SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2501.05040 · 2025

SWE-Fixer: Training Open-Source LLMs for Effective and Efficient GitHub Issue Resolution

arXiv · PDF · Open in the Atlas

Code that ran

13 of 15 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 13 have no contract verdict; 0 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/2501.05040.json · badge: /api/ran/2501.05040.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
jorge-martinez-gil/graphcodebert-feature-integration community (classifier label)
classifier basis not recorded on the link
13 ran · 2 no recorded run
FunctionStatusWhere it lives
bootstrap_metric_ci Ran
ran on a synthesized input
jorge-martinez-gil/graphcodebert-feature-integration/src/featfuse/stats.py
code served (permissive licence) · get_code("5556819fa8611c78")
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
from sklearn.metrics import accuracy_score, f1_score

# Define a simple metric function that takes y_true and y_pred
def simple_metric(y_true, y_pred):
    """Simple accuracy-like metric"""
    return np.mean(y_true == y_pred)

# Create small sample data
np.random.seed(42)
y_true = np.array([0, 1, 1, 0, 1, 0, 1, 1])
y_pred = np.array([0, 1, 0, 0, 1, 0, 1, 1])

ARGS = {
    'metric_fn': simple_metric,
    'y_true': y_true,
    'y_pred_or_score': y_pred,
    'n_boot': 100,
    'confidence': 0.95,
    'seed': 42
}
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 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
compute_metrics Ran
ran on a synthesized input
jorge-martinez-gil/graphcodebert-feature-integration/legacy/fine-tunning-graphcodebert-karnalim-with-features.py
code served (permissive licence) · get_code("c17d5d78451bbd57")
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 numpy as np

torch.manual_seed(0)

# The function expects p to be a tuple of (predictions, labels)
# where predictions is a 2D array/tensor with shape (batch_size, num_classes)
# and labels is a 1D array/tensor with shape (batch_size,)
# For binary classification, we need 2 classes

batch_size = 2
num_classes = 2

# Create predictions as a 2D array with logits/scores for each class
predictions = np.random.randn(batch_size, num_classes).astype(np.float32)

# Create binary labels (0 or 1)
labels = np.array([0, 1], dtype=np.int64)

# The function expects p to be a tuple of (predictions, labels)
p = (predictions, labels)

ARGS = {
    'p': p
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, sklearn, torch — 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
config_hash Ran
ran on a synthesized input
jorge-martinez-gil/graphcodebert-feature-integration/src/featfuse/reproducibility.py
code served (permissive licence) · get_code("937c3bfd40cb3ca5")
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
from typing import Any, Dict, Optional

# The function config_hash takes a config dict and returns a stable SHA-256 hash.
# It uses json.dumps with sort_keys=True and default=str to serialize the dict.
# No special attributes or methods are called on the dict itself.

# Create a simple configuration dictionary with various types of values
config = {
    "model_name": "test_model",
    "learning_rate": 0.001,
    "batch_size": 32,
    "num_epochs": 10,
    "hidden_dim": 256,
    "dropout": 0.1,
    "use_attention": True,
    "layers": [64, 128, 256],
    "nested_config": {
        "activation": "relu",
        "norm_type": "batch",
        "epsilon": 1e-5
    }
}

ARGS = {
    "config": config
}
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
expected_calibration_error Ran
ran on a synthesized input
jorge-martinez-gil/graphcodebert-feature-integration/src/featfuse/metrics.py
code served (permissive licence) · get_code("4dc062f89cd64e1d")
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

# Create sample data for expected_calibration_error
np.random.seed(0)

# y_true: binary labels (0 or 1)
y_true = np.array([0, 1, 1, 0, 1, 0, 1, 1])

# y_prob: predicted probabilities (between 0 and 1)
y_prob = np.array([0.1, 0.9, 0.8, 0.2, 0.7, 0.3, 0.85, 0.95])

# n_bins: number of bins for calibration error calculation
n_bins = 5

ARGS = {
    "y_true": y_true,
    "y_prob": y_prob,
    "n_bins": n_bins,
}
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 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
get_classifier Ran
ran on a synthesized input
jorge-martinez-gil/graphcodebert-feature-integration/src/featfuse/experiment.py
code served (permissive licence) · get_code("306fd49af7f42a7e")
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 LogisticRegression. 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 sklearn.ensemble import GradientBoostingClassifier, RandomForestClassifier
from sklearn.linear_model import LogisticRegression
from sklearn.neural_network import MLPClassifier
from sklearn.svm import SVC

torch.manual_seed(0)

# The function takes two parameters: name and seed
# name: a string that maps to a classifier type
# seed: an integer for random state (default 42)

ARGS = {
    'name': 'logistic_regression',
    'seed': 42
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, sklearn, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type LogisticRegression — a LogisticRegression output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a LogisticRegression 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
make_manifest Ran
ran on a synthesized input
jorge-martinez-gil/graphcodebert-feature-integration/src/featfuse/reproducibility.py
code served (permissive licence) · get_code("2119a197049c004a")
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 numpy as np
from datetime import datetime, timezone

