SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2004.09167 · 2020

CheXbert: Combining Automatic Labelers and Expert Annotations for Accurate Radiology Report Labeling Using BERT

arXiv · PDF · Open in the Atlas

Code that ran

15 of 17 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 9 have no contract verdict; 6 failures of our instrument; 2 no recorded run), measured 2026-11-21; 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/2004.09167.json · badge: /api/ran/2004.09167.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
wjhou/recap community (classifier label)
classifier basis not recorded on the link
7 ran · 0 no recorded run
wjhou/icon community (classifier label)
classifier basis not recorded on the link
5 ran · 2 no recorded run
wjhou/organ community (classifier label)
classifier basis not recorded on the link
3 ran · 0 no recorded run
FunctionStatusWhere it lives
cook_refs Ran
ran; our contract draft was wrong, not the code
wjhou/recap/pycocoevalcap/cider/cider_scorer.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from wenhuchen/data-to-text-evaluation-metric, is pointer only (licence: NOASSERTION) · get_code("4c1b107103b7c856")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (Apache-2.0), but get_code serves the stored copy, first harvested from wenhuchen/data-to-text-evaluation-metric, and GitHub could not classify the licence of that repository (NOASSERTION). The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

cook_refs Ran
ran; our contract draft was wrong, not the code
wjhou/recap/pycocoevalcap/bleu/bleu_scorer.py
code served (permissive licence) · get_code("7586e1d32f7325f7")
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 1, 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.

from collections import defaultdict

ARGS = {
    'refs': [
        'the quick brown fox jumps over the lazy dog',
        'a quick brown fox leaps over a lazy dog'
    ],
    'eff': 'shortest',
    'n': 4
}
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 tuple, ndim 1, 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
cook_test Ran
ran; our contract draft was wrong, not the code
wjhou/recap/pycocoevalcap/cider/cider_scorer.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from wenhuchen/data-to-text-evaluation-metric, is pointer only (licence: NOASSERTION) · get_code("3573c4bc85f65dfd")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (Apache-2.0), but get_code serves the stored copy, first harvested from wenhuchen/data-to-text-evaluation-metric, and GitHub could not classify the licence of that repository (NOASSERTION). The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

cook_test Ran
ran; our contract draft was wrong, not the code
wjhou/recap/pycocoevalcap/bleu/bleu_scorer.py
code served (permissive licence) · get_code("bc5342cb883ef9da")
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.

from collections import defaultdict

# First, let's understand what cook_test needs:
# - test: a string (test sentence)
# - refs: a tuple of (reflen, refmaxcounts) where reflen is a list and refmaxcounts is a dict
# - eff: optional string parameter (can be 'closest' or None)
# - n: integer for n-gram size (default 4)

# Build refs by calling precook on reference sentences
def precook(s, n=4, out=False):
    """Takes a string as input and returns an object that can be given to
    either cook_refs or cook_test. This is optional: cook_refs and cook_test
    can take string arguments as well."""
    words = s.split()
    counts = defaultdict(int)
    for k in range(1, n + 1):
        for i in range(len(words) - k + 1):
            ngram = tuple(words[i:i + k])
            counts[ngram] += 1
    return (len(words), counts)

# Create reference data
# refs should be a tuple of (reflen, refmaxcounts)
# reflen is a list of reference lengths
# refmaxcounts is a dict of max counts across all references

ref1_len, ref1_counts = precook("the cat sat on the mat", n=4)
ref2_len, ref2_counts = precook("a cat is sitting on a mat", n=4)

# Create refmaxcounts by taking max of counts for each ngram
refmaxcounts = defaultdict(int)
for ngram, count in ref1_counts.items():
    refmaxcounts[ngram] = max(refmaxcounts[ngram], count)
for ngram, count in ref2_counts.items():
    refmaxcounts[ngram] = max(refmaxcounts[ngram], count)

# refs format: (list_of_reflengths, dict_of_maxcounts)
refs = ([ref1_len, ref2_len], refmaxcounts)

# Create test sentence
test = "the cat sat on the mat"

# Create parameters
ARGS = {
    'test': test,
    'refs': refs,
    'eff': None,
    'n': 4
}
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 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
extract_temporal_info Ran
ran on a synthesized input
wjhou/icon/src_stage1/dataset_ende.py
code served (permissive licence) · get_code("a58f9ee70337dd6e")
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 1, 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 torch
import numpy as np
from tqdm import tqdm

torch.manual_seed(0)

