epifanio / adc-pycsw Goto Github PK

The docker set-up uses the entrypoint.py to start pycsw
which has pointers to a repository table:

https://github.com/geopython/pycsw/blob/master/docker/entrypoint.py#L88
which returns:

DEBUG:__main__:Setting up pycsw's data repository...
DEBUG:__main__:Repository URL: None
Traceback (most recent call last):
  File "/usr/local/lib/python3.8/configparser.py", line 789, in get
    value = d[option]
  File "/usr/local/lib/python3.8/collections/__init__.py", line 898, in __getitem__
    return self.__missing__(key)            # support subclasses that define __missing__
  File "/usr/local/lib/python3.8/collections/__init__.py", line 890, in __missing__
    raise KeyError(key)
KeyError: 'table'

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "/usr/local/bin/entrypoint.py", line 188, in <module>
    launch_pycsw(config, workers=workers, reload=args.reload)
  File "/usr/local/bin/entrypoint.py", line 88, in launch_pycsw
    pycsw_config.get("repository", "table"),
  File "/usr/local/lib/python3.8/configparser.py", line 792, in get
    raise NoOptionError(option, section)
configparser.NoOptionError: No option 'table' in section: 'repository'

Trying to add a table entry in the pycsw.cfg for debugging, I then got:

912ax DEBUG:__main__:Setting up pycsw's data repository...
912ax DEBUG:__main__:Repository URL: None
912ax Traceback (most recent call last):
912ax   File "/usr/local/bin/entrypoint.py", line 188, in <module>
912ax     launch_pycsw(config, workers=workers, reload=args.reload)
912ax   File "/usr/local/bin/entrypoint.py", line 86, in launch_pycsw
912ax     db_handler(
912ax TypeError: 'NoneType' object is not callable
8k9fd DEBUG:__main__:Setting up pycsw's data repository...
8k9fd DEBUG:__main__:Repository URL: None
8k9fd Traceback (most recent call last):
8k9fd   File "/usr/local/bin/entrypoint.py", line 188, in <module>
8k9fd     launch_pycsw(config, workers=workers, reload=args.reload)
8k9fd   File "/usr/local/bin/entrypoint.py", line 86, in launch_pycsw
8k9fd     db_handler(
8k9fd TypeError: 'NoneType' object is not callable

I guess to fix this we have to modify the entrypoint.py , should suffice to just avoid to execute the db_handler() if the repository.database is set to None in the config?

Filtering of data from the same solr core should be supported, based on the mmd collection keyword.
The configuration file should declare the list of collection keyword which are used as filter to expose data in different pycsw endpoint.

For both anytext and BBOX we have dedicated methods in: https://github.com/epifanio/adc-pycsw/blob/main/plugins/repository/solr_helper.py we need to isolate the time query as well - at the moment it is repeated in the main plug-in code

We need to add support for multiple query like:

• multiple anytext query for different fields (priority)
• multiple BBOX (mid priority)
• multiple Time range ( low priority / optional)

For how our plug-in has been implemented, when a query is not "understood", the code will ignore the whole filter, thus returning all the records. The same was happening for the queries we just implemented, which are now fixed. This is misleading and we require a way to signal this behavior (not implemented) to the user.

Following the feedback received by querying our catalogue - will document here the Query that fails, their CSW syntax and mapping to SOLR

At the moment the plugin files for the repository are stored in:

pycsw/plugins/repository/solr_metno.py

will be good to have those files in a dedicated directory SOLR

so that we will change the configuration from:

source=pycsw.plugins.repository.solr_metno.SOLRMETNORepository

to:

source=pycsw.plugins.repository.SOLR.solr_metno.SOLRMETNORepository

We export the path used for the XSLT conversion from MMD to ISO:

export MMD_TO_ISO="path/to/mmd-ti-iso.xsl"

We should export just the PATH to an XSLT directory instead,

MMD_XSLT_DIR="/path/toMMD_XSLT_DIR"

and reuse such env in the code for the several output profiles

The following line will need to be set to series when doc['isParent] = True

And we will need to add:
record['parentidentifier'] = doc['related_dataset'] when doc[isChild = True]

I think this fixes will add parent/child relations for the default outputschema results at least.