open-icu 0.0.1__tar.gz

open_icu-0.0.1/.gitignore

@@ -0,0 +1,166 @@
+# Byte-compiled / optimized / DLL files
+__pycache__/
+*.py[cod]
+*$py.class
+
+# C extensions
+*.so
+
+# Distribution / packaging
+.Python
+build/
+develop-eggs/
+dist/
+downloads/
+eggs/
+.eggs/
+lib/
+lib64/
+parts/
+sdist/
+var/
+wheels/
+share/python-wheels/
+*.egg-info/
+.installed.cfg
+*.egg
+MANIFEST
+
+# PyInstaller
+#  Usually these files are written by a python script from a template
+#  before PyInstaller builds the exe, so as to inject date/other infos into it.
+*.manifest
+*.spec
+
+# Installer logs
+pip-log.txt
+pip-delete-this-directory.txt
+
+# Unit test / coverage reports
+htmlcov/
+.tox/
+.nox/
+.coverage
+.coverage.*
+.cache
+nosetests.xml
+coverage.xml
+*.cover
+*.py,cover
+.hypothesis/
+.pytest_cache/
+cover/
+
+# Translations
+*.mo
+*.pot
+
+# Django stuff:
+*.log
+local_settings.py
+db.sqlite3
+db.sqlite3-journal
+
+# Flask stuff:
+instance/
+.webassets-cache
+
+# Scrapy stuff:
+.scrapy
+
+# Sphinx documentation
+docs/_build/
+
+# PyBuilder
+.pybuilder/
+target/
+
+# Jupyter Notebook
+.ipynb_checkpoints
+
+# IPython
+profile_default/
+ipython_config.py
+
+# pyenv
+#   For a library or package, you might want to ignore these files since the code is
+#   intended to run in multiple environments; otherwise, check them in:
+# .python-version
+
+# pipenv
+#   According to pypa/pipenv#598, it is recommended to include Pipfile.lock in version control.
+#   However, in case of collaboration, if having platform-specific dependencies or dependencies
+#   having no cross-platform support, pipenv may install dependencies that don't work, or not
+#   install all needed dependencies.
+#Pipfile.lock
+
+# poetry
+#   Similar to Pipfile.lock, it is generally recommended to include poetry.lock in version control.
+#   This is especially recommended for binary packages to ensure reproducibility, and is more
+#   commonly ignored for libraries.
+#   https://python-poetry.org/docs/basic-usage/#commit-your-poetrylock-file-to-version-control
+#poetry.lock
+
+# pdm
+#   Similar to Pipfile.lock, it is generally recommended to include pdm.lock in version control.
+#pdm.lock
+#   pdm stores project-wide configurations in .pdm.toml, but it is recommended to not include it
+#   in version control.
+#   https://pdm.fming.dev/latest/usage/project/#working-with-version-control
+.pdm.toml
+.pdm-python
+.pdm-build/
+
+# PEP 582; used by e.g. github.com/David-OConnor/pyflow and github.com/pdm-project/pdm
+__pypackages__/
+
+# Celery stuff
+celerybeat-schedule
+celerybeat.pid
+
+# SageMath parsed files
+*.sage.py
+
+# Environments
+.venv
+env/
+venv/
+ENV/
+env.bak/
+venv.bak/
+
+# Spyder project settings
+.spyderproject
+.spyproject
+
+# Rope project settings
+.ropeproject
+
+# mkdocs documentation
+/site
+
+# mypy
+.mypy_cache/
+.dmypy.json
+dmypy.json
+
+# Pyre type checker
+.pyre/
+
+# pytype static type analyzer
+.pytype/
+
+# Cython debug symbols
+cython_debug/
+
+# PyCharm
+#  JetBrains specific template is maintained in a separate JetBrains.gitignore that can
+#  be found at https://github.com/github/gitignore/blob/main/Global/JetBrains.gitignore
+#  and can be added to the global gitignore or merged into this file.  For a more nuclear
+#  option (not recommended) you can uncomment the following to ignore the entire idea folder.