# Create sample dictionaries with required structure
samples = [
    {
        'id': 'sample_1',
        'image_path': '/path/to/image1.jpg',
        'report': 'Initial report for sample 1'
    },
    {
        'id': 'sample_2',
        'image_path': '/path/to/image2.jpg',
        'report': 'Initial report for sample 2'
    },
    {
        'id': 'sample_3',
        'image_path': '/path/to/image3.jpg',
        'report': 'Initial report for sample 3'
    }
]

# Create reference samples
ref_samples = [
    {
        'id': 'ref_sample_1',
        'image_path': '/path/to/ref_image1.jpg',
        'report': 'Reference report for ref sample 1'
    },
    {
        'id': 'ref_sample_2',
        'image_path': '/path/to/ref_image2.jpg',
        'report': 'Reference report for ref sample 2'
    }
]

# Create temporal_ids mapping
temporal_ids = {
    'sample_1': {
        'object_id': 'sample_2',
        'predicate': 'precedes'
    },
    'sample_2': {
        'object_id': 'ref_sample_1',
        'predicate': 'follows'
    },
    'sample_3': {
        'object_id': 'sample_1',
        'predicate': 'concurrent_with'
    }
}

ARGS = {
    'samples': samples,
    'ref_samples': ref_samples,
    'temporal_ids': temporal_ids
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, torch, tqdm — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type list, ndim 1, 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
f1 Ran
ran on a synthesized input
wjhou/icon/src_retrieval/retrieval.py
code served (permissive licence) · get_code("6d01795fce054855")
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 torch
import numpy as np
from typing import Set

# Construct valid sets for the function
x_set = {1, 2, 3, 4, 5}
y_set = {3, 4, 5, 6, 7}

ARGS = {
    "x": x_set,
    "y": y_set
}
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 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
label_func Ran
ran on a synthesized input
wjhou/icon/eval/eval_consistency.py
code served (permissive licence) · get_code("1cecec4bb2430637")
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. 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 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 torch.nn as nn

torch.manual_seed(0)

class MockTokenizer:
    def __call__(self, texts, return_tensors=None, padding=None):
        batch_size = len(texts)
        seq_length = 5
        return {
            'input_ids': torch.randint(0, 100, (batch_size, seq_length)),
            'attention_mask': torch.ones(batch_size, seq_length)
        }

class MockModel(nn.Module):
    def __init__(self):
        super().__init__()
        self.linear = nn.Linear(10, 8)
    
    def forward(self, source_padded, attention_mask):
        batch_size = source_padded.shape[0]
        logit1 = self.linear(torch.randn(batch_size, 10))
        logit2 = self.linear(torch.randn(batch_size, 10))
        return [logit1, logit2]

texts = ["report one", "report two", "report three"]
model = MockModel()
tokenizer = MockTokenizer()

ARGS = {
    'texts': texts,
    'model': model,
    'tokenizer': tokenizer,
    'max_batch_size': 2
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch, tqdm — 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
load_images Ran
ran on a synthesized input
wjhou/icon/src_stage1/dataset_ende.py
code served (permissive licence) · get_code("b878a058d12646b0")
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 os
import tempfile
from PIL import Image
import numpy as np

# Create a temporary directory with test images
temp_dir = tempfile.mkdtemp()

# Create some test images
img_array_1 = np.random.randint(0, 256, (10, 10, 3), dtype=np.uint8)
img_array_2 = np.random.randint(0, 256, (15, 15, 3), dtype=np.uint8)

img_1 = Image.fromarray(img_array_1, 'RGB')
img_2 = Image.fromarray(img_array_2, 'RGB')

img_1_path = os.path.join(temp_dir, 'test_image_1.png')
img_2_path = os.path.join(temp_dir, 'test_image_2.png')

img_1.save(img_1_path)
img_2.save(img_2_path)