# Build the config dict
config = {
    "model_name": "test_model",
    "learning_rate": 0.001,
    "batch_size": 32,
    "num_epochs": 10,
    "hidden_dim": 256,
    "dropout": 0.1,
    "optimizer": "adam"
}

# Set seed for reproducibility
seed = 42

# Build optional extra dict
extra = {
    "run_name": "test_run",
    "experiment_id": "exp_001",
    "notes": "test execution"
}

# Define ARGS mapping all parameters
ARGS = {
    "config": config,
    "seed": seed,
    "extra": extra
}
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 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
mcnemar_test Ran
ran on a synthesized input
jorge-martinez-gil/graphcodebert-feature-integration/src/featfuse/stats.py
code served (permissive licence) · get_code("6189dbb4e2522b64")
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

# McNemar's test compares two classifiers on the same examples
# It needs:
# - y_true: ground truth labels
# - pred_a: predictions from classifier A
# - pred_b: predictions from classifier B

# Create small test data
np.random.seed(0)
n_samples = 20

# Ground truth labels (binary classification)
y_true = np.array([0, 1, 0, 1, 1, 0, 1, 0, 0, 1, 
                   1, 0, 1, 0, 1, 1, 0, 0, 1, 0])

# Predictions from classifier A (some correct, some incorrect)
pred_a = np.array([0, 1, 0, 1, 1, 1, 1, 0, 0, 1,
                   1, 0, 0, 0, 1, 1, 0, 0, 1, 0])

# Predictions from classifier B (different mistakes than A)
pred_b = np.array([0, 1, 1, 1, 1, 0, 1, 0, 0, 0,
                   1, 0, 1, 0, 1, 1, 0, 1, 1, 0])

ARGS = {
    'y_true': y_true,
    'pred_a': pred_a,
    'pred_b': pred_b,
}
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 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
paired_bootstrap_diff Ran
ran on a synthesized input
jorge-martinez-gil/graphcodebert-feature-integration/src/featfuse/stats.py
code served (permissive licence) · get_code("6fd9fc26f9ea8174")
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
from sklearn.metrics import accuracy_score

def simple_metric(y_true, y_pred):
    """Simple accuracy metric for testing"""
    return accuracy_score(y_true, y_pred)

# Create small test data
np.random.seed(42)
n_samples = 10
y_true = np.random.randint(0, 2, n_samples)  # Binary labels
pred_a = np.random.randint(0, 2, n_samples)  # Predictions from model A
pred_b = np.random.randint(0, 2, n_samples)  # Predictions from model B

ARGS = {
    'metric_fn': simple_metric,
    'y_true': y_true,
    'pred_a': pred_a,
    'pred_b': pred_b,
    'n_boot': 100,
    'confidence': 0.95,
    'seed': 42
}
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 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
plot_calibration Ran
ran on a synthesized input
jorge-martinez-gil/graphcodebert-feature-integration/src/featfuse/viz.py
code served (permissive licence) · get_code("293e3be63397cd3b")
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 numpy as np
import tempfile
import os

# Create a temporary directory for saving the plot
temp_dir = tempfile.gettempdir()
temp_path = os.path.join(temp_dir, "test_calibration_plot.png")

# Generate synthetic calibration data
np.random.seed(0)
n_samples = 100
y_true = np.random.randint(0, 2, n_samples)  # Binary labels
y_prob = np.random.uniform(0, 1, n_samples)  # Predicted probabilities

ARGS = {
    'y_true': y_true,
    'y_prob': y_prob,
    'path': temp_path,
    'n_bins': 10,
    'title': 'Reliability diagram'
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
matplotlib, numpy — 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
plot_feature_importance Ran
ran on a synthesized input
jorge-martinez-gil/graphcodebert-feature-integration/src/featfuse/viz.py
code served (permissive licence) · get_code("111459a93cf871ff")
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 tempfile
import numpy as np

ARGS = {
    'names': ['feature_a', 'feature_b', 'feature_c', 'feature_d'],
    'importances': [0.15, 0.42, 0.28, 0.09],
    'path': tempfile.mktemp(suffix='.png'),
    'title': 'Feature importance'
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
matplotlib, numpy — 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
plot_metric_bars Ran
ran on a synthesized input
jorge-martinez-gil/graphcodebert-feature-integration/src/featfuse/viz.py
code served (permissive licence) · get_code("02bdd0d30c99b07a")
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 numpy as np
import tempfile
import os

# Create a temporary directory for saving the plot
temp_dir = tempfile.mkdtemp()
temp_path = os.path.join(temp_dir, "test_plot.png")

# Define arguments for plot_metric_bars
ARGS = {
    'method_metrics': {
        'Method_A': {
            'accuracy': 0.85,
            'precision': 0.82,
            'recall': 0.88,
            'f1': 0.85
        },
        'Method_B': {
            'accuracy': 0.91,
            'precision': 0.89,
            'recall': 0.93,
            'f1': 0.91
        },
        'Method_C': {
            'accuracy': 0.78,
            'precision': 0.75,
            'recall': 0.81,
            'f1': 0.78
        }
    },
    'metrics': ['accuracy', 'precision', 'recall', 'f1'],
    'path': temp_path,
    'title': 'Method comparison'
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
matplotlib, numpy — 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
set_seed Ran
ran on a synthesized input
jorge-martinez-gil/graphcodebert-feature-integration/src/featfuse/reproducibility.py
code served (permissive licence) · get_code("0dfa44f5bdd15b72")
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. The code mentions cuda, and the harness aliases cuda to the CPU for every sample; whether this run needed that is 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 os
import random
import numpy as np

