This utility updates metadata on koordiantes platforms.
For installation instructions see INSTALL.md
- Downloads the xml metadata from the Data Service, for one or many user defined data sources
- Takes a back-up of the metadata file (if text replacement is to be made)
- Finds and replaces text within the metadata file
- Updates the metadata with the newly edited XML.
A config.yaml file must be provided. This can be created by editing the provided config_tempate.yaml file.
Connection:
Api_key = <ADMIN API KEY> # See notes below on API Keys
Domain = <Data Service Domain> # e.g. data.linz.govt.nz
Text:
Mapping:
1: # The order the text replacements are to occur
# DONT NOT DUPLICATE NUMBERS !!!
# ENUSRE NUMBERING IS SEQUENTIAL STARTING AT 1 !!!
search: the terrace # The text to search for replacement
# !!! FORMAT: Python Regular Expression
target_element: './/gmd:contact/gmd:CI_ResponsibleParty/gmd:individualName/gco:CharacterString' #
# The element you wish to target. if Null and search and replace
# is performed across then entire file.
replace: The Road # The text that is to replace the search text
ignore_case: True # True or False. If True, search text case is ignored
2:
search: Land Info New Zealand
replace: Land Information New Zealand
ignore_case: True
Output:
Destination: <Directory> # The directory where to write
# metadata file backups
Datasets:
Layers: <Layers to Process> # A list of Layers/Table ids or "All"
# All will process All Tables and Layers
# e.g. [93639,93648, 93649] or "All"
Sets: # Sets are not currently supported
Docs: # Documents are not currently supported
Test:
Dry_run: False # True or False
# If True, metadata xml documents are
# edited and stored but no changes to the
# Data Service made
Overwrite_files: True # True or Flase
# Useful for dev and testing. When True
# metadata files that are already in
# the destination dir will be over-written
Text Mapping
The order that text is searched and replaced is important. For this reason the mappings must be formatted as above. That is, each mapping must be sequentially number, starting at 1, in the order the search and replace is to be executed.
The script uses re.sub. The search text format must therefore be in the Python Regular Expression format. Replace text is a standard plain text string that is to replace the regular expression match.
If target_element the edits for the mapping are only made against the referenced XML element.
API Key
The (LINZ) Data Service API key must be generated with the required permissions to update metadata. It is recommend that a API key is created specifically for this task.
The API KEY must have the following permission enabled against it. You will need admin rights to be able to enable all of the below
- Query layer data
- Search and view tables and layers
- Create, edit and delete tables and layers
- View the data catalog and access catalog services (CS-W)
For LDS users, your API key can be managed here
There are two options for storing your API Key where the script can utilise the key
for authentication. The API key can be entered in the config.yml or stored as
an environmental variable. Storing the API Key as an environmental variable is
the safest and therefore recommended way to do this. The environment variable
the key is to be assigned to must be LDS_APIKEY=<lds_apikey>
Once the config.yaml file has been updated simply run
metadata_updater (if installed via the recommended setup.py method)
For all edited metadata the script will output a backup and edited version of the metadata xml file to the destination directory (see configuration notes). These files are output for the purposes of record keeping, understanding the changes made and debugging purposes.
Each output file is named <Data Type>_<Data Id>_<data Ttile>.iso.xml
with the backup being appended with "._bak"
For Example: Layer 4567 would be:
layer_50772_nz-primary-parcels.iso.xmllayer_50772_nz-primary-parcels.iso.xml._bak
Important; a log will be output to the metadata_updater.log file.
If when the script is finished it reports a number of errors
For example:
'Process failed with 5 error(s). Please see log for critical messages'
The log must be grepped/searched for critical errors. Each critical error indicates a layer / table failed to be updated. The log will indicate which layers / tables and why a failure occurred.
Also of importance: The current scope of this script is to only handle tables and layers. Therefore when documents and sets are encountered they are skipped over but their ids are logged out. If you are running the script you may want to grep / search for these in the log and edit the metadata manually.
These are of log "WARNING" level and formatted as per the below example:
2018-02-26 12:12:45,931 [WARNING]: Dataset 1234: Data is of
"Document" type. This process only handles tables/layers'
The script uses the Koordinates Python API client
While the Python API client documentation is clear on method use it does not provide well defined user work flows. The Editing of the data is based on the work flow of; Create a Draft >> Edit the Draft >> Import the draft >> and then publish the layers together as a pulish group.
This work flow of creating a draft through to publishing is easy enough to follow in the source code of this project.
Unit tests are provided to test all methods not making requests via the Koordinates Python client
Integration tests are provided to test all methods making requests via the Koordinates Python client
Both these test can be ran by executing ~/lds-metadata/updater/tests/test.py
Integration test require that the LDS_APIKEY envi var is set.
The tests are run with every pull request and also on push to master
This is so far an initial minimum viable product release.
There are a number of proposed enhancements. This includes utilising an XML Tree to be able target tags and only update text in specific fields
See the list of enhancements as store against the GitHub issues of this project
Please supply any feedback and bug reports to the projects GitHub Issues page