feat: new dataset structure
Browse filesThis view is limited to 50 files because it contains too many changes. Β
See raw diff
- LOADING_TROUBLESHOOTING.md +0 -49
- README.md +56 -39
- TESTING.md +0 -173
- {train/texts β articles}/PMC10038974.md +0 -0
- {train/texts β articles}/PMC10085626.md +0 -0
- {train/texts β articles}/PMC10091789.md +0 -0
- {train/texts β articles}/PMC10099095.md +0 -0
- {train/texts β articles}/PMC10139129.md +0 -0
- {train/texts β articles}/PMC10145266.md +0 -0
- {train/texts β articles}/PMC10152845.md +0 -0
- {train/texts β articles}/PMC10154044.md +0 -0
- {train/texts β articles}/PMC10159199.md +0 -0
- {train/texts β articles}/PMC10163902.md +0 -0
- {test/texts β articles}/PMC10179231.md +0 -0
- {train/texts β articles}/PMC10196221.md +0 -0
- {train/texts β articles}/PMC10214567.md +0 -0
- {train/texts β articles}/PMC10230242.md +0 -0
- {train/texts β articles}/PMC10244018.md +0 -0
- {train/texts β articles}/PMC10275785.md +0 -0
- {test/texts β articles}/PMC10278212.md +0 -0
- {train/texts β articles}/PMC1029622.md +0 -0
- {train/texts β articles}/PMC10298263.md +0 -0
- {train/texts β articles}/PMC10309098.md +0 -0
- {train/texts β articles}/PMC10327396.md +0 -0
- {train/texts β articles}/PMC10337687.md +0 -0
- {train/texts β articles}/PMC10349379.md +0 -0
- {train/texts β articles}/PMC10349800.md +0 -0
- {train/texts β articles}/PMC10377184.md +0 -0
- {train/texts β articles}/PMC10381361.md +0 -0
- {train/texts β articles}/PMC10409991.md +0 -0
- {test/texts β articles}/PMC10418744.md +0 -0
- {train/texts β articles}/PMC10452379.md +0 -0
- {train/texts β articles}/PMC10463210.md +0 -0
- {train/texts β articles}/PMC10478012.md +0 -0
- {train/texts β articles}/PMC10483403.md +0 -0
- {val/texts β articles}/PMC10495004.md +0 -0
- {test/texts β articles}/PMC10499425.md +0 -0
- {train/texts β articles}/PMC10501538.md +0 -0
- {test/texts β articles}/PMC10502099.md +0 -0
- {train/texts β articles}/PMC10526247.md +0 -0
- {train/texts β articles}/PMC10527451.md +0 -0
- {train/texts β articles}/PMC10529681.md +0 -0
- {train/texts β articles}/PMC10532840.md +0 -0
- {train/texts β articles}/PMC10532907.md +0 -0
- {train/texts β articles}/PMC10537526.md +0 -0
- {train/texts β articles}/PMC10557961.md +0 -0
- {train/texts β articles}/PMC10565537.md +0 -0
- {test/texts β articles}/PMC10566653.md +0 -0
- {train/texts β articles}/PMC10582663.md +0 -0
- {train/texts β articles}/PMC10583240.md +0 -0
LOADING_TROUBLESHOOTING.md
DELETED
@@ -1,49 +0,0 @@
|
|
1 |
-
# Loading Troubleshooting
|
2 |
-
|
3 |
-
## CastError: Column Names Don't Match
|
4 |
-
|
5 |
-
If you encounter a `CastError` about column names not matching when loading the dataset from Hugging Face Hub, this is caused by schema conflicts between different file formats in the repository.