+#.idea/
+
+# pipeline output
+examples/output/
+.env
+.claude/*

open_icu-0.0.1/LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2024 AIDH MS
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

open_icu-0.0.1/PKG-INFO

@@ -0,0 +1,310 @@
+Metadata-Version: 2.4
+Name: open-icu
+Version: 0.0.1
+Summary: This is a project to unify and analyse data from different ICU data sources.
+Project-URL: homepage, https://github.com/aidh-ms/OpenICU
+Project-URL: repository, https://github.com/aidh-ms/OpenICU
+Project-URL: changelog, https://github.com/aidh-ms/OpenICU/releases
+Project-URL: issues, https://github.com/aidh-ms/OpenICU/issues
+Project-URL: documentation, https://github.com/aidh-ms/OpenICU/wiki
+Author: Paul-B98, TobiasVeronaLaser, prockenschaub
+Maintainer: Paul-B98, TobiasVeronaLaser, prockenschaub
+License-Expression: MIT
+License-File: LICENSE
+Keywords: Data,Healthcare,ICU
+Classifier: License :: OSI Approved :: MIT License
+Classifier: Operating System :: OS Independent
+Classifier: Programming Language :: Python :: 3
+Requires-Python: >=3.13
+Requires-Dist: meds>=0.4.0
+Requires-Dist: polars>=1.41.2
+Requires-Dist: pyarrow>=24.0.0
+Requires-Dist: pydantic>=2.13.4
+Requires-Dist: pyyaml>=6.0.2
+Requires-Dist: types-pyyaml>=6.0.12.20260518
+Description-Content-Type: text/markdown
+
+# OpenICU
+
+[![CI](https://github.com/aidh-ms/OpenICU/actions/workflows/continuous_integration.yml/badge.svg)](https://github.com/aidh-ms/OpenICU/actions/workflows/continuous_integration.yml)
+[![Coverage Status](https://coveralls.io/repos/github/aidh-ms/OpenICU/badge.svg?branch=main)](https://coveralls.io/github/aidh-ms/OpenICU?branch=main)
+[![Python](https://img.shields.io/badge/python-3.13%2B-blue)](https://github.com/aidh-ms/OpenICU/blob/main/pyproject.toml)
+[![License: MIT](https://img.shields.io/badge/license-MIT-green.svg)](LICENSE)
+
+**OpenICU** is an open-source Python framework for extracting and harmonising intensive care unit (ICU) data from heterogeneous sources — such as [MIMIC-IV](https://physionet.org/content/mimiciv/), [eICU-CRD](https://physionet.org/content/eicu-crd/), and your own institutional database exports — into the standardised [MEDS](https://github.com/Medical-Event-Data-Standard/meds) (Medical Event Data Standard) format.
+
+Instead of writing one-off SQL or pandas scripts for every dataset and every study, you describe *what* to extract in declarative YAML configurations and let OpenICU handle the *how*: typed reading, joins, timestamp reconstruction, unit-aware event codes, and cross-dataset concept harmonisation. OpenICU ships with curated configurations for public ICU datasets and a growing dictionary of clinical concepts (vital signs, laboratory values, vasopressors, ventilation, and more), so the same study code can run against any supported dataset.
+
+OpenICU is the spiritual successor to dataset-harmonisation tools like [`ricu`](https://github.com/eth-mds/ricu) (R), rebuilt in Python on a modern stack: [Polars](https://pola.rs) for fast, out-of-core streaming processing, [Pydantic](https://docs.pydantic.dev) for validated configuration, and MEDS for an interoperable output format that plugs directly into the wider MEDS ecosystem of modelling and evaluation tools.
+
+---
+
+**Source code:** <https://github.com/aidh-ms/OpenICU> · **Issues:** <https://github.com/aidh-ms/OpenICU/issues>
+
+---
+
+## Why OpenICU?
+
+- **One concept, many datasets.** Define a clinical concept (e.g. *heart rate*, *norepinephrine rate*, *antibiotics*) once; map it per dataset with small YAML files. Extraction code stays identical across MIMIC-IV, eICU-CRD, NWICU, and custom sources.
