SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2406.13864 · ICLR · 2024

Evaluating Representation Learning on the Protein Structure Universe

Pietro Liò, Jian Tang, Zuobai Zhang, Alex Morehead, Jianlin Cheng, Chaitanya Joshi, Simon Mathis, Arian Jamasb, Charles Harris, Kieran Didi, Tom Blundell

arXiv · PDF · Open in the Atlas

Code that ran

14 of 16 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; 2 honoured the contract we drafted, 3 violated it and 1 has no contract verdict; 8 failures of our instrument; 2 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/2406.13864.json · badge: /api/ran/2406.13864.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
a-r-j/ProteinWorkshop canonical (classifier label)
classifier basis: census
8 ran · 0 no recorded run
LPDI-EPFL/masif canonical (classifier label)
classifier basis: census
6 ran · 2 no recorded run
FunctionStatusWhere it lives
compute_dists Ran
ran, violated the contract we drafted
LPDI-EPFL/masif/source/masif_modules/train_ppi_search.py
code served (permissive licence) · get_code("c038d262db53bffe")
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 (2,) 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

# The function compute_dists expects two descriptor arrays
# It computes Euclidean distances between corresponding rows
# descs1 and descs2 should be 2D numpy arrays with the same number of rows
# and the same number of columns (descriptor dimension)

descs1 = np.array([
    [1.0, 2.0, 3.0, 4.0],
    [5.0, 6.0, 7.0, 8.0]
], dtype=np.float32)

descs2 = np.array([
    [1.5, 2.5, 3.5, 4.5],
    [5.5, 6.5, 7.5, 8.5]
], dtype=np.float32)

ARGS = {
    "descs1": descs1,
    "descs2": descs2
}
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 1, dtype family float, shape (2,) — shape (2,) 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
construct_batch Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
LPDI-EPFL/masif/source/masif_modules/train_ppi_search.py
code served (permissive licence) · get_code("7462f95ee754ba4c")
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

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

# Create small numpy arrays for binder data
# Shape: (batch_size, num_points, features)
binder_rho_wrt_center = np.random.rand(10, 8).astype(np.float32)
binder_theta_wrt_center = np.random.rand(10, 8).astype(np.float32)
binder_input_feat = np.random.rand(10, 8, 5).astype(np.float32)
binder_mask = np.random.rand(10, 8).astype(np.float32)

# Create small numpy arrays for positive data
pos_rho_wrt_center = np.random.rand(10, 8).astype(np.float32)
pos_theta_wrt_center = np.random.rand(10, 8).astype(np.float32)
pos_input_feat = np.random.rand(10, 8, 5).astype(np.float32)
pos_mask = np.random.rand(10, 8).astype(np.float32)

# Create small numpy arrays for negative data
neg_rho_wrt_center = np.random.rand(10, 8).astype(np.float32)
neg_theta_wrt_center = np.random.rand(10, 8).astype(np.float32)
neg_input_feat = np.random.rand(10, 8, 5).astype(np.float32)
neg_mask = np.random.rand(10, 8).astype(np.float32)

# Create index arrays for training
c_pos_training_idx = np.array([0, 1, 2, 3], dtype=np.int32)
c_neg_training_idx = np.array([0, 1, 2, 3], dtype=np.int32)
c_neg_training_idx_2 = np.array([4, 5, 6, 7], dtype=np.int32)

ARGS = {
    'binder_rho_wrt_center': binder_rho_wrt_center,
    'binder_theta_wrt_center': binder_theta_wrt_center,
    'binder_input_feat': binder_input_feat,
    'binder_mask': binder_mask,
    'c_pos_training_idx': c_pos_training_idx,
    'pos_rho_wrt_center': pos_rho_wrt_center,
    'pos_theta_wrt_center': pos_theta_wrt_center,
    'pos_input_feat': pos_input_feat,
    'pos_mask': pos_mask,
    'c_neg_training_idx': c_neg_training_idx,
    'neg_rho_wrt_center': neg_rho_wrt_center,
    'neg_theta_wrt_center': neg_theta_wrt_center,
    'neg_input_feat': neg_input_feat,
    'neg_mask': neg_mask,
    'c_neg_training_idx_2': c_neg_training_idx_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
construct_batch_val_test Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
LPDI-EPFL/masif/source/masif_modules/train_ppi_search.py
code served (permissive licence) · get_code("c921767debc219a9")
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