|
6 |
-
|
7 |
-
### Quick Fix
|
8 |
-
|
9 |
-
Use explicit JSONL data files:
|
10 |
-
|
11 |
-
```python
|
12 |
-
from datasets import load_dataset
|
13 |
-
|
14 |
-
dataset = load_dataset('json', data_files={
|
15 |
-
'train': 'hf://datasets/shlokn/autogkb/train.jsonl',
|
16 |
-
'validation': 'hf://datasets/shlokn/autogkb/val.jsonl',
|
17 |
-
'test': 'hf://datasets/shlokn/autogkb/test.jsonl'
|
18 |
-
})
|
19 |
-
```
|
20 |
-
|
21 |
-
### Why This Happens
|
22 |
-
|
23 |
-
The dataset repository contains both:
|
24 |
-
- `.jsonl` files with the processed dataset (recommended for most use cases)
|
25 |
-
- `.tsv` files with raw annotations (for manual inspection)
|
26 |
-
|
27 |
-
The Hugging Face datasets library auto-detects both formats and tries to merge them, but they have slightly different column names:
|
28 |
-
- JSONL files: `isplural`, `is_plural` (processed field names)
|
29 |
-
- TSV files: `isPlural` (original column name)
|
30 |
-
|
31 |
-
### The Fix
|
32 |
-
|
33 |
-
We've added a `.huggingfaceignore` file to prevent auto-detection of the TSV files. However, it may take some time for this change to propagate on the Hub.
|
34 |
-
|
35 |
-
### Alternative Loading Methods
|
36 |
-
|
37 |
-
See `load_dataset_examples.py` for multiple ways to load the dataset:
|
38 |
-
1. **Explicit JSONL files** (recommended, most reliable)
|
39 |
-
2. **Standard HF loading** (works after .huggingfaceignore is processed)
|
40 |
-
3. **Local loading** (if you have the repo cloned)
|
41 |
-
4. **Manual loading** (using pandas directly)
|
42 |
-
|
43 |
-
### Verification
|
44 |
-
|
45 |
-
Run the test script to verify loading works:
|
46 |
-
|
47 |
-
```bash
|
48 |
-
python test_hf_loading.py
|
49 |
-
```
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
README.md
CHANGED
@@ -20,8 +20,16 @@ multilinguality:
|
|
20 |
pretty_name: AutoGKB Annotation Benchmark
|
21 |
dataset_info:
|
22 |
features:
|
23 |
-
- name:
|
|
|
|
|
|
|
|
|
|
|
|
|
24 |
dtype: string
|
|
|
|
|
25 |
- name: variant_haplotypes
|
26 |
dtype: string
|
27 |
- name: gene
|
@@ -29,7 +37,7 @@ dataset_info:
|
|
29 |
- name: drugs
|
30 |
dtype: string
|
31 |
- name: pmid
|
32 |
-
dtype:
|
33 |
- name: phenotype_category
|
34 |
dtype: string
|
35 |
- name: significance
|
@@ -64,8 +72,6 @@ dataset_info:
|
|
64 |
dtype: string
|
65 |
- name: comparison_metabolizer_types
|
66 |
dtype: string
|
67 |
-
- name: text
|
68 |
-
dtype: string
|
69 |
splits:
|
70 |
- name: train
|
71 |
num_examples: 3124
|
@@ -102,25 +108,33 @@ Each example contains:
|
|
102 |
Example:
|
103 |
```json
|
104 |
{
|
105 |
-
"
|
106 |
-
"
|
107 |
-
"
|
108 |
-
"
|
109 |
-
"
|
110 |
-
"
|
111 |
-
"
|
112 |
-
"
|
113 |
-
"
|
114 |
-
"
|
115 |
-
"
|
116 |
-
"
|
117 |
-
"
|
|
|
|
|
|
|
|
|
118 |
}
|
119 |
```
|
120 |
|
121 |
### Data Fields
|
122 |
|
123 |
#### Core Fields
|
|
|
|
|
|
|
|
|
124 |
- `variant_annotation_id`: Unique identifier for each annotation
|
125 |
- `variant_haplotypes`: Genetic variant identifier (e.g., rs numbers, haplotypes)
|
126 |
- `gene`: Gene symbol (e.g., CYP2D6, ABCB1)
|
@@ -149,8 +163,6 @@ Example:
|
|
149 |
- `comparison_alleles_or_genotypes`: Reference genotypes for comparison
|
150 |
- `comparison_metabolizer_types`: Reference metabolizer types
|
151 |
|
152 |
-
#### Text Data
|
153 |
-
- `text`: Full text of the source scientific article in markdown format
|
154 |
|
155 |
### Data Splits
|
156 |
|
@@ -234,10 +246,10 @@ This dataset is released under the Apache License 2.0.
