Duplicate from snoroozi/pmoa-tts

Co-authored-by: Shahriar <snoroozi@users.noreply.huggingface.co>

.gitattributes

@@ -0,0 +1,60 @@
+*.7z filter=lfs diff=lfs merge=lfs -text
+*.arrow filter=lfs diff=lfs merge=lfs -text
+*.bin filter=lfs diff=lfs merge=lfs -text
+*.bz2 filter=lfs diff=lfs merge=lfs -text
+*.ckpt filter=lfs diff=lfs merge=lfs -text
+*.ftz filter=lfs diff=lfs merge=lfs -text
+*.gz filter=lfs diff=lfs merge=lfs -text
+*.h5 filter=lfs diff=lfs merge=lfs -text
+*.joblib filter=lfs diff=lfs merge=lfs -text
+*.lfs.* filter=lfs diff=lfs merge=lfs -text
+*.lz4 filter=lfs diff=lfs merge=lfs -text
+*.mds filter=lfs diff=lfs merge=lfs -text
+*.mlmodel filter=lfs diff=lfs merge=lfs -text
+*.model filter=lfs diff=lfs merge=lfs -text
+*.msgpack filter=lfs diff=lfs merge=lfs -text
+*.npy filter=lfs diff=lfs merge=lfs -text
+*.npz filter=lfs diff=lfs merge=lfs -text
+*.onnx filter=lfs diff=lfs merge=lfs -text
+*.ot filter=lfs diff=lfs merge=lfs -text
+*.parquet filter=lfs diff=lfs merge=lfs -text
+*.pb filter=lfs diff=lfs merge=lfs -text
+*.pickle filter=lfs diff=lfs merge=lfs -text
+*.pkl filter=lfs diff=lfs merge=lfs -text
+*.pt filter=lfs diff=lfs merge=lfs -text
+*.pth filter=lfs diff=lfs merge=lfs -text
+*.rar filter=lfs diff=lfs merge=lfs -text
+*.safetensors filter=lfs diff=lfs merge=lfs -text
+saved_model/**/* filter=lfs diff=lfs merge=lfs -text
+*.tar.* filter=lfs diff=lfs merge=lfs -text
+*.tar filter=lfs diff=lfs merge=lfs -text
+*.tflite filter=lfs diff=lfs merge=lfs -text
+*.tgz filter=lfs diff=lfs merge=lfs -text
+*.wasm filter=lfs diff=lfs merge=lfs -text
+*.xz filter=lfs diff=lfs merge=lfs -text
+*.zip filter=lfs diff=lfs merge=lfs -text
+*.zst filter=lfs diff=lfs merge=lfs -text
+*tfevents* filter=lfs diff=lfs merge=lfs -text
+# Audio files - uncompressed
+*.pcm filter=lfs diff=lfs merge=lfs -text
+*.sam filter=lfs diff=lfs merge=lfs -text
+*.raw filter=lfs diff=lfs merge=lfs -text
+# Audio files - compressed
+*.aac filter=lfs diff=lfs merge=lfs -text
+*.flac filter=lfs diff=lfs merge=lfs -text
+*.mp3 filter=lfs diff=lfs merge=lfs -text
+*.ogg filter=lfs diff=lfs merge=lfs -text
+*.wav filter=lfs diff=lfs merge=lfs -text
+# Image files - uncompressed
+*.bmp filter=lfs diff=lfs merge=lfs -text
+*.gif filter=lfs diff=lfs merge=lfs -text
+*.png filter=lfs diff=lfs merge=lfs -text
+*.tiff filter=lfs diff=lfs merge=lfs -text
+# Image files - compressed
+*.jpg filter=lfs diff=lfs merge=lfs -text
+*.jpeg filter=lfs diff=lfs merge=lfs -text
+*.webp filter=lfs diff=lfs merge=lfs -text
+# Video files - compressed
+*.mp4 filter=lfs diff=lfs merge=lfs -text
+*.webm filter=lfs diff=lfs merge=lfs -text
+*.jsonl filter=lfs diff=lfs merge=lfs -text

README.md

@@ -0,0 +1,172 @@
+---
+task_categories:
+- text-classification
+- time-series-forecasting
+language:
+- en
+tags:
+- clinical,
+- time-series,
+- biomedical
+- text
+pretty_name: PMOA-TTS
+size_categories:
+- 100K<n<1M
+dataset_info:
+  features:
+  - name: pmc_id
+    dtype: string
+  - name: case_report_id
+    dtype: string
+  - name: textual_timeseries
+    list:
+    - name: event
+      dtype: string
+    - name: time
+      dtype: int64
+  - name: demographics
+    struct:
+    - name: age
+      dtype: string
+    - name: sex
+      dtype: string
+    - name: ethnicity
+      dtype: string
+  - name: diagnoses
+    sequence: string
+  - name: death_info
+    struct:
+    - name: observed_time
+      dtype: float64
+    - name: death_event_indicator
+      dtype: int64
+  splits:
+  - name: train
+    num_bytes: 240122853
+    num_examples: 124349
+  - name: case_study_100
+    num_bytes: 205311
+    num_examples: 88
+  - name: case_study_25k_L33
+    num_bytes: 47652572
+    num_examples: 24746
+  - name: case_study_25k_DSR1
+    num_bytes: 63989161
+    num_examples: 24746
+  download_size: 211632200
+  dataset_size: 351969897
+configs:
+- config_name: default
+  data_files:
+  - split: train
+    path: data/train-*
+  - split: case_study_100
+    path: data/case_study_100-*
+  - split: case_study_25k_L33
+    path: data/case_study_25k_L33-*
+  - split: case_study_25k_DSR1
+    path: data/case_study_25k_DSR1-*
+license: cc-by-nc-sa-4.0
+---
+# PMOA-TTS: Textual Time Series from PubMed Case Reports
+
+**PMOA-TTS** is a dataset of structured textual time series derived from 124k clinical case reports published in PubMed Open Access. Each data point corresponds to a single patient case and includes a sequence of timestamped clinical events extracted using large language models. This dataset is intended to support research in temporal modeling, survival analysis, event forecasting, and multimodal representation learning.
+
+This release contains the full dataset preprocessed from raw annotated files generated by LLaMA 3.3. A future version will include annotations from DeepSeek-R1 for the full dataset to enable comparative modeling. The current DeepSeek-R1 annotations are for the 25k subset presented in the survival experiments of our paper.
+
+---
+
+## 💾 Dataset Summary
+
+Each data point corresponds to a case report and includes:
+
+- A **textual time series** of clinical events with timestamps.
+- **Demographic information**: age, sex, and ethnicity (when available).
+- A set of **diagnoses** extracted from the same report.
+- A **death phenotype label** indicating whether the patient died or was censored, and the observed time.
+
+---
+
+## 📂 Dataset Splits
+
+- **`train` [default]**: The full dataset of 124k single-patient case reports automatically extracted using LLaMA 3.3. This is the default split and contains the primary annotations used in downstream tasks.
+- **`case_study_100`**: A curated benchmark of 100 case reports (88 present in this dataset), used in the paper to evaluate LLM-based vs. metadata-based case report identification. Each sample has been manually reviewed for single-case validity across five different diagnoses.
+- **`case_study_25k_L33`**: A 25k-case subset of LLaMA 3.3-annotated reports used for downstream survival analysis experiments in the paper. This split is designed to support modeling and evaluation of time-to-event outcomes on a more computationally manageable subset.
+- **`case_study_25k_DSR1`**: A 25k-case subset of DeepSeek-R1-annotated reports used for downstream survival analysis experiments in the paper. The `textual_timeseries` and `death_info` fields in this split are reconstructed from DeepSeek-R1 outputs, enabling comparative analysis between models across the same patient cohort.
+
+## 📁 Data Fields
+
+Each row in the dataset is a JSON object with the following fields:
+
+| Field | Type | Description |
+|-------|------|-------------|
+| `pmc_id` | `str` | Folder-level ID for the PubMed case report (`PMC000xxxxxx` to `PMC011xxxxxx`) |
+| `case_report_id` | `str` | Full filename for the case (e.g., `PMC6417290`) |
+| `textual_timeseries` | `List[Dict]` | Time series of `{ "event": str, "time": int }` |
+| `demographics` | `Dict` | `{"age": str or "Not Specified", "sex": str or "Not Specified", "ethnicity": str or "Not Specified"}` |
+| `diagnoses` | `List[str]` | List of diagnosis terms extracted per case |
+| `death_info` | `Dict` |`{"observed_time": int, "death_event_indicator": 0 or 1}` |
+
+### Example JSON entry
+
+```json
+{
+  "pmc_id": "PMC006xxxxxx",
+  "case_report_id": "PMC6417290",
+  "textual_timeseries": [
+    {"event": "56 years old", "time": 0}, 
+    {"event": "male", "time": 0}, 
+    {"event": "HIV-positive", "time": 0}, 
+    {"event": "admitted to the hospital", "time": 0}, 
+    {"event": "knee arthralgia", "time": 0}, 
+    ...
+    {"event": "postural headache", "time": 48}, 
+    ...
+    {"event": "neurological examination", "time": 120}, 
+    ...
+    {"event": "persistence of headache", "time": 144}, 
+    {"event": "brain computed tomography scan", "time": 144}
+    ...
+    {"event": "symptom free", "time": 4320}, 
+    {"event": "CT scan", "time": 4320}, 
+    {"event": "complete resolution of subdural hematoma", "time": 4320}, 
+    {"event": "no brain shift", "time": 4320}
+  ],
+  "diagnoses": [
+    "HIV", 
+    "Knee arthralgia", 
+    "Subdural hematoma", 
+    "Postdural puncture headache (PDPH)"
+  ],
+  "death_info": {
+    "observed_time": 4230,
+    "death_event_indicator": 0
+  },
+  "demographics": {
+    "age": 56,
+    "sex": "Male",
+    "ethnicity": "Not Specified"
+  }
+}
+```
+
+## 🔍 Intended Use
+
+This dataset is intended for research in:
+
+- Time series modeling from unstructured text
+- Temporal representation learning
+- Survival analysis (e.g., using `death_info`)
+- Clinical forecasting (next-event prediction)
+- Multimodal clinical modeling (text, time, demographics, outcomes)
+
+## 📝 License
+
+The dataset is derived from publicly available PubMed Open Access case reports. All annotations and metadata are released under the **CC BY NC SA 4.0** license.
+
+## 🧩 Coming Soon
+- DeepSeek-R1 annotations for full 125k dataset. (DeepSeek-R1 annotations for the 25k subset is already available).
+
+## 🙏 Acknowledgments
+
+This dataset was generated as part of research at National Library of Medicine (NLM) at National Institutes of Health (NIH) and Carnegie Mellon University. We thank the PubMed Open Access initiative and the authors of case reports that made this work possible.

