[Wikidata-bugs] [Maniphest] [Updated] T116298: SPARQL endpoint should gracefully handle cycles and loops in transitive properties

2016-03-10 Thread Smalyshev
Smalyshev removed a project: Discovery-Wikidata-Query-Service-Sprint.

TASK DETAIL
  https://phabricator.wikimedia.org/T116298

EMAIL PREFERENCES
  https://phabricator.wikimedia.org/settings/panel/emailpreferences/

To: Smalyshev
Cc: Jheald, Aklapper, daniel, debt, Gehel, D3r1ck01, FloNight, Izno, jkroll, 
Smalyshev, Wikidata-bugs, Jdouglas, aude, Deskana, Manybubbles, Mbch331



___
Wikidata-bugs mailing list
Wikidata-bugs@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wikidata-bugs


[Wikidata-bugs] [Maniphest] [Updated] T116298: SPARQL endpoint should gracefully handle cycles and loops in transitive properties

2015-10-26 Thread Smalyshev
Smalyshev added a project: Discovery-Wikidata-Query-Service-Sprint.

TASK DETAIL
  https://phabricator.wikimedia.org/T116298

EMAIL PREFERENCES
  https://phabricator.wikimedia.org/settings/panel/emailpreferences/

To: Smalyshev
Cc: Jheald, Aklapper, daniel, jkroll, Smalyshev, Wikidata-bugs, Jdouglas, aude, 
Deskana, Manybubbles



___
Wikidata-bugs mailing list
Wikidata-bugs@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wikidata-bugs


[Wikidata-bugs] [Maniphest] [Updated] T116298: SPARQL endpoint should gracefully handle cycles and loops in transitive properties

2015-10-25 Thread Jheald
Jheald added a subscriber: Jheald.
Jheald added a comment.

Transitive properties aren't necessary a-cyclic  -- for example, consider 
https://phabricator.wikimedia.org/P460 "said to be the same as", which is 
transitive, currently used to link given names together, if they are considered 
to be variants of the same name in different languages.

The page at

  https://www.wikidata.org/wiki/Wikidata:WikiProject_Names/given-name_variants

includes two columns of queries that find all the variants of each name (ie 
extract the whole equivalence class), using a 
https://phabricator.wikimedia.org/P460* path query from a given starting name, 
and count up the number of occurrences of each one.

https://phabricator.wikimedia.org/P460 is transitive, but it is not directed, 
and so there are huge numbers of loops in the equivalence classes. But this 
isn't a problem.  The path search just keeps going until no new 
https://phabricator.wikimedia.org/P460 link leads it to a node it has not 
already seen.  So the engine //can// handles path loops, without a problem.

Similarly, looking at the query Daniel posted,

  SELECT (COUNT(DISTINCT(?city)) AS ?count) WHERE {
?city wdt:P31/wdt:P279* wd:Q515 .# find instances of subclasses of city
?city wdt:P131* wd:Q1202 .
  }

if one comments out either one or the other of the two requirement lines, the 
query runs without a hitch.  So I don't think there is anything to do with the 
nature of the data (eg whether it contains cycles or not) that affects whether 
or not Blazegraph can handle it.

The problem appears to arise with how Blazegraph combines the two requirements.

Related to this may be the fact that Blazegraph also fails with path searches 
of the form

  SELECT (COUNT(DISTINCT ?a) AS ?count) WHERE { 
  BIND (wd:Q3305213 AS ?class) .# paintings 
  ?a wdt:P31/wdt:P279* ?class .
  }

even though the same search works perfectly if the variable ?class is 
eliminated, and wd:Q3305213 instead specified inline.

I filed the latter with Blazegraph as issue 1543 three weeks ago, but no 
commentary on it from them as yet

  https://jira.blazegraph.com/browse/BLZG-1543


TASK DETAIL
  https://phabricator.wikimedia.org/T116298

EMAIL PREFERENCES
  https://phabricator.wikimedia.org/settings/panel/emailpreferences/

To: Smalyshev, Jheald
Cc: Jheald, Aklapper, daniel, jkroll, Smalyshev, Wikidata-bugs, Jdouglas, aude, 
Deskana, Manybubbles



___
Wikidata-bugs mailing list
Wikidata-bugs@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wikidata-bugs