|
|
234 |
### Citation Information
|
235 |
|
236 |
```bibtex
|
237 |
-
@misc{
|
238 |
-
title={
|
239 |
-
author={
|
240 |
-
year={
|
241 |
note={A benchmark for pharmacogenomic variant-drug annotation extraction from scientific literature}
|
242 |
}
|
243 |
```
|
@@ -257,7 +269,8 @@ This dataset contributes to the biomedical NLP community by providing:
|
|
257 |
```python
|
258 |
from datasets import load_dataset
|
259 |
|
260 |
-
dataset
|
|
|
261 |
|
262 |
# Access different splits
|
263 |
train_data = dataset["train"]
|
@@ -268,6 +281,14 @@ test_data = dataset["test"]
|
|
268 |
efficacy_examples = train_data.filter(
|
269 |
lambda x: "Efficacy" in x["phenotype_category"]
|
270 |
)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
271 |
```
|
272 |
|
273 |
### Evaluation
|
@@ -288,21 +309,17 @@ python evaluate.py baseline_predictions.tsv val/annotations.tsv --output results
|
|
288 |
## File Structure
|
289 |
|
290 |
```
|
291 |
-
|
292 |
-
βββ
|
293 |
-
β βββ
|
294 |
-
β
|
295 |
-
|
296 |
-
|
297 |
-
|
298 |
-
βββ test
|
299 |
-
|
300 |
-
|
301 |
-
βββ dataset_statistics.json
|
302 |
-
βββ evaluate.py # Evaluation script
|
303 |
-
βββ baseline_model.py # Rule-based baseline
|
304 |
-
βββ variant_drug_benchmark.py # HuggingFace dataset script
|
305 |
-
βββ dataset_infos.json # Dataset metadata
|
306 |
βββ LICENSE # Apache 2.0 license
|
307 |
βββ README.md # This file
|
308 |
```
|
|
|
20 |
pretty_name: AutoGKB Annotation Benchmark
|
21 |
dataset_info:
|
22 |
features:
|
23 |
+
- name: pmcid
|
24 |
+
dtype: string
|
25 |
+
- name: article_title
|
26 |
+
dtype: string
|
27 |
+
- name: article_path
|
28 |
+
dtype: string
|
29 |
+
- name: article_text
|
30 |
dtype: string
|
31 |
+
- name: variant_annotation_id
|
32 |
+
dtype: int64
|
33 |
- name: variant_haplotypes
|
34 |
dtype: string
|
35 |
- name: gene
|
|
|
37 |
- name: drugs
|
38 |
dtype: string
|
39 |
- name: pmid
|
40 |
+
dtype: int64
|
41 |
- name: phenotype_category
|
42 |
dtype: string
|
43 |
- name: significance
|
|
|
72 |
dtype: string
|
73 |
- name: comparison_metabolizer_types
|
74 |
dtype: string
|
|
|
|
|
75 |
splits:
|
76 |
- name: train
|
77 |
num_examples: 3124
|
|
|
108 |
Example:
|
109 |
```json
|
110 |
{
|
111 |
+
"pmcid": "PMC6714673",
|
112 |
+
"article_title": "Warfarin Dose Model for the Prediction of Stable Maintenance Dose in Indian Patients",
|
113 |
+
"article_path": "articles/PMC6714673.md",
|
114 |
+
"article_text": "# Warfarin Dose Model for the Prediction of Stable Maintenance Dose in Indian Patients\n\n## Abstract\n\nWarfarin is a commonly used anticoagulant...",
|
115 |
+
"variant_annotation_id": 1449192282,
|
116 |
+
"variant_haplotypes": "rs1799853",
|
117 |
+
"gene": "CYP2C9",
|
118 |
+
"drugs": "warfarin",
|
119 |
+
"pmid": 28049362,
|
120 |
+
"phenotype_category": "Dosage",
|
121 |
+
"significance": "yes",
|
122 |
+
"sentence": "Genotype CT is associated with decreased dose of warfarin as compared to genotype CC.",
|
123 |
+
"alleles": "CT",
|
124 |
+
"is_is_not_associated": "Associated with",
|
125 |
+
"direction_of_effect": "decreased",
