Tools to work with Amsterdam Schema.
Project description
amsterdam-schema-tools
Set of libraries and tools to work with Amsterdam schema.
Install the package with: pip install amsterdam-schema-tools. This installs
the library and a command-line tool called schema, with various subcommands.
A listing can be obtained from schema --help.
Subcommands that talk to a PostgreSQL database expect either a DATABASE_URL
environment variable or a command line option --db-url with a DSN.
Many subcommands want to know where to find schema files. Most will look in a
directory of schemas denoted by the SCHEMA_URL environment variable or the
--schema-url command line option. E.g.,
schema create tables --schema-url=myschemas mydataset
will try to load the schema for mydataset from
myschemas/mydataset/dataset.json.
Generate amsterdam schema from existing database tables
The --prefix argument controls whether table prefixes are removed in the schema, because that is required for Django models.
As example we can generate a BAG schema. Point DATABASE_URL to bag_v11 database and then run :
schema show tablenames | sort | awk '/^bag_/{print}' | xargs schema introspect db bag --prefix bag_ | jq
The jq formats it nicely and it can be redirected to the correct directory in the schemas repository directly.
Express amsterdam schema information in relational tables
Amsterdam schema is expressed as jsonschema. However, to make it easier for people with a more relational mind- or toolset it is possible to express amsterdam schema as a set of relational tables. These tables are meta_dataset, meta_table and meta_field.
It is possible to convert a jsonschema into the relational table structure and vice-versa.
This command converts a dataset from an existing dataset in jsonschema format:
schema import schema <id of dataset>
To convert from relational tables back to jsonschema:
schema show schema <id of dataset>
Generating amsterdam schema from existing GeoJSON files
The following command can be used to inspect and import the GeoJSON files:
schema introspect geojson <dataset-id> *.geojson > schema.json
edit schema.json # fine-tune the table names
schema import geojson schema.json <table1> file1.geojson
schema import geojson schema.json <table2> file2.geojson
Importing GOB events
The schematools library has a module that reads GOB events into database tables that are defines by an Amsterdam schema. This module can be used to read GOB events from a Kafka stream. It is also possible to read GOB events from a batch file with line-separeted events using:
schema import events <path-to-dataset> <path-to-file-with-events>
Schema Tools as a pre-commit hook
Included in the project is a pre-commit hook
that can validate schema files
in a project such as amsterdam-schema
To configure it
extend the .pre-commit-config.yaml
in the project with the schema file defintions as follows:
- repo: https://github.com/Amsterdam/schema-tools
rev: v3.5.0
hooks:
- id: validate-schema
args: ['https://schemas.data.amsterdam.nl/schema@v1.2.0#']
exclude: |
(?x)^(
schema.+| # exclude meta schemas
datasets/index.json
)$
args is a one element list
containing the URL to the Amsterdam Meta Schema.
validate-schema will only process json files.
However not all json files are Amsterdam schema files.
To exclude files or directories use exclude with pattern.
pre-commit depends on properly tagged revisions of its hooks.
Hence, we should not only bump version numbers on updates to this package,
but also commit a tag with the version number; see below.
Doing a release
(This is for schema-tools developers.)
We use GitHub pull requests. If your PR should produce a new release of
schema-tools, make sure one of the commit increments the version number in
setup.cfg appropriately. Then,
- merge the commit in GitHub, after review;
- pull the code from GitHub and merge it into the master branch,
git checkout master && git fetch origin && git merge --ff-only origin/master; - tag the release X.Y.Z with
git tag -a vX.Y.Z -m "Bump to vX.Y.Z"; - push the tag to GitHub with
git push origin --tags; - release to PyPI:
make upload(requires the PyPI secret).
Mocking data
The schematools library contains two Django management commands to generate
mock data. The first one is create_mock_data which generates mock data for all
the datasets that are found at the configured schema location SCHEMA_URL
(where SCHEMA_URL can be configure to point to a path at the local filesystem).
The create_mock_data command processes all datasets. However, it is possible
to limit this by adding positional arguments. These positional arguments can be
dataset ids or paths to the location of the dataset.json on the local filesystem.
Furthermore, the command has some options, e.g. to change
the default number of generated records (--size) or to reverse meaning of the positional
arguments using --exclude.
To avoid duplicate primary keys on subsequent runs the --start-at options can be used
to start autonumbering of primary keys at an offset.
E.g. to generate 5 records for the bag and gebieden datasets, starting the
autonumbering of primary keys at 50.
django create_mock_data bag gebieden --size 5 --start-at 50
To generate records for all datasets, except for the fietspaaltjes dataset:
django create_mock_data fietspaaltjes --exclude # or -x
To generate records for the bbga dataset, by loading the schema from the local filesystem:
django create_mock_data <path-to-bbga-schema>/datasets.json
During record generation in create_mock_data, the relations are not added,
so foreign key fields will be filled with NULL values.
There is a second management command relate_mock_data that can be used to
add the relations. This command support positional arguments for datasets
in the same way as create_mock_data.
Furthermore, the command also has the --exclude option to reverse the meaning
of the positional dataset arguments.
E.g. to add relations to all datasets:
django relate_mock_data
To add relations for bag and gebieden only:
django relate_mock_data bag gebieden
To add relations for all datasets except meetbouten:
django relate_mock_data meetbouten --exclude # or -x
NB. When only a subset of the datasets is being mocked, the command can fail when datasets that are involved in a relation are missing, so make sure to include all relevant datasets.
For convenience an additional management command truncate_tables has been added,
to truncate all tables.
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
File details
Details for the file amsterdam-schema-tools-5.6.2.tar.gz.
File metadata
- Download URL: amsterdam-schema-tools-5.6.2.tar.gz
- Upload date:
- Size: 126.3 kB
- Tags: Source
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/3.6.0 importlib_metadata/4.8.2 pkginfo/1.7.1 requests/2.26.0 requests-toolbelt/0.9.1 tqdm/4.62.3 CPython/3.9.16
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 | f2e16ef8ce0b00829ddbf3f94512ff0a8a46f795c0c9dd0ce8cbb7ff0c9fe9b0 |
|
| MD5 | 3c8f8b68b22a2caf83872a98a7afd0de |
|
| BLAKE2b-256 | 3e10c0033685267d438e9419817f63d5f7206b19606bfdbb81eafa0be65965ea |
File details
Details for the file amsterdam_schema_tools-5.6.2-py3-none-any.whl.
File metadata
- Download URL: amsterdam_schema_tools-5.6.2-py3-none-any.whl
- Upload date:
- Size: 158.7 kB
- Tags: Python 3
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/3.6.0 importlib_metadata/4.8.2 pkginfo/1.7.1 requests/2.26.0 requests-toolbelt/0.9.1 tqdm/4.62.3 CPython/3.9.16
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 | 8c3830baa77103e4274b6bfd53ed098944187771339ddfa9036c72cb9b2345cc |
|
| MD5 | 7783ba1e9db1b0526212c70df96af751 |
|
| BLAKE2b-256 | bf1861eeea8b1ee96db788761020636bf2cce3953c48476f28cd8a02aa316e35 |