+- **MEDS-native output.** All output is written as MEDS-compliant Parquet (`subject_id`, `time`, `code`, `numeric_value`, `text_value`, plus configurable extension columns) with full metadata (`dataset.json`, `codes.parquet`) — ready for MEDS-compatible downstream tooling.
+- **Declarative, versioned, reproducible.** Every table, event, and concept is a versioned YAML config with a stable identifier. The exact merged configuration used for a run is snapshotted into the output project, so results can be traced and reproduced.
+- **Versions and variants as diffs.** New dataset versions and variants (like the eICU demo) `extend` a reference version and state only their differences — no copied configs to keep in sync. The entire eICU demo configuration is two small files.
+- **Fully offline.** Designed for sensitive medical data: no network access required, nothing leaves your secure perimeter.
+- **Scales down and up.** Polars lazy streaming lets you process full MIMIC-IV on a 16–32 GB laptop, without a database server or cluster.
+- **Extensible.** Add new datasets by writing YAML (no Python required); add new transformation callbacks or complex concept logic in Python where YAML isn't enough.
+
+## How it works
+
+OpenICU organises processing as a pipeline of **steps** that operate inside a **project** directory:
+
+```mermaid
+flowchart LR
+    A[Raw source data<br>CSV/CSV.GZ] -->|ExtractionStep| B[Dataset-level MEDS events<br>e.g. mimic-iv//chartevents//220045//Heart Rate//bpm]
+    B -->|ConceptStep| C[Harmonised concepts<br>e.g. heart_rate//bpm]
+    C --> D[Your analysis /<br>MEDS ecosystem tools]
+```
+
+1. **Extraction step** — reads the raw tables of each dataset (typed CSV scan, lookup-table joins, timestamp reconstruction) and emits one MEDS event stream per configured event. Event codes preserve full provenance: `dataset//table//itemid//label//unit`.
+2. **Concept step** — maps dataset-specific event codes onto shared, dataset-agnostic clinical concepts using pattern matching (*simple* concepts), computation over other concepts (*derived* concepts), or custom Python transformers (*complex* concepts). Concept dependencies are resolved automatically.
+3. **Sharding step** *(in development)* — re-partitions harmonised data into analysis-ready shards.
+
+Each step reads its inputs and writes its outputs as a self-contained MEDS dataset inside the project, so intermediate results are inspectable and individual steps can be re-run in isolation.
+
+## Supported datasets
+
+OpenICU ships with ready-to-use extraction and concept configurations under [`config/`](config/):
+
+| Dataset | Version | Extraction configs | Concept mappings |
+| --- | --- | --- | --- |
+| [MIMIC-IV](https://physionet.org/content/mimiciv/3.1/) | 3.1, 2.2 | 19 tables (ICU + hosp) | ~80 concepts |
+| [MIMIC-IV demo](https://physionet.org/content/mimic-iv-demo/2.2/) | 2.2 | inherited from MIMIC-IV via `extends` | inherited |
+| [eICU-CRD](https://physionet.org/content/eicu-crd/2.0/) | 2.0 | 14 tables | in progress |
+| [eICU-CRD demo](https://physionet.org/content/eicu-crd-demo/2.0/) | 2.0 | inherited from eICU-CRD via `extends` | inherited |
+| [NWICU](https://physionet.org/content/nwicu/0.1.0/) | 0.1.0 | 9 tables | in progress |
+
+The shared concept dictionary in [`config/concepts/`](config/concepts/) currently covers ~90 concepts across vital signs, blood gas, clinical chemistry, hematology, medications (incl. vasopressors and antibiotics), neurological scores, respiratory parameters, fluid output, and demographics.
+
+> [!NOTE]
+> Access to the public datasets themselves requires the usual credentialing on [PhysioNet](https://physionet.org/). OpenICU works on the downloaded CSV/CSV.GZ files — no database setup needed.