|
126 |
+
"pd_pk_terms": "dose of",
|
127 |
+
"comparison_alleles_or_genotypes": "CC"
|
128 |
}
|
129 |
```
|
130 |
|
131 |
### Data Fields
|
132 |
|
133 |
#### Core Fields
|
134 |
+
- `pmcid`: PubMed Central identifier of the source article
|
135 |
+
- `article_title`: Title of the source scientific article
|
136 |
+
- `article_path`: Relative path to the article file (markdown format)
|
137 |
+
- `article_text`: Full text of the scientific article in markdown format
|
138 |
- `variant_annotation_id`: Unique identifier for each annotation
|
139 |
- `variant_haplotypes`: Genetic variant identifier (e.g., rs numbers, haplotypes)
|
140 |
- `gene`: Gene symbol (e.g., CYP2D6, ABCB1)
|
|
|
163 |
- `comparison_alleles_or_genotypes`: Reference genotypes for comparison
|
164 |
- `comparison_metabolizer_types`: Reference metabolizer types
|
165 |
|
|
|
|
|
166 |
|
167 |
### Data Splits
|
168 |
|
|
|
246 |
### Citation Information
|
247 |
|
248 |
```bibtex
|
249 |
+
@misc{autogkb_annotation_benchmark_2025,
|
250 |
+
title={AutoGKB Annotation Benchmark},
|
251 |
+
author={Shlok Natarajan},
|
252 |
+
year={2025},
|
253 |
note={A benchmark for pharmacogenomic variant-drug annotation extraction from scientific literature}
|
254 |
}
|
255 |
```
|
|
|
269 |
```python
|
270 |
from datasets import load_dataset
|
271 |
|
272 |
+
# Load the dataset from Hugging Face Hub
|
273 |
+
dataset = load_dataset("autogkb/autogkb-annotation-benchmark")
|
274 |
|
275 |
# Access different splits
|
276 |
train_data = dataset["train"]
|
|
|
281 |
efficacy_examples = train_data.filter(
|
282 |
lambda x: "Efficacy" in x["phenotype_category"]
|
283 |
)
|
284 |
+
|
285 |
+
# Example: Access article text for a specific annotation
|
286 |
+
first_example = train_data[0]
|
287 |
+
print(f"PMC ID: {first_example['pmcid']}")
|
288 |
+
print(f"Article Title: {first_example['article_title']}")
|
289 |
+
print(f"Gene: {first_example['gene']}")
|
290 |
+
print(f"Drug: {first_example['drugs']}")
|
291 |
+
print(f"Full Article Text: {first_example['article_text'][:500]}...")
|
292 |
```
|
293 |
|
294 |
### Evaluation
|
|
|
309 |
## File Structure
|
310 |
|
311 |
```
|
312 |
+
autogkb/
|
313 |
+
βββ articles/ # Full article texts in markdown format
|
314 |
+
β βββ PMC10038974.md
|
315 |
+
β βββ PMC10085626.md
|
316 |
+
β βββ ... # 1,431 articles total
|
317 |
+
βββ train.jsonl # Training annotations (3,124 examples)
|
318 |
+
βββ val.jsonl # Validation annotations (796 examples)
|
319 |
+
βββ test.jsonl # Test annotations (596 examples)
|
320 |
+
βββ autogkb_annotation_benchmark.py # HuggingFace dataset script
|
321 |
+
βββ dataset_infos.json # Dataset metadata
|
322 |
+
βββ dataset_statistics.json # Dataset statistics
|
|
|
|
|
|
|
|
|
323 |
βββ LICENSE # Apache 2.0 license
|
324 |
βββ README.md # This file
|
325 |
```
|
TESTING.md
DELETED
@@ -1,173 +0,0 @@
|
|
1 |
-
# AutoGKB Dataset Testing Guide
|
2 |
-
|
3 |
-
This directory contains comprehensive testing utilities for the AutoGKB Annotation Benchmark dataset to ensure reliable loading and data integrity.
