Dereferencer #11
Dereferencer #11
Conversation
nicer container-names
make it easy to get the jupyer token
…ask/basic-ttl-ingest
but moving the log elsewhere
reverses symlink direction to make docker happy (and the other usage satisfied) applies the poetry build inside docker inage build to fix #1
further enhances the request from #3 some path changes allong the way (conform the reeversed symlinks)
this marks an important milestone for #4 as we can now insert triples!
…-2023 into task/basic-ttl-ingest
completes the goal of "test driving the ingest of rdf into graphdb" fixes #4
Know issue with inserting wktstrings, look into postgres for solving this
TODO: discuss with marc what the correct approach is for this
- added re testing - deleted make_tasks tests because function is no longer in use - added edmo as test for the script tag search - added pytest.ini so when doing pytests I can see what went wrong
| ?s ?p schema:Dataset . | ||
| } | ||
| assert-paths: | ||
| cache_lifetime: 1 |
There was a problem hiding this comment.
please remind me what this does?
| cache_lifetime: 18000 | ||
| ``` | ||
|
|
||
| In this example, the subjects are the literal values of the URIs. The `assert-paths` are the property paths to follow in the results. The `cache_lifetime` is the lifetime of the cache in minutes. |
There was a problem hiding this comment.
still does not explain what is being cached, and thus does not help me to consider what should drive my decision as a user to set, increase, decrease this thingy?
There was a problem hiding this comment.
if this is a _test file should we not put it somewhere else?
at least grouped under ./data/tests for now but maybe even ./tests/data (although that might come natural if we repackage this as k-gap project without other data)
| cc:prohibits cc:CCCommercialUse . | ||
|
|
||
| <https://data.arms-mbon.org/> a schema:Project . |
There was a problem hiding this comment.
looks like this too is being added for testing purposes?
if so please consider moving to separate file and under ./data/tests
| SPARQL: > | ||
| select DISTINCT ?s where { | ||
| ?s ?p ?o . | ||
| FILTER regex(str(?s), "^http://marineregions.org/mrgid/[0-9]{1,5}$") |
There was a problem hiding this comment.
violates the adagium "sparql regex on URI considered bad"
please rephrase in terms of selecting subjects of a certain type maybe?
| :return: The results of the query | ||
| :rtype: list | ||
| """ | ||
| template = "deref_property_trajectory.sparql" |
There was a problem hiding this comment.
select_property_path.sparql
as it should not be tied to us using this in a context of deciding to maybe deref and fetch
| "subject": uri, | ||
| "property_trajectory": pp_trajectory, | ||
| "prefixes": prefixes} | ||
| query = J2RDF.build_syntax(template, **vars) |
There was a problem hiding this comment.
you are not using template var anywhere else - pls consider inlining for readability?
alternative - make decent local constants for the provided templates you use in this file?
There was a problem hiding this comment.
drop deref in this name, there is nothing in this template that ties its use to the conicidence of using it here in the deref component context
or put differently why not add deref_ to all folders and files in this space ¯_(ツ)_/¯
There was a problem hiding this comment.
also -- is this still in use / useful?
seems to be a precursor / double of the more complete next one in the list?
|
|
||
| SELECT ?o | ||
| WHERE { | ||
| <{{ subject }}> <{{ property }}> ?o . |
There was a problem hiding this comment.
rename property to path or property_path ?
There was a problem hiding this comment.
another deref_ prefix to drop
There was a problem hiding this comment.
and kill that _trajectory thing as well - not meaningful, only confusing
There was a problem hiding this comment.
now spotted the subtle < > difference --> does that make sense / enough motivation to keep it around?
|
|
||
| SELECT ?o | ||
| WHERE { | ||
| <{{ subject }}> {{ property_trajectory }} ?o . |
There was a problem hiding this comment.
property_trajectory should be property_path
There was a problem hiding this comment.
now spotted the subtle < > difference --> does that make sense / enough motivation to keep it around?
| WHERE { | ||
| ?s rdf:type schema:Person . | ||
| } | ||
| assert-pathsy: |
There was a problem hiding this comment.
typo? assert-pathsy has a trailing y
There was a problem hiding this comment.
strange - what is the use?
if temporary disabling of a deref-config is the goal then it now requires renaming anyway -- so just make the arrangement to add a .bak suffix of sorts rather then alter the dereference* prefix ?
There was a problem hiding this comment.
this and all following files in this directory
does not make sense that the PR branch for lwua-deref is making all these changes to deref-ingest ...
remove / move to different PR if needed ?
There was a problem hiding this comment.
generally hate foldernames with suffix-s
./config/ will do nicely
motivation: there is only one folder with that name (so it is singular :) -- , and all folders are meant to possibly contain multiple files ... so what gives?
There was a problem hiding this comment.
why is this general stuff in this specific PR?
There was a problem hiding this comment.
why is jupyter folder included in this PR for deref?
added some showcases?
Module dereferencer added that will perform tasks given a config file.
Set config file looks like this: