25 healthy volunteers and DOD-O including 55 patients suffering from obstructive sleep apnea (OSA). Both datasets have been scored by 5 sleep technologists from different sleep centers. We developed a framework to compare automated approaches to a consensus of multiple human scorers. Using this framework, we benchmarked and compared the main literature approaches. We also developed and benchmarked a new deep learning method, SimpleSleepNet, inspired by current state-of-the-art. We demonstrated that many methods can reach human-level performance on both datasets. SimpleSleepNet achieved an F1 of 89.9 % vs 86.8 % on average for human scorers on DOD-H, and an F1 of 88.3 % vs 84.8 % on DOD-O. Our study highlights that using state-of-the-art automated sleep staging outperforms human scorers performance for healthy volunteers and patients suffering from OSA. Consideration could be made to use automated approaches in the clinical setting.

For agents

The same record, over MCP at https://syntology.ai/mcp:

get_harvested_code_for_paper("1911.03221")
get_code_for_paper("1911.03221")
have("1911.03221")

The run record, dated, one paper per request, free:

curl https://syntology.ai/api/ran/1911.03221.json

A badge for a README (the split and the date, never a ratio):

[![Syntology run record](https://syntology.ai/api/ran/1911.03221.svg)](https://syntology.ai/paper/1911.03221)

Connect an agent — have() is free.