|
4 |
-
|
5 |
-
## Quick Start
|
6 |
-
|
7 |
-
```bash
|
8 |
-
# Run quick test to verify everything works
|
9 |
-
python load_dataset_safe.py test
|
10 |
-
|
11 |
-
# Run comprehensive simple tests
|
12 |
-
python test_dataset_simple.py
|
13 |
-
|
14 |
-
# Run full test suite (may take longer)
|
15 |
-
python test_dataset_loading.py
|
16 |
-
```
|
17 |
-
|
18 |
-
## Test Files
|
19 |
-
|
20 |
-
### `test_dataset_simple.py`
|
21 |
-
**Recommended for routine testing**
|
22 |
-
- Fast execution (< 30 seconds)
|
23 |
-
- Tests TSV structure and headers
|
24 |
-
- Tests dataset loading with small samples
|
25 |
-
- Tests data content quality
|
26 |
-
- Good for CI/CD pipelines
|
27 |
-
|
28 |
-
### `test_dataset_loading.py`
|
29 |
-
**Comprehensive test suite**
|
30 |
-
- Full unittest framework
|
31 |
-
- Tests cache handling
|
32 |
-
- Tests multiple loading methods
|
33 |
-
- Tests data integrity across all splits
|
34 |
-
- More thorough but slower
|
35 |
-
|
36 |
-
### `load_dataset_safe.py`
|
37 |
-
**Production-ready loading utility**
|
38 |
-
- Handles cache issues automatically
|
39 |
-
- Multiple fallback loading methods
|
40 |
-
- Safe for use in development and production
|
41 |
-
- Includes quick testing function
|
42 |
-
|
43 |
-
## Common Issues and Solutions
|
44 |
-
|
45 |
-
### Cache Issues
|
46 |
-
The HuggingFace datasets library can have caching conflicts. Our utilities automatically:
|
47 |
-
- Disable caching when needed
|
48 |
-
- Clear problematic cache directories
|
49 |
-
- Provide fallback loading methods
|
50 |
-
|
51 |
-
### Loading Methods
|
52 |
-
The safe loader tries multiple approaches:
|
53 |
-
1. **Manual generation**: Direct data loading without caching
|
54 |
-
2. **Temporary disk**: Save to temp location and reload
|
55 |
-
3. **CSV fallback**: Direct pandas loading as last resort
|
56 |
-
|
57 |
-
## Usage Examples
|
58 |
-
|
59 |
-
### Basic Loading
|
60 |
-
```python
|
61 |
-
from load_dataset_safe import load_autogkb_safe
|
62 |
-
|
63 |
-
# Load small sample for testing
|
64 |
-
dataset = load_autogkb_safe(split="train", max_examples=10)
|
65 |
-
|
66 |
-
# Load specific split
|
67 |
-
train_data = load_autogkb_safe(split="train")
|
68 |
-
|
69 |
-
# Load all splits
|
70 |
-
dataset_dict = load_autogkb_safe()
|
71 |
-
```
|
72 |
-
|
73 |
-
### Advanced Options
|
74 |
-
```python
|
75 |
-
# Disable caching (recommended for development)
|
76 |
-
dataset = load_autogkb_safe(use_cache=False)
|
77 |
-
|
78 |
-
# Load from specific directory
|
79 |
-
dataset = load_autogkb_safe(data_dir="/path/to/dataset")
|
80 |
-
|
81 |
-
# Quick test function
|
82 |
-
from load_dataset_safe import quick_test
|
83 |
-
success = quick_test(max_examples=5)
|
84 |
-
```
|
85 |
-
|
86 |
-
## Dataset Validation
|
87 |
-
|
88 |
-
The tests verify:
|
89 |
-
|
90 |
-
### β
File Structure
|
91 |
-
- All required TSV files exist (`train/`, `val/`, `test/`)
|
92 |
-
- Correct column headers (22 columns matching schema)
|
93 |
-
- Non-empty files
|
94 |
-
|
95 |
-
### β
Data Loading
|
96 |
-
- Dataset builder initialization
|
97 |
-