# The set_seed function takes two parameters:
# 1. seed: int (default 42)
# 2. deterministic: bool (default True)

# Both parameters are simple scalars, so we just need to provide valid values
ARGS = {
    "seed": 42,
    "deterministic": True,
}
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 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
similarity Ran
ran on a synthesized input
jorge-martinez-gil/graphcodebert-feature-integration/legacy/utils/java_sim_exec_opt.py
code served (permissive licence) · get_code("0785ae298e13dd83")
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 subprocess
import os
from difflib import SequenceMatcher

def execute_java_code(java_code):
    class_name = 'Test'
    java_filename = f'{class_name}.java'
    class_filename = f'{class_name}.class'
    full_java_code = f'public class {class_name} {{\n{java_code}\n}}'
    with open(java_filename, 'w') as file:
        file.write(full_java_code)
    try:
        subprocess.run(['javac', java_filename], check=True, stderr=subprocess.PIPE)
        result = subprocess.run(['java', class_name], stdout=subprocess.PIPE, stderr=subprocess.PIPE, text=True)
    except subprocess.CalledProcessError as e:
        print(f'Error during compilation/execution: {e.stderr.decode()}')
        return None
    finally:
        os.remove(java_filename)
        if os.path.exists(class_filename):
            os.remove(class_filename)
    return result.stdout.strip()

def similarity(java_code1, java_code2):
    try:
        output1 = execute_java_code(java_code1)
        output2 = execute_java_code(java_code2)
        if output1 is None or output2 is None:
            print('Error executing one or both of the Java code snippets.')
            return
        similarity_ratio = SequenceMatcher(None, output1, output2).ratio()
        return similarity_ratio
    except Exception as e:
        return 0

# Setup code to construct valid arguments
java_code_1 = """
    public static void main(String[] args) {
        System.out.println("Hello World");
    }
"""

java_code_2 = """
    public static void main(String[] args) {
        System.out.println("Hello World");
    }
"""

ARGS = {
    'java_code1': java_code_1,
    'java_code2': java_code_2
}
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 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
classification_metrics Not yet run
unverified (harvested, no recorded run)
jorge-martinez-gil/graphcodebert-feature-integration/src/featfuse/metrics.py
code served (permissive licence) · get_code("1b12db6bf011d6fa")
How it ran

no invocation recorded.

execute_java_code Not yet run
unverified (harvested, no recorded run)
jorge-martinez-gil/graphcodebert-feature-integration/legacy/utils/java_sim_exec_opt.py
code served (permissive licence) · get_code("69d614ed324ad1a6")
How it ran

no invocation recorded.

Also present in other papers' code

Of 15 samples lifted from this paper's linked repositories, 15 carry a function identity and 0 do not. Of the 15 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 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.

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

Large Language Models (LLMs) have demonstrated remarkable proficiency across a variety of complex tasks. One significant application of LLMs is in tackling software engineering challenges, particularly in resolving real-world tasks on GitHub by fixing code based on the issues reported by the users. However, many current approaches rely on proprietary LLMs, which limits reproducibility, accessibility, and transparency. The critical components of LLMs for addressing software engineering issues and how their capabilities can be effectively enhanced remain unclear. To address these challenges, we introduce SWE-Fixer, a novel open-source framework designed to effectively and efficiently resolve GitHub issues. SWE-Fixer comprises two essential modules: a code file retrieval module and a code editing module. The retrieval module employs BM25 along with a lightweight model to achieve coarse-to-fine file retrieval. Subsequently, the code editing module utilizes the other model to generate patches for the identified files. To mitigate the lack of publicly available datasets, we compile an extensive dataset that includes 110K GitHub issues along with their corresponding patches and train the two models of SWE-Fixer separately. We assess our approach on the SWE-Bench Lite and Verified benchmarks, achieving competitive performance among open-source models with scores of 22.0% and 30.2%. Furthermore, SWE-Fixer reaches state-of-the-art performance (24.7% on Lite and 32.8% on Verified) with PASS_TO_PASS (P2P) filtering. Additionally, our approach requires only two model calls per instance, making it significantly more efficient than existing methods. These results highlight the effectiveness of SWE-Fixer in real-world code-fixing scenarios. We will make our model, dataset, and code publicly available at https://github.com/InternLM/SWE-Fixer.

For agents

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

get_harvested_code_for_paper("2501.05040")
get_code_for_paper("2501.05040")
have("2501.05040")

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

curl https://syntology.ai/api/ran/2501.05040.json

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

[![Syntology run record](https://syntology.ai/api/ran/2501.05040.svg)](https://syntology.ai/paper/2501.05040)

Connect an agent — have() is free.