opencaesar / owl-tools

A set of analysis tools for OWL
Apache License 2.0
5 stars 1 forks source link

OwlQueryTask needs to describe where a failure has occurred #43

Open mprather opened 1 year ago

mprather commented 1 year ago

Description

If the OwlQueryTask is used in folder mode and one of the queries contains an error, the information displayed to the user is not helpful. The task does not tell you what file had the problem.

If the --info flag is used, the additional info does not help.

Steps to Reproduce

Configure a task to use OwlQueryTask in "folder" mode.

task runAllQueries(type:io.opencaesar.owl.query.OwlQueryTask, dependsOn: owlLoadTestModels) {
    endpointURL = "http://localhost:3030/$serviceName"
    queryPath   = file('src/sparql')
    resultPath  = file('build/queries')
    format      = 'csv'
}

If a query fails, the system tends to give a really obtuse error message such as

=================================================================
                        S T A R T
                     OWL Query 1.1.4
=================================================================
Endpoint URL: http://localhost:3030/monsid-demo
Query path: C:\src\hrs\src\sparql
Result location: C:\src\hrs\build\queries
Format Type: csv
:runAllQueries (Thread[Daemon worker Thread 5,5,main]) completed. Took 1.554 secs.

FAILURE: Build failed with an exception.

* What went wrong:
Execution failed for task ':runAllQueries'.
> org.apache.jena.riot.thrift.RiotThriftException: org.apache.thrift.protocol.TProtocolException: don't know what type: 13

Immediate debug challenges are present: 1) what file? 2) what query or portion of the query?

If the '--info' flag is used when running gradle, the user is presented with something like...

Skipping task ':owlLoad' as it is up-to-date.
:owlLoad (Thread[Execution worker for ':',5,main]) completed. Took 0.024 secs.
:runAllQueries (Thread[Daemon worker Thread 5,5,main]) started.
Query Type: Select
Query Type: Select
Query Type: Select
Query Type: Select
Query Type: Select
Query Type: Select
Query Type: Select
Query Type: Select
Query Type: Select
Query Type: Select
Result saved at: C:\src\hrs\build\queries\a.csv
Query Type: Select
Result saved at: C:\src\hrs\build\queries\f.csv
Result saved at: C:\src\hrs\build\queries\s.csv
Result saved at: C:\src\hrs\build\queries\j.csv
Result saved at: C:\src\hrs\build\queries\p.csv
Result saved at: C:\src\hrs\build\queries\t.csv
Result saved at: C:\src\hrs\build\queries\b.csv
Result saved at: C:\src\hrs\build\queries\c.csv
Result saved at: C:\src\hrs\build\queries\m.csv
Result saved at: C:\src\hrs\build\queries\e.csv

> Task :runAllQueries FAILED
Caching disabled for task ':owlQueryModels' because:
  Build cache is disabled

Issues with this output:

  1. The select doesn't state what it selected.
  2. There is no way to cleanly correlate the order of the "Select" statements with the actual saved list. The order changes for the saved results (which is why I left the names as A,B,C to illustrate the difficulty).
  3. In several runs, if we expected 14 files but only find 13, then it becomes easier to isolate the faulty query. However, this is an unreliable method to isolate bad queries. We have seen situations where the task errors out, shows less than 14 saved results but 14 files exist nonetheless.

Expected Behavior

The