- Manual data generation
|
98 |
-
- Cache-aware loading
|
99 |
-
- Fallback methods
|
100 |
-
|
101 |
-
### β
Data Integrity
|
102 |
-
- No duplicate IDs within splits
|
103 |
-
- Required fields are non-empty
|
104 |
-
- Correct data types (all strings)
|
105 |
-
- PMID format validation
|
106 |
-
|
107 |
-
### β
Text Matching
|
108 |
-
- PMID to text file matching
|
109 |
-
- Text content availability
|
110 |
-
- Cross-reference validation
|
111 |
-
|
112 |
-
## Expected Test Output
|
113 |
-
|
114 |
-
```
|
115 |
-
π§ͺ AutoGKB Dataset Simple Test Suite
|
116 |
-
==================================================
|
117 |
-
π Testing TSV file structure...
|
118 |
-
β
train/annotations.tsv: 22 columns
|
119 |
-
β
val/annotations.tsv: 22 columns
|
120 |
-
β
test/annotations.tsv: 22 columns
|
121 |
-
β
TSV structure test passed!
|
122 |
-
|
123 |
-
π¦ Testing dataset loading...
|
124 |
-
β
train: 3 examples loaded
|
125 |
-
β
validation: 3 examples loaded
|
126 |
-
β
test: 3 examples loaded
|
127 |
-
β
Dataset loading test passed!
|
128 |
-
|
129 |
-
π Testing full dataset loading...
|
130 |
-
β
Full dataset test passed!
|
131 |
-
Splits: ['train', 'validation', 'test']
|
132 |
-
train: 5 examples
|
133 |
-
validation: 5 examples
|
134 |
-
test: 5 examples
|
135 |
-
|
136 |
-
π Testing data content quality...
|
137 |
-
β
Data content test passed! Checked 10 examples
|
138 |
-
|
139 |
-
π All tests passed!
|
140 |
-
```
|
141 |
-
|
142 |
-
## Troubleshooting
|
143 |
-
|
144 |
-
### If tests fail:
|
145 |
-
|
146 |
-
1. **Check file structure**: Ensure `train/`, `val/`, `test/` directories exist with `annotations.tsv` files
|
147 |
-
2. **Clear cache**: Run `python -c "from load_dataset_safe import clear_datasets_cache; clear_datasets_cache()"`
|
148 |
-
3. **Check dependencies**: Ensure `datasets`, `pandas` are installed
|
149 |
-
4. **Run with verbose output**: Add print statements or use the comprehensive test suite
|
150 |
-
|
151 |
-
### Common error messages:
|
152 |
-
|
153 |
-
- **"Required file missing"**: Check that TSV files exist in correct locations
|
154 |
-
- **"Column mismatch"**: Verify TSV headers match schema (run column header test)
|
155 |
-
- **"Cache issues"**: Use `use_cache=False` or clear cache directory
|
156 |
-
- **"Loading failed"**: Try fallback methods or check file permissions
|
157 |
-
|
158 |
-
## Integration with CI/CD
|
159 |
-
|
160 |
-
For automated testing:
|
161 |
-
|
162 |
-
```bash
|
163 |
-
# Add to your CI pipeline
|
164 |
-
python test_dataset_simple.py
|
165 |
-
if [ $? -eq 0 ]; then
|
166 |
-
echo "Dataset tests passed"
|
167 |
-
else
|
168 |
-
echo "Dataset tests failed"
|
169 |
-
exit 1
|
170 |
-
fi
|
171 |
-
```
|
172 |
-
|
173 |
-
The simple test suite is designed to be fast and reliable for continuous integration environments.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
{train/texts β articles}/PMC10038974.md
RENAMED
File without changes
|
{train/texts β articles}/PMC10085626.md
RENAMED
File without changes
|
{train/texts β articles}/PMC10091789.md
RENAMED
File without changes
|