+
+## Installation
+
+OpenICU requires **Python 3.13+**. Until the first PyPI release, install from GitHub:
+
+```bash
+# with pip
+pip install git+https://github.com/aidh-ms/OpenICU
+
+# with uv
+uv add git+https://github.com/aidh-ms/OpenICU
+```
+
+For development, clone the repository and use the included dev container or [`uv`](https://docs.astral.sh/uv/):
+
+```bash
+git clone https://github.com/aidh-ms/OpenICU.git
+cd OpenICU
+uv sync --all-groups
+```
+
+## Quickstart
+
+A pipeline is two small YAML files plus a few lines of Python.
+
+**1. Configure the extraction step** (`config/extraction.yml`) — point OpenICU at the bundled dataset configs and your local data:
+
+```yaml
+name: Extraction
+version: 1.0.0
+
+config_files:
+  - path: /path/to/OpenICU/config/datasets/mimic-iv/3.1/tables/
+
+config:
+  data:
+    - name: mimic-iv
+      path: /path/to/physionet.org/files/mimiciv/3.1
+```
+
+**2. Configure the concept step** (`config/concept.yml`) — select the concept dictionary and per-dataset mappings:
+
+```yaml
+name: Concept
+version: 1.0.0
+
+config_files:
+  - path: /path/to/OpenICU/config/concepts
+
+config:
+  extraction_step: Extraction
+  dataset_configs:
+    - name: mimic-iv
+      path: /path/to/OpenICU/config/datasets/mimic-iv/3.1/mappings/
+```
+
+**3. Run the pipeline:**
+
+```python
+from pathlib import Path
+
+from open_icu import OpenICUProject, ExtractionStep, ConceptStep
+
+config_path = Path.cwd() / "config"
+project_path = Path.cwd() / "output" / "project"
+
+with OpenICUProject(project_path) as project:
+    extraction_step = ExtractionStep.load(project, config_path / "extraction.yml")
+    extraction_step.run()
+
+    concept_step = ConceptStep.load(project, config_path / "concept.yml")
+    concept_step.run()
+```
+
+**4. Use the results.** The project directory now contains one MEDS dataset per step:
+
+```
+output/project/
+├── configs/        # snapshot of every config used (reproducibility)
+├── workspace/      # intermediate per-step files
+└── datasets/
+    ├── extraction/
+    │   ├── data/mimic-iv/3.1/<table>/<EVENT>.parquet
+    │   └── metadata/{dataset.json, codes.parquet}
+    └── concept/
+        ├── data/<concept>/<version>/<dataset>.parquet
+        └── metadata/{dataset.json, codes.parquet}
+```
+
+Read it with any Parquet-capable tool, e.g.:
+
+```python
+import polars as pl
+
+heart_rate = pl.scan_parquet(
+    project_path / "datasets" / "concept" / "data" / "heart_rate" / "1.0.0" / "mimic-iv.parquet"
+).collect()
+```
+
+A complete runnable example lives in [`example/pipeline.ipynb`](example/pipeline.ipynb).
+
+## Configuration at a glance
+
+OpenICU is configured at three levels — see the [documentation](#documentation) for the full reference.
+
+**Dataset/table configs** (`config/datasets/<dataset>/<version>/tables/*.yml`) describe how to turn one raw table into MEDS events: typed columns, joins, and event definitions.
+
+```yaml
+path: icu/chartevents.csv.gz
+
+columns:
+  - name: subject_id
+    type: int64
+  - name: charttime
+    type: datetime
+    params: { format: "%Y-%m-%d %H:%M:%S" }
+  - name: itemid
+    type: int64
+  - name: value
+    type: string
+  - name: valuenum
+    type: float32
+  - name: valueuom
+    type: string
+
+join:
+  - path: icu/d_items.csv.gz
+    both_on: [itemid]
+    columns: [{ name: itemid, type: int64 }, { name: label, type: string }]
+
+events:
+  - name: CHART
+    columns:
+      subject_id: col(subject_id)
+      time: col(charttime)
+      code: [col(itemid), col(label), col(valueuom)]
+      numeric_value: col(valuenum)
+      text_value: col(value)
+```
+
+`name` is a technical event identifier used by the extraction step. It is not included in the generated MEDS `code`; the code is built from the automatic dataset/table prefix, optional `code_prefix`, configured `columns.code` components, and optional `code_suffix`.
