SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2207.12201 · 2022

Calibrated One-class Classification for Unsupervised Time Series Anomaly Detection

arXiv · PDF · Open in the Atlas

Code that ran

11 of 14 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 11 have no contract verdict; 0 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/2207.12201.json · badge: /api/ran/2207.12201.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
xuhongzuo/couta canonical (classifier label)
classifier basis not recorded on the link
11 ran · 3 no recorded run
FunctionStatusWhere it lives
adjust_scores Ran
ran on a synthesized input
xuhongzuo/couta/src/utils_eval.py
code served (permissive licence) · get_code("5bfa4b4ceafb84c9")
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 (12,) 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 sample data for testing adjust_scores
# label: ground-truth binary labels indicating normal (0) and anomaly (1) segments
# score: anomaly scores for each time point

# Create a simple label array with multiple segments:
# - 0-2: normal (label=0)
# - 3-5: anomaly (label=1)
# - 6-8: normal (label=0)
# - 9-11: anomaly (label=1)
label = np.array([0, 0, 0, 1, 1, 1, 0, 0, 0, 1, 1, 1])

# Create corresponding anomaly scores
# Higher scores for anomalies, lower for normal points
score = np.array([0.1, 0.2, 0.15, 0.8, 0.6, 0.9, 0.1, 0.25, 0.05, 0.85, 0.75, 0.95])

ARGS = {
    "label": label,
    "score": score,
}
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 (12,) — shape (12,) 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
cosine Ran
ran on a synthesized input
xuhongzuo/couta/src/experiments/data_generator_collective_case.py
code served (permissive licence) · get_code("1be0714f24c5f3db")
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 (100,) 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

np.random.seed(0)

ARGS = {
    'length': 100,
    'freq': 0.04,
    'coef': 1.5,
    'offset': 0.0,
    'noise_amp': 0.05,
}
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 (100,) — shape (100,) 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
create_batch_neg Ran
ran on a synthesized input
xuhongzuo/couta/src/algorithms/couta_algo.py
code served (permissive licence) · get_code("3f2b46c7dfce194f")
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 torch
import numpy as np

torch.manual_seed(0)

# Create a small batch of sequences
# batch_size=2, sequence_length=15, dimension=4
batch_seqs_torch = torch.randn(2, 15, 4)
batch_seqs_np = np.random.randn(2, 15, 4)

# Test with torch tensor version
ARGS = {
    'batch_seqs': batch_seqs_torch,
    'max_cut_ratio': 0.5,
    'seed': 0,
    'return_mul_label': False,
    'ss_type': 'FULL'
}
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 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
data_standardize Ran
ran on a synthesized input
xuhongzuo/couta/src/utils_general.py
code served (permissive licence) · get_code("d95b338432e5df83")
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, 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 pandas as pd
import numpy as np

# Create sample DataFrames with numeric columns
np.random.seed(0)

# Create training data with 4 columns and 2 rows
X_train = pd.DataFrame({
    'col1': [1.0, 2.0],
    'col2': [10.0, 20.0],
    'col3': [100.0, 200.0],
    'col4': [5.0, 5.0]  # constant column for testing remove/keep logic
})

# Create test data with same structure
X_test = pd.DataFrame({
    'col1': [1.5, 1.8],
    'col2': [15.0, 18.0],
    'col3': [150.0, 180.0],
    'col4': [5.0, 5.0]
})

ARGS = {
    'X_train': X_train,
    'X_test': X_test,
    'remove': False,
    'verbose': False,
    'max_clip': 5,
    'min_clip': -4
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, pandas — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple, ndim 3, 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
get_data_lst Ran
ran on a synthesized input
xuhongzuo/couta/main_utils.py
code served (permissive licence) · get_code("6695ea1de391d0cc")
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 os
import tempfile
import pandas as pd
import numpy as np

# Create a temporary directory structure with CSV files
temp_dir = tempfile.mkdtemp()
data_root = temp_dir + "/"
data_name = "test_data"
data_path = os.path.join(data_root, data_name)
os.makedirs(data_path, exist_ok=True)

# Create machine subdirectories with train/test CSV files
machine1_path = os.path.join(data_path, "machine1")
machine2_path = os.path.join(data_path, "machine2")
os.makedirs(machine1_path, exist_ok=True)
os.makedirs(machine2_path, exist_ok=True)