{train/texts β articles}/PMC10099095.md
RENAMED
File without changes
|
{train/texts β articles}/PMC10139129.md
RENAMED
File without changes
|
{train/texts β articles}/PMC10145266.md
RENAMED
File without changes
|
{train/texts β articles}/PMC10152845.md
RENAMED
File without changes
|
{train/texts β articles}/PMC10154044.md
RENAMED
File without changes
|
{train/texts β articles}/PMC10159199.md
RENAMED
File without changes
|
{train/texts β articles}/PMC10163902.md
RENAMED
File without changes
|
{test/texts β articles}/PMC10179231.md
RENAMED
File without changes
|
{train/texts β articles}/PMC10196221.md
RENAMED
File without changes
|
{train/texts β articles}/PMC10214567.md
RENAMED
File without changes
|
{train/texts β articles}/PMC10230242.md
RENAMED
File without changes
|
{train/texts β articles}/PMC10244018.md
RENAMED
File without changes
|
{train/texts β articles}/PMC10275785.md
RENAMED
File without changes
|
{test/texts β articles}/PMC10278212.md
RENAMED
File without changes
|
{train/texts β articles}/PMC1029622.md
RENAMED
File without changes
|
{train/texts β articles}/PMC10298263.md
RENAMED
File without changes
|
{train/texts β articles}/PMC10309098.md
RENAMED
File without changes
|
{train/texts β articles}/PMC10327396.md
RENAMED
File without changes
|
{train/texts β articles}/PMC10337687.md
RENAMED
File without changes
|
{train/texts β articles}/PMC10349379.md
RENAMED
File without changes
|
{train/texts β articles}/PMC10349800.md
RENAMED
File without changes
|
{train/texts β articles}/PMC10377184.md
RENAMED
File without changes
|
{train/texts β articles}/PMC10381361.md
RENAMED
File without changes
|
{train/texts β articles}/PMC10409991.md
RENAMED
File without changes
|
{test/texts β articles}/PMC10418744.md
RENAMED
File without changes
|
{train/texts β articles}/PMC10452379.md
RENAMED
File without changes
|
{train/texts β articles}/PMC10463210.md
RENAMED
File without changes
|
{train/texts β articles}/PMC10478012.md
RENAMED
File without changes
|
{train/texts β articles}/PMC10483403.md
RENAMED
File without changes
|
{val/texts β articles}/PMC10495004.md
RENAMED
File without changes
|
{test/texts β articles}/PMC10499425.md
RENAMED
File without changes
|
{train/texts β articles}/PMC10501538.md
RENAMED
File without changes
|
{test/texts β articles}/PMC10502099.md
RENAMED
File without changes
|
{train/texts β articles}/PMC10526247.md
RENAMED
File without changes
|
{train/texts β articles}/PMC10527451.md
RENAMED
File without changes
|
{train/texts β articles}/PMC10529681.md
RENAMED
File without changes
|
{train/texts β articles}/PMC10532840.md
RENAMED
File without changes
|
{train/texts β articles}/PMC10532907.md
RENAMED
File without changes
|
{train/texts β articles}/PMC10537526.md
RENAMED
File without changes
|
{train/texts β articles}/PMC10557961.md
RENAMED
File without changes
|
{train/texts β articles}/PMC10565537.md
RENAMED
File without changes
|
{test/texts β articles}/PMC10566653.md
RENAMED
File without changes
|
{train/texts β articles}/PMC10582663.md
RENAMED
File without changes
|
{train/texts β articles}/PMC10583240.md
RENAMED
File without changes
|