# Create relative paths (without the root_path prefix)
rel_path_1 = 'test_image_1.png'
rel_path_2 = 'test_image_2.png'

# Create image_paths as a list of lists (nested structure)
image_paths = [
    [rel_path_1, rel_path_2]
]

ARGS = {
    'root_path': temp_dir,
    'image_paths': image_paths
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
PIL, numpy, tqdm — 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
load_mentions Ran
ran on a synthesized input
wjhou/organ/src/plan_extraction.py
code served (permissive licence) · get_code("673d3788abc824eb")
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 os
import tempfile
import shutil

# Create a temporary directory with test files
temp_dir = tempfile.mkdtemp()

# Create test files with mention data
test_files = {
    'obs1.txt': 'mention1\nmention2\nmention3\n',
    'obs2.txt': 'mention4\nmention5\n',
    'obs3.txt': 'mention1\nmention6\n'
}

for filename, content in test_files.items():
    filepath = os.path.join(temp_dir, filename)
    with open(filepath, 'w', encoding='utf-8') as f:
        f.write(content)

ARGS = {
    'folder': temp_dir
}
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 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
load_pmi_mentions Ran
ran on a synthesized input
wjhou/organ/src/plan_extraction.py
code served (permissive licence) · get_code("4cb2e174305de20f")
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 defaultdict. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import json
import tempfile
from collections import defaultdict

# Create a temporary JSON file with PMI mentions data
pmi_data = {
    "pneumonia_1": ["mention_a", "mention_b", "mention_c", "mention_d", "mention_e"],
    "tuberculosis_2": ["mention_x", "mention_y", "mention_z"],
    "No Finding_chest": ["mention_irrelevant"],
    "cardiomegaly_3": ["mention_p", "mention_q", "mention_r", "mention_s"]
}

# Write to temporary file
temp_file = tempfile.NamedTemporaryFile(mode='w', suffix='.json', delete=False, encoding='utf-8')
json.dump(pmi_data, temp_file)
temp_file.close()

ARGS = {
    'path': temp_file.name,
    'topk': 5
}
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 defaultdict — a defaultdict output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a defaultdict 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
precook Ran
ran; our contract draft was wrong, not the code
wjhou/recap/pycocoevalcap/bleu/bleu_scorer.py
code served (permissive licence) · get_code("ec4448373d72d636")
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 tuple, ndim 1, 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.

ARGS = {
    "s": "the quick brown fox jumps over the lazy dog",
    "n": 4,
    "out": False
}
Setup block drafted by
a language model, Claude Sonnet 4.5, prompt python_fixture_v1
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 tuple, ndim 1, 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
precook Ran
ran; our contract draft was wrong, not the code
wjhou/recap/pycocoevalcap/cider/cider_scorer.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from wenhuchen/data-to-text-evaluation-metric, is pointer only (licence: NOASSERTION) · get_code("9560e3e3e41bc1b2")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (Apache-2.0), but get_code serves the stored copy, first harvested from wenhuchen/data-to-text-evaluation-metric, and GitHub could not classify the licence of that repository (NOASSERTION). The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

process_examples Ran
ran on a synthesized input
wjhou/icon/src_stage1/dataset_ende.py
code served (permissive licence) · get_code("47141bbf9a391e9d")
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 2, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import numpy as np

# Construct a minimal examples dict with the required keys
examples = {
    'id': ['report_001', 'report_002'],
    'image_path': ['/path/to/image1.jpg', '/path/to/image2.jpg']
}

ARGS = {
    'examples': examples
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, torch, tqdm — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple, ndim 2, dtype family float — a tuple output has no array shape of its own; none is stated
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
process_examples Ran
ran on a synthesized input
wjhou/recap/src_stage1/data_process_ende.py
code served (permissive licence) · get_code("a56bdab712ac729b")
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 2, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import numpy as np

torch.manual_seed(0)

ARGS = {
    'examples': {
        'id': ['report_001', 'report_002'],
        'image_path': ['/path/to/image1.jpg', '/path/to/image2.jpg'],
        'temporal_image_path': ['/path/to/temporal1.jpg', '/path/to/temporal2.jpg'],
        'temporal_predicate': ['before', 'after']
    }
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, torch, tqdm — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple, ndim 2, dtype family float — a tuple output has no array shape of its own; none is stated
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
stemming Ran
ran on a synthesized input
wjhou/organ/src/plan_extraction.py
code served (permissive licence) · get_code("41fd9b56fd1a9569")
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 1, 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.