# Create sample train and test CSVs for machine1
train_df1 = pd.DataFrame({
    'feature1': [1.0, 2.0, 3.0],
    'feature2': [4.0, 5.0, 6.0],
    'label': [0, 1, 0]
})
test_df1 = pd.DataFrame({
    'feature1': [7.0, 8.0],
    'feature2': [9.0, 10.0],
    'label': [1, 0]
})
train_df1.index.name = 'idx'
test_df1.index.name = 'idx'
train_df1.to_csv(os.path.join(machine1_path, "train.csv"))
test_df1.to_csv(os.path.join(machine1_path, "test.csv"))

# Create sample train and test CSVs for machine2
train_df2 = pd.DataFrame({
    'feature1': [11.0, 12.0, 13.0],
    'feature2': [14.0, 15.0, 16.0],
    'label': [1, 0, 1]
})
test_df2 = pd.DataFrame({
    'feature1': [17.0, 18.0],
    'feature2': [19.0, 20.0],
    'label': [0, 1]
})
train_df2.index.name = 'idx'
test_df2.index.name = 'idx'
train_df2.to_csv(os.path.join(machine2_path, "train.csv"))
test_df2.to_csv(os.path.join(machine2_path, "test.csv"))

ARGS = {
    'data': 'test_data',
    'data_root': data_root,
    'entities': 'machine1,machine2'
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, pandas — 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
get_model_configs Ran
ran on a synthesized input
xuhongzuo/couta/main_utils.py
code served (permissive licence) · get_code("0d6ed0e36e58f5e7")
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 torch
import yaml
import tempfile
import os

# Create a temporary directory and a minimal YAML config file
temp_dir = tempfile.mkdtemp()
config_dir = os.path.join(temp_dir, 'configs')
os.makedirs(config_dir, exist_ok=True)

yaml_path = os.path.join(config_dir, 'COUTA.yaml')
config_content = {
    'DSADS': {
        'model_type': 'lstm',
        'hidden_size': 64,
        'num_layers': 2
    },
    'Epilepsy': {
        'model_type': 'cnn',
        'num_filters': 32,
        'kernel_size': 3
    },
    'ASD': {
        'model_type': 'transformer',
        'num_heads': 4
    }
}

with open(yaml_path, 'w') as f:
    yaml.dump(config_content, f)

# Change to temp directory so the relative path works
original_dir = os.getcwd()
os.chdir(temp_dir)

try:
    # Define arguments for get_model_configs
    ARGS = {
        'algo': 'couta',
        'data': 'DSADS_motion_data'
    }
finally:
    # Restore original directory
    os.chdir(original_dir)
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch, yaml — 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
get_sequences_data Ran
ran on a synthesized input
xuhongzuo/couta/src/utils_general.py
code served (permissive licence) · get_code("f1e66a69ff07171c")
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
import tempfile
import os

# Create a temporary directory structure with mock data files
temp_dir = tempfile.mkdtemp()
data_root = temp_dir + "/"
data_name = "test_data"
data_dir = os.path.join(data_root, data_name)
os.makedirs(data_dir, exist_ok=True)

# Create mock numpy arrays with distinct shapes
train_x = np.random.randn(10, 4).astype(np.float32)
test_x = np.random.randn(5, 4).astype(np.float32)
train_y = np.random.randint(0, 2, 10).astype(np.int64)
test_y = np.random.randint(0, 2, 5).astype(np.int64)

# Save the mock data files
np.save(os.path.join(data_dir, f"{data_name}_trainx.npy"), train_x)
np.save(os.path.join(data_dir, f"{data_name}_testx.npy"), test_x)
np.save(os.path.join(data_dir, f"{data_name}_trainy.npy"), train_y)
np.save(os.path.join(data_dir, f"{data_name}_testy.npy"), test_y)

ARGS = {
    "data": data_name,
    "data_root": data_root,
}
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
get_sub_seqs Ran
ran on a synthesized input
xuhongzuo/couta/src/algorithms/algorithm_utils.py
code served (permissive licence) · get_code("370f5ea00f912a03")
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 (4, 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

# Construct x_arr: a time-series array with shape (time, channels)
# Using shape (10, 3) - 10 time points, 3 channels
x_arr = np.random.randn(10, 3).astype(np.float32)

# Define seq_len: the window size for subsequences
seq_len = 4

# Define stride: how many time points to move between subsequences
stride = 2

# Define start_discont: starting points of discontinuities
# Using empty array as default, but can also test with values
start_discont = np.array([])

