Metadata-Version: 2.5
Name: geomend
Version: 0.0.8
Summary: Combine GEO metadata tables, review suggested variable mappings, and export data with source and review history
Project-URL: Documentation, https://prestonraab.github.io/geomend/
Project-URL: Repository, https://github.com/prestonraab/geomend
Project-URL: Issues, https://github.com/prestonraab/geomend/issues
License: MIT License
        
        Copyright (c) 2026 Preston Raab
        
        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.
License-File: LICENSE
Keywords: GEO,bioinformatics,data curation,metadata harmonization,ontology mapping,schema matching
Classifier: Development Status :: 3 - Alpha
Classifier: Intended Audience :: Science/Research
Classifier: License :: OSI Approved :: MIT License
Classifier: Operating System :: OS Independent
Classifier: Programming Language :: Python :: 3 :: Only
Classifier: Programming Language :: Python :: 3.13
Classifier: Topic :: Scientific/Engineering :: Bio-Informatics
Requires-Python: >=3.13
Requires-Dist: bagit<2,>=1.9
Requires-Dist: biopython<2,>=1.85
Requires-Dist: bioregistry<1,>=0.14
Requires-Dist: diskcache<6,>=5.6
Requires-Dist: duckdb<2,>=1.5
Requires-Dist: fastapi<1,>=0.115
Requires-Dist: filelock<4,>=3.16
Requires-Dist: frictionless<6,>=5.18
Requires-Dist: geoparse<3,>=2.0
Requires-Dist: httpx<1,>=0.28
Requires-Dist: huey<3,>=2.5
Requires-Dist: joblib<2,>=1.4
Requires-Dist: linkml-map==0.5.4
Requires-Dist: numpy<3,>=2.1
Requires-Dist: oaklib<1,>=0.6
Requires-Dist: packaging<27,>=24
Requires-Dist: pandas<3,>=2.2
Requires-Dist: pint<1,>=0.24
Requires-Dist: pooch<2,>=1.9
Requires-Dist: pydantic<3,>=2.12
Requires-Dist: pyyaml<7,>=6.0
Requires-Dist: rapidfuzz<4,>=3.10
Requires-Dist: scikit-learn<2,>=1.5
Requires-Dist: scipy<2,>=1.14
Requires-Dist: sse-starlette<4,>=3.5
Requires-Dist: sssom<1,>=0.4
Requires-Dist: uvicorn<1,>=0.34
Provides-Extra: cohort
Requires-Dist: diskcache<6,>=5.6; extra == 'cohort'
Requires-Dist: duckdb<2,>=1.5; extra == 'cohort'
Requires-Dist: fastapi<1,>=0.115; extra == 'cohort'
Requires-Dist: frictionless<6,>=5.18; extra == 'cohort'
Requires-Dist: httpx<1,>=0.28; extra == 'cohort'
Requires-Dist: huey<3,>=2.5; extra == 'cohort'
Requires-Dist: joblib<2,>=1.4; extra == 'cohort'
Requires-Dist: linkml-map==0.5.4; extra == 'cohort'
Requires-Dist: numpy<3,>=2.1; extra == 'cohort'
Requires-Dist: packaging<27,>=24; extra == 'cohort'
Requires-Dist: pint<1,>=0.24; extra == 'cohort'
Requires-Dist: pooch<2,>=1.9; extra == 'cohort'
Requires-Dist: rapidfuzz<4,>=3.10; extra == 'cohort'
Requires-Dist: scikit-learn<2,>=1.5; extra == 'cohort'
Requires-Dist: scipy<2,>=1.14; extra == 'cohort'
Requires-Dist: sse-starlette<4,>=3.5; extra == 'cohort'
Requires-Dist: uvicorn<1,>=0.34; extra == 'cohort'
Provides-Extra: geo-verify
Requires-Dist: geoparse<3,>=2.0; extra == 'geo-verify'
Requires-Dist: pandas<3,>=2.2; extra == 'geo-verify'
Provides-Extra: learned
Requires-Dist: huggingface-hub<1,>=0.28; extra == 'learned'
Requires-Dist: torch<3,>=2.6; extra == 'learned'
Requires-Dist: transformers<5,>=4.48; extra == 'learned'
Provides-Extra: ontology
Requires-Dist: bioregistry<1,>=0.14; extra == 'ontology'
Requires-Dist: oaklib<1,>=0.6; extra == 'ontology'
Requires-Dist: pooch<2,>=1.9; extra == 'ontology'
Requires-Dist: sssom<1,>=0.4; extra == 'ontology'
Provides-Extra: reconcile
Requires-Dist: biopython<2,>=1.85; extra == 'reconcile'
Requires-Dist: frictionless<6,>=5.18; extra == 'reconcile'
Requires-Dist: httpx<1,>=0.28; extra == 'reconcile'
Requires-Dist: joblib<2,>=1.4; extra == 'reconcile'
Requires-Dist: numpy<3,>=2.1; extra == 'reconcile'
Requires-Dist: packaging<27,>=24; extra == 'reconcile'
Requires-Dist: pint<1,>=0.24; extra == 'reconcile'
Requires-Dist: pooch<2,>=1.9; extra == 'reconcile'
Requires-Dist: rapidfuzz<4,>=3.10; extra == 'reconcile'
Requires-Dist: scikit-learn<2,>=1.5; extra == 'reconcile'
Requires-Dist: scipy<2,>=1.14; extra == 'reconcile'
Provides-Extra: server
Requires-Dist: fastapi<1,>=0.115; extra == 'server'
Requires-Dist: sse-starlette<4,>=3.5; extra == 'server'
Requires-Dist: uvicorn<1,>=0.34; extra == 'server'
Provides-Extra: transform
Requires-Dist: linkml-map==0.5.4; extra == 'transform'
Description-Content-Type: text/markdown