data/case_study_100-00000-of-00001.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b2310285504256c3e5bfda4f11d1b63132a3618841cb80d780e7441ec4f9b575
+size 104917

data/case_study_25k_DSR1-00000-of-00001.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:39732d6faf55abcd5d15484502649acc789509a41626bcac425eb415bb2eb026
+size 28839947

data/case_study_25k_L33-00000-of-00001.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:96dcb78c2c0f317a7f5efc33408256a5931b8c9607baf2af23558747024ec862
+size 22019579

data/train-00000-of-00001.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6a4b828a9ffda9230629b2fa29d81dda24a72e5c6911b4f31562b75cb404fb9f
+size 109822131

tts_dataset.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1aac58db4cd2c1a874e4713130fdddb900b1a8c9aab87e396b50c80318088486
+size 346093140

tts_dataset_case_study_100.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ddfb678a75857159c865c2d13ffcbb0d9a061d63a4240c9584a4a5dbb6ef14ac
+size 293355

tts_dataset_case_study_25k_DSR1.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:42eb53efa18da788d5e4415169479c034088b76715dbf44cbf95ddc9937d4117
+size 91104833

tts_dataset_case_study_25k_L33.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6c13531242e4c0880403046bcf2cdbaeec2671668d984aab6d99bcecf7a74198
+size 68686344