- Python 100%
| Filename | Latest commit message | Latest commit date |
|---|---|---|
| docs | ||
| src | ||
| .gitignore | ||
| COPYING | ||
| COPYING.LESSER | ||
| MANIFEST.in | ||
| pyproject.toml | ||
| README.md | ||
splent_feature_local_publication
The local_publication feature of a SPLENT product line.
A SPLENT feature is a Python package that a product composes with others:
its models, routes, services, templates and hooks are registered when the
product starts. What this feature provides and requires is listed in the
[tool.splent.contract] section of pyproject.toml, which
splent feature:contract --write keeps in sync with the source.
This one publishes a hub's datasets on the hub itself, for a hub without
Zenodo (F5-6 of the OpenHubs plan). It is a publication provider of the
dataset feature (dataset's DepositionProvider, 0.10.0 and later), next
to zenodo, and owns no table and no route: what it publishes are
dataset's rows, at dataset's addresses.
What a publication gives a dataset
- An identifier of the hub's own,
pid_scheme = "local": the dataset's public id (CO-5), the ULID it has had since it was made (26 characters of Crockford's base 32, sorted by the time it was made, that name neither the dataset nor its owner). It is no DOI and never looks like one: a DOI with a prefix nobody registered would be a false one. - Its landing at dataset's
/records/<public_id>/, which answers what/doi/<doi>/answers: the page, the FAIR metadata by content negotiation (JSON-LD, Turtle, DataCite), and the tombstone of a withdrawn dataset. The metadata names the landing as the identifier's address, never doi.org, and the pages call it an ID, not a DOI. - Its files where hubfile keeps them: nothing is deposited anywhere. Their
pages and raw addresses are hubfile's, under the landing
(
/records/<public_id>/files/<name>and.../files/raw/<name>, hubfile 0.10.0 and later), and/hubfiles/raw/<id>/<name>.
0.1.0 minted an identifier of its own and answered /records/<pid>/
itself, a rule dataset's landing declares too, which the contract check of
product:validate counted as a route collision. Since 0.2.0 the identifier
is the public id and the landing is dataset's alone. A dataset 0.1.0 published
keeps that identifier, and its landing went with the route; only the
development data of filehub_splent, which no release ever ran, has any.
API v1's publish answers it with doi null and the identifier in pid,
pid_scheme and landing_url. A metadata change needs no sync: the
landing reads the dataset as it is.
Use it in a product
splent feature:add splent_feature_local_publication
A product that installs zenodo as well publishes each dataset through one of
them: the one the publication asks for, else dataset's
DATASET_PUBLICATION_PROVIDER, else the only one installed. A dataset keeps
the provider it was published through, its new versions included.
The OpenHubs model gets the node with the change in
docs/openhubs_spl.uvl.patch: publicacion becomes a mandatory or-group of
zenodo and local_publication, with local_publication => dataset.
Develop
splent feature:test splent_feature_local_publication
splent feature:contract splent_feature_local_publication
uvlhub does not declare it, so its tests run in a neutral app, with the
newest releases of dataset and hubfile; filehub_splent declares it, so
SPLENT_APP=filehub_splent splent feature:test splent_feature_local_publication
runs them in the file hub. Its tests publish a CSV, a file of hubfile's
generic type.
Release
splent feature:release splent_feature_local_publication --dry-run --bump patch
splent feature:release splent_feature_local_publication
The first command rehearses the release and changes nothing. The second publishes it to the git forge and to PyPI, and a PyPI version is permanent.