# Create small sample data
batch_size = 2
height = 3
width = 4
feat_dim = 5

# c_idx: indices to select from batch
c_idx = np.array([0, 1])

# rho_wrt_center: radial coordinate data, shape should allow indexing by c_idx
# After indexing: (batch_size, height, width), then expand_dims -> (batch_size, height, width, 1)
rho_wrt_center = np.random.rand(batch_size, height, width).astype(np.float32)

# theta_wrt_center: angular coordinate data, same shape as rho_wrt_center
theta_wrt_center = np.random.rand(batch_size, height, width).astype(np.float32)

# input_feat: feature data, shape (batch_size, height, width, feat_dim)
# The function checks if shape[2] == 5 or 3, so we use 5
input_feat = np.random.rand(batch_size, height, width, feat_dim).astype(np.float32)

# mask: binary mask, shape (batch_size, height, width)
# After indexing and expand_dims -> (height, width, 1)
mask = np.random.randint(0, 2, size=(batch_size, height, width)).astype(np.float32)

# flip parameter
flip = False

ARGS = {
    "c_idx": c_idx,
    "rho_wrt_center": rho_wrt_center,
    "theta_wrt_center": theta_wrt_center,
    "input_feat": input_feat,
    "mask": mask,
    "flip": flip
}
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
default Ran
ran, honoured the contract we drafted
a-r-j/ProteinWorkshop/proteinworkshop/models/utils.py
code served (permissive licence) · get_code("7b8f9a99ba265533")
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 int, ndim 0, dtype family int, 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 torch

torch.manual_seed(0)

ARGS = {
    "val": 42,
    "d": "default_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 int, ndim 0, dtype family int, 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
exists Ran
ran, violated the contract we drafted
a-r-j/ProteinWorkshop/proteinworkshop/models/utils.py
code served (permissive licence) · get_code("77f79a4a1c1acb84")
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 bool, 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 torch

def exists(val):
    return val is not None

# The function takes a single parameter 'val' of type Any
# It simply checks if val is not None
# We can test it with various values

ARGS = {
    'val': 42
}
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 bool, 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
get_data_path Ran
ran; our contract draft was wrong, not the code
a-r-j/ProteinWorkshop/proteinworkshop/datasets/atom3d_datamodule.py
code served (permissive licence) · get_code("09ec7a346cca7bcd")
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 torch
import numpy as np
from beartype import beartype as typechecker

torch.manual_seed(0)

ARGS = {
    'dataset': 'PSR',
    'lba_split': 30,
    'ppi_split': 'DIPS-split'
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
beartype, numpy, torch — 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
get_task_split Ran
ran; our contract draft was wrong, not the code
a-r-j/ProteinWorkshop/proteinworkshop/datasets/atom3d_datamodule.py
code served (permissive licence) · get_code("6f28864098d23c94")
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 torch
import numpy as np

ARGS = {
    'task': 'PSR',
    'lba_split': 30,
    'ppi_split': 'DIPS',
    'res_split': 'cath-topology',
    'msp_split': 30
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
beartype, numpy, torch — 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
get_test_data_path Ran
ran on a synthesized input
a-r-j/ProteinWorkshop/proteinworkshop/datasets/atom3d_datamodule.py
code served (permissive licence) · get_code("69b7c9afed7ab8de")
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 torch
import numpy as np

# Construct arguments for get_test_data_path
ARGS = {
    'dataset': 'PSR',
    'lba_split': 30,
    'ppi_split': 'DIPS-split',
    'test_phase': 'test',
    'use_dips_for_testing': False
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
beartype, numpy, torch — 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
load_config_from_checkpoint Ran
ran; our contract draft was wrong, not the code
a-r-j/ProteinWorkshop/proteinworkshop/model_io.py
code served (permissive licence) · get_code("5b87cb0fe092cb16")
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 DictConfig. 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 os
import tempfile
import torch
import omegaconf

torch.manual_seed(0)