+
+**Concept configs** (`config/concepts/<category>/*.yml`) define dataset-agnostic clinical concepts:
+
+```yaml
+name: heart_rate
+version: 1.0.0
+unit: bpm
+```
+
+**Concept mappings** (`config/datasets/<dataset>/<version>/mappings/*.yml`) connect a concept to dataset-specific codes:
+
+```yaml
+type: simple
+mappings:
+  - pattern:
+      table: chartevents
+      event: CHART  # technical extraction event identifier, not part of `code`
+      code: (220045//Heart Rate)
+    columns:
+      numeric_value: col(numeric_value)
+```
+
+Here, `event` selects the technical extraction event stream; it is not matched as part of the MEDS `code`.
+
+Wherever values are computed, configs use a small **expression language** with composable callbacks — e.g. reconstructing absolute timestamps from eICU's relative offsets:
+
+```yaml
+callbacks:
+  - to_datetime(hospitaldischargeyear, 1, 1, hospitaladmittime24, hospitaladmitoffset, "minutes", output=admission_timestamp)
+post_join_callbacks:
+  - add_offset(admission_timestamp, labresultoffset, output=event_timestamp)
+```
+
+Arithmetic, comparisons, and boolean logic work as ordinary expressions (`col(weight) / (col(height) * col(height))`), and custom callbacks can be registered from Python.
+
+## Documentation
+
+- [Package overview](docs/getting_started/overview.md)
+- [Installation](docs/getting_started/installation.md)
+- [Basic usage](docs/getting_started/basic_usage.md)
+- User guide: [pipeline & projects](docs/user_guide/pipeline.md) · [extraction configs](docs/user_guide/extraction.md) · [dataset versions & variants](docs/user_guide/versioning.md) · [concept configs](docs/user_guide/concepts.md) · [expression language](docs/user_guide/expressions.md)
+- [Contributing](docs/getting_started/contributing.md)
+- Architecture documentation following [arc42](https://docs.arc42.org/home/) in [`docs/arc/`](docs/arc/)
+
+## Project status & roadmap
+
+OpenICU is in active development (pre-1.0); configuration formats may still change between minor versions. Current focus areas:
+
+- Completing concept mappings for eICU-CRD and NWICU
+- The sharding step for analysis-ready data partitioning
+- Additional datasets (e.g. MIMIC-IV demo, HiRID, AmsterdamUMCdb)
+- A first PyPI release
+
+Contributions of dataset configurations and concept definitions are especially welcome — they are pure YAML and require no changes to the framework itself.
+
+## Contributing
+
+We welcome contributions! Please use the dev container for a ready-to-go environment, follow [Conventional Commits](https://www.conventionalcommits.org/) (`feat:`, `fix:`, `docs:`, …), and make sure checks pass before opening a PR:
+
+```bash
+uv run ruff format && uv run ruff check    # format & lint
+uv run ty check .                          # type checking
+uv run pytest                              # tests
+```
+
+See the [contributing guide](docs/getting_started/contributing.md) for details, and our [Code of Conduct](.github/CODE_OF_CONDUCT.md).
+
+## Citation
+
+If you use OpenICU in your research, please cite it via the metadata in [`CITATION.cff`](CITATION.cff) (use the *"Cite this repository"* button on GitHub).
+
+## Related projects
+
+- [MEDS](https://github.com/Medical-Event-Data-Standard/meds) — the Medical Event Data Standard that OpenICU targets
+- [`ricu`](https://github.com/eth-mds/ricu) — R package for ICU data harmonisation that inspired OpenICU's concept dictionary
+- [YAIB](https://github.com/rvandewater/YAIB) — Yet Another ICU Benchmark, a complementary benchmarking framework
+
+## License
+
+OpenICU is released under the [MIT License](LICENSE).
+Developed by the [AIDH MS](https://github.com/aidh-ms) team at the University of Münster and the Medical University of Innsbruck.