# Alternative with discontinuities (optional, but shown as possibility):
# start_discont = np.array([6, 8])

ARGS = {
    "x_arr": x_arr,
    "seq_len": seq_len,
    "stride": stride,
    "start_discont": start_discont,
}
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 (4, 4, 3) — shape (4, 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
minmax_norm Ran
ran on a synthesized input
xuhongzuo/couta/src/utils_general.py
code served (permissive licence) · get_code("588a35c94648bbe1")
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 (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

ARGS = {
    "arr": [1.0, 2.5, 3.7, 4.2, 5.1]
}
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 (5,) — shape (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
series_segmentation Ran
ran on a synthesized input
xuhongzuo/couta/src/experiments/data_generator_collective_case.py
code served (permissive licence) · get_code("8561f20e4c8119f3")
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. 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

# The function takes 'data' and an optional 'stepsize' parameter
# 'data' is used with np.diff() and np.split(), so it should be a 1D numpy array
# 'stepsize' is used in comparison with np.diff(data), so it should be numeric

# Create a simple 1D array where we have consistent differences of 1
# This will create segments that the function can split
data = np.array([1, 2, 3, 4, 6, 7, 8, 10, 11], dtype=np.int32)

stepsize = 1

ARGS = {
    "data": data,
    "stepsize": stepsize
}
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 — a list 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
sine Ran
ran on a synthesized input
xuhongzuo/couta/src/experiments/data_generator_collective_case.py
code served (permissive licence) · get_code("d67d19b30290c8e8")
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 (100,) 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

ARGS = {
    "length": 100,
    "freq": 0.04,
    "coef": 1.5,
    "offset": 0.0,
    "noise_amp": 0.05,
}
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 (100,) — shape (100,) 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_best_f1 Not yet run
unverified (harvested, no recorded run)
xuhongzuo/couta/src/utils_eval.py
code served (permissive licence) · get_code("a69f0fe91418b83e")
How it ran

no invocation recorded.

get_logger Not yet run
unverified (harvested, no recorded run)
xuhongzuo/couta/main_utils.py
code served (permissive licence) · get_code("22c9707971cf2a58")
How it ran

no invocation recorded.

get_metrics Not yet run
unverified (harvested, no recorded run)
xuhongzuo/couta/src/utils_eval.py
code served (permissive licence) · get_code("ad6de4f08952b158")
How it ran

no invocation recorded.

Also present in other papers' code

Of 14 samples lifted from this paper's linked repositories, 14 carry a function identity and 0 do not. Of the 14 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, 0 later) and 13 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
get_sub_seqs 1 other paper
1 dated earlier
2109.11428 (2021-09)
An Evaluation of Anomaly Detection and Diagnosis in Multivariate Time Series

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

Time series anomaly detection is instrumental in maintaining system availability in various domains. Current work in this research line mainly focuses on learning data normality deeply and comprehensively by devising advanced neural network structures and new reconstruction/prediction learning objectives. However, their one-class learning process can be misled by latent anomalies in training data (i.e., anomaly contamination) under the unsupervised paradigm. Their learning process also lacks knowledge about the anomalies. Consequently, they often learn a biased, inaccurate normality boundary. To tackle these problems, this paper proposes calibrated one-class classification for anomaly detection, realizing contamination-tolerant, anomaly-informed learning of data normality via uncertainty modeling-based calibration and native anomaly-based calibration. Specifically, our approach adaptively penalizes uncertain predictions to restrain irregular samples in anomaly contamination during optimization, while simultaneously encouraging confident predictions on regular samples to ensure effective normality learning. This largely alleviates the negative impact of anomaly contamination. Our approach also creates native anomaly examples via perturbation to simulate time series abnormal behaviors. Through discriminating these dummy anomalies, our one-class learning is further calibrated to form a more precise normality boundary. Extensive experiments on ten real-world datasets show that our model achieves substantial improvement over sixteen state-of-the-art contenders.

For agents

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

get_harvested_code_for_paper("2207.12201")
get_code_for_paper("2207.12201")
have("2207.12201")

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

curl https://syntology.ai/api/ran/2207.12201.json

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

[![Syntology run record](https://syntology.ai/api/ran/2207.12201.svg)](https://syntology.ai/paper/2207.12201)

Connect an agent — have() is free.