# Create a temporary checkpoint file with the expected structure
temp_dir = tempfile.mkdtemp()
ckpt_path = os.path.join(temp_dir, "checkpoint.pt")

# Create a config dict
config_dict = {
    "learning_rate": 0.001,
    "batch_size": 32,
    "epochs": 10,
    "model_type": "transformer"
}

# Create the checkpoint structure that the function expects
checkpoint = {
    "hyper_parameters": {
        "cfg": config_dict
    }
}

# Save the checkpoint
torch.save(checkpoint, ckpt_path)

ARGS = {
    "ckpt_path": ckpt_path
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
omegaconf, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type DictConfig — a DictConfig output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a DictConfig output the recorded ndim 1 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
mask_input_feat Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
LPDI-EPFL/masif/source/masif_modules/train_masif_site.py
code served (permissive licence) · get_code("de681cc738ca5b43")
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

# Create input_feat: a 3D numpy array
# The function deletes along axis=2, so shape should be (batch, features, channels)
# Using distinct sizes: batch=2, features=4, channels=5
input_feat = np.random.randn(2, 4, 5)

# Create mask: array of 0s and 1s where 0s indicate indices to delete
# Should have same length as axis=2 of input_feat (length 5)
mask = np.array([1.0, 0.0, 1.0, 0.0, 1.0])

ARGS = {
    "input_feat": input_feat,
    "mask": mask,
}
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
max_neg_value Ran
ran; our contract draft was wrong, not the code
a-r-j/ProteinWorkshop/proteinworkshop/models/utils.py
code served (permissive licence) · get_code("f2fe51a4a5646c6f")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Sonnet 4.5, prompt python_fixture_v1) 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 torch

torch.manual_seed(0)

# The function expects a tensor and returns the negative of the maximum value for that tensor's dtype
# We need to provide a tensor with a floating point dtype so torch.finfo works
t = torch.randn(2, 4)

ARGS = {
    't': t
}
Setup block drafted by
a language model, Claude Sonnet 4.5, prompt python_fixture_v1
Imports
torch — 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
pad_indices Ran
ran, honoured the contract we drafted
LPDI-EPFL/masif/source/masif_modules/train_masif_site.py
code served (permissive licence) · get_code("fa57bb24283f3605")
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 int, shape (2, 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

# Analyze the function:
# - indices: a list/sequence of arrays, each array contains indices
# - max_verts: an integer representing the maximum number of vertices
# The function pads each array in indices to length max_verts

# Create sample indices - a list of numpy arrays with different lengths
indices = [
    np.array([0, 1, 2]),
    np.array([5, 6]),
]

max_verts = 5

ARGS = {
    "indices": indices,
    "max_verts": max_verts,
}
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 int, shape (2, 5) — shape (2, 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
run_masif_site Ran
ran; our contract draft was wrong, not the code
LPDI-EPFL/masif/source/masif_modules/train_masif_site.py
code served (permissive licence) · get_code("851def46ba261ee9")
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 list, ndim 3, 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
from unittest.mock import Mock

# Create a mock learning object with the required attributes
learning_obj = Mock()

# Create mock TensorFlow placeholders/variables
learning_obj.rho_coords = Mock(name='rho_coords')
learning_obj.theta_coords = Mock(name='theta_coords')
learning_obj.input_feat = Mock(name='input_feat')
learning_obj.mask = Mock(name='mask')
learning_obj.indices_tensor = Mock(name='indices_tensor')

# Create a mock session that returns a score
learning_obj.session = Mock()
learning_obj.full_score = Mock(name='full_score')
learning_obj.session.run.return_value = [np.array([[0.5, 0.3, 0.7]])]