# GEOmend

GEOmend combines metadata from GEO studies and local tables. Start with one
project: **add datasets → inspect suggestions → refine or review → export**.
Source artifacts, fixed review evidence, decisions, and output history are retained.

## Start

Python 3.13+ is required. The default install includes standard workflow dependencies.

```bash
python -m pip install geomend
geomend project create samples-a.tsv samples-b.tsv --project my-project
geomend project summary my-project
geomend project tasks my-project --json
geomend project exports plan my-project --mode best_guess --json
geomend project exports table my-project --mode best_guess --output metadata.tsv --profile analysis
```

Inputs can be GEO accessions, family-SOFT files, or CSV/TSV metadata. Initial
suggestions and `best_guess` output are provisional. Choose `--mode reviewed`
explicitly for a cohort table restricted to reviewed mappings. Review and
transformation completeness do not prove source meaning.

## Python and browser clients

```python
from geomend.project import ProjectClient

project = ProjectClient.open("my-project")
summary = project.summary()
work = project.tasks(limit=20)
fields = project.mappings.fields(limit=20)
plan = project.exports.plan(mode="best_guess")
project.exports.table("metadata.csv", mode="best_guess", expected_plan_sha256=plan.plan_sha256)
```

The starting interface has `create`, `open`, `summary`, `tasks`, and `capabilities`.
Expand `inputs`, `mappings`, `concepts`, `transformations`, `review`, `exports`, or
`administration` when needed. Discovery and navigation are read-only; they never
accept suggestions, refresh reviews, or materialize output.

Use `geomend project administration serve my-project` for the authenticated
loopback service under `/api/v1`. Download its browser client, which exposes the
same starting and domain operations. GUI clients own presentation and drafts;
GEOmend owns scientific decisions and their validation.

## Review with others

```bash
geomend project review assign my-project --output-dir round-one --reviewer a --reviewer b --json
```

Reviewers use `geomend.reviewer.ReviewerClient` or the portable review CLI to
inspect fixed evidence, preview judgments, and save responses. Import responses
with `project review import`, compare with `project review compare`, and issue
synthesis with `project review synthesize`. Round IDs identify the original
issuer. Review hierarchy organizes work without weighting reviewers by title.

See the [two project journeys](docs/project.md), [reviewer protocol](docs/reviewer-protocol.md),
and [advanced reference](docs/project-reference.md). `geomend project --help` shows
domains; `geomend project --expanded-help` shows every option.

## Advanced workflows

Standalone reconciliation supports curated registries and immutable reviewed
workspaces through `geomend reconcile plan` followed by `geomend reconcile run`.
Native data and audit exports retain distinct contracts. Ontology, transformation,
identity, snapshot, fork, and storage operations remain available through focused
interfaces. Never edit review SQLite or response records directly.

Learned retrieval adds dependencies through `geomend[learned]`. Local Ollama is
optional and never downloaded automatically. Training and scientific benchmarks
live in `geomend-research`; see the [development boundary](docs/development-boundary.md).

For checkout development, use the existing environment from the repository root:

```bash
PYTHONPATH="$PWD/src" .venv/bin/geomend capabilities --json
```

Source artifacts remain immutable. Suggestions, reviewer preferences, effective
mappings, and export readiness are distinct states. Scores rank evidence; they
are not probabilities of correctness. GEOmend does not infer scientific
equivalence from matching strings or deduplicate samples across studies automatically.