from nltk.stem.porter import PorterStemmer

stemmer = PorterStemmer()
tokens = ["running", "quickly", "happily", "jumped"]

ARGS = {
    "stemmer": stemmer,
    "tokens": tokens,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
nltk — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type list, ndim 1, 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
compute_ce_metric Not yet run
unverified (harvested, no recorded run)
wjhou/icon/src_stage1/chexbert_eval.py
code served (permissive licence) · get_code("a5d48d0faff8e506")
How it ran

no invocation recorded.

load_chexbert Not yet run
unverified (harvested, no recorded run)
wjhou/icon/src_stage1/chexbert_eval.py
code served (permissive licence) · get_code("6b14457458f396df")
How it ran

no invocation recorded.

Also present in other papers' code

Of 17 samples lifted from this paper's linked repositories, 17 carry a function identity and 0 do not. Of the 17 distinct functions with an identity, 6 are also present in other papers whose code we hold (5 with at least one paper dated earlier, 0 the same month, 6 later) and 11 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
cook_refs 25 other papers
11 dated earlier, 14 later
1411.5726 (2014-11)
CIDEr: Consensus-based Image Description Evaluation
1502.03044 (2015-02)
Show, Attend and Tell: Neural Image Caption Generation with Visual Attention
1503.01817 (2015-03)
YFCC100M: The New Data in Multimedia Research
and 22 more not named here
cook_test 22 other papers
8 dated earlier, 14 later
1411.5726 (2014-11)
CIDEr: Consensus-based Image Description Evaluation
1502.03044 (2015-02)
Show, Attend and Tell: Neural Image Caption Generation with Visual Attention
1503.01817 (2015-03)
YFCC100M: The New Data in Multimedia Research
and 19 more not named here
precook 22 other papers
9 dated earlier, 13 later
1411.4555 (2014-11)
Show and Tell: A Neural Image Caption Generator
1502.03044 (2015-02)
Show, Attend and Tell: Neural Image Caption Generation with Visual Attention
1611.09268 (2016-11)
MS MARCO: A Human Generated MAchine Reading COmprehension Dataset
and 19 more not named here
cook_refs 21 other papers
8 dated earlier, 13 later
1502.03044 (2015-02)
Show, Attend and Tell: Neural Image Caption Generation with Visual Attention
1611.09268 (2016-11)
MS MARCO: A Human Generated MAchine Reading COmprehension Dataset
1611.09571 (2016-11)
Predicting Human Eye Fixations via an LSTM-based Saliency Attentive Model
and 18 more not named here
precook 14 other papers
4 dated earlier, 10 later
1411.5726 (2014-11)
CIDEr: Consensus-based Image Description Evaluation
1502.03044 (2015-02)
Show, Attend and Tell: Neural Image Caption Generation with Visual Attention
1806.00807 (2018-06)
Learning Semantic Sentence Embeddings using Sequential Pair-wise Discriminator
and 11 more not named here
cook_test 1 other paper
1 later
aaai_32596 (2025 (year only))
HC-LLM: Historical-Constrained Large Language Models for Radiology Report Generation

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 2026-11-21.

Repositories linked to this paper

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

Abstract

The extraction of labels from radiology text reports enables large-scale training of medical imaging models. Existing approaches to report labeling typically rely either on sophisticated feature engineering based on medical domain knowledge or manual annotations by experts. In this work, we introduce a BERT-based approach to medical image report labeling that exploits both the scale of available rule-based systems and the quality of expert annotations. We demonstrate superior performance of a biomedically pretrained BERT model first trained on annotations of a rule-based labeler and then finetuned on a small set of expert annotations augmented with automated backtranslation. We find that our final model, CheXbert, is able to outperform the previous best rules-based labeler with statistical significance, setting a new SOTA for report labeling on one of the largest datasets of chest x-rays.

For agents

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

get_harvested_code_for_paper("2004.09167")
get_code_for_paper("2004.09167")
have("2004.09167")

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

curl https://syntology.ai/api/ran/2004.09167.json

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

[![Syntology run record](https://syntology.ai/api/ran/2004.09167.svg)](https://syntology.ai/paper/2004.09167)

Connect an agent — have() is free.