# Create params dict
params = {
    'learning_rate': 0.001,
    'batch_size': 2,
    'num_features': 4
}

# Create rho_wrt_center: (batch_size, num_points, 1)
rho_wrt_center = np.random.rand(2, 5, 1).astype(np.float32)

# Create theta_wrt_center: (batch_size, num_points, 2)
theta_wrt_center = np.random.rand(2, 5, 2).astype(np.float32)

# Create input_feat: (batch_size, num_points, num_features)
input_feat = np.random.rand(2, 5, 4).astype(np.float32)

# Create mask: (batch_size, max_verts)
mask = np.ones((2, 6), dtype=np.float32)

# Create indices: list of arrays with variable lengths
indices = [
    np.array([0, 1, 2], dtype=int),
    np.array([1, 2, 3, 4], dtype=int)
]

ARGS = {
    'params': params,
    'learning_obj': learning_obj,
    'rho_wrt_center': rho_wrt_center,
    'theta_wrt_center': theta_wrt_center,
    'input_feat': input_feat,
    'mask': mask,
    'indices': indices
}
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 list, ndim 3, dtype family float — a list 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
str2bool Ran
ran, violated the contract we drafted
a-r-j/ProteinWorkshop/proteinworkshop/datasets/flip_datamodule.py
code served (permissive licence) · get_code("c7130aa8aaddfd58")
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 bool, 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 torch

ARGS = {
    "v": "true"
}
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 bool, 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
extract_patch_and_coord Not yet run
unverified (harvested, no recorded run)
LPDI-EPFL/masif/source/masif_modules/read_data_from_surface.py
code served (permissive licence) · get_code("914a02a1cfbb46e1")
How it ran

no invocation recorded.

get_masif_ligand_prediction Not yet run
unverified (harvested, no recorded run)
LPDI-EPFL/masif/comparison/masif_ligand/parse_json_tmscore_rocauc.py
code served (permissive licence) · get_code("1f11ce42bb37e65d")
How it ran

no invocation recorded.

Also present in other papers' code

Of 16 samples lifted from this paper's linked repositories, 16 carry a function identity and 0 do not. Of the 16 distinct functions with an identity, 1 are also present in other papers whose code we hold (1 with at least one paper dated earlier, 0 the same month, 1 later) and 15 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.

FunctionAlso present inEarliest-datedOthers named
max_neg_value 26 other papers
18 dated earlier, 8 later · 23 only as an assembly brick
2112.05682 (2021-12)
Self-attention Does Not Need $O(n^2)$ Memory
2206.00923 (2022-06)
Modeling Image Composition for Complex Scene Generation
2208.01626 (2022-08)
Prompt-to-Prompt Image Editing with Cross Attention Control
and 23 more not named here

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

We introduce ProteinWorkshop, a comprehensive benchmark suite for representation learning on protein structures with Geometric Graph Neural Networks. We consider large-scale pre-training and downstream tasks on both experimental and predicted structures to enable the systematic evaluation of the quality of the learned structural representation and their usefulness in capturing functional relationships for downstream tasks. We find that: (1) large-scale pretraining on AlphaFold structures and auxiliary tasks consistently improve the performance of both rotation-invariant and equivariant GNNs, and (2) more expressive equivariant GNNs benefit from pretraining to a greater extent compared to invariant models. We aim to establish a common ground for the machine learning and computational biology communities to rigorously compare and advance protein structure representation learning. Our open-source codebase reduces the barrier to entry for working with large protein structure datasets by providing: (1) storage-efficient dataloaders for large-scale structural databases including AlphaFoldDB and ESM Atlas, as well as (2) utilities for constructing new tasks from the entire PDB. ProteinWorkshop is available at: github.com/a-r-j/ProteinWorkshop.

For agents

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

get_harvested_code_for_paper("2406.13864")
get_code_for_paper("2406.13864")
have("2406.13864")

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

curl https://syntology.ai/api/ran/2406.13864.json

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

[![Syntology run record](https://syntology.ai/api/ran/2406.13864.svg)](https://syntology.ai/paper/2406.13864)

Connect an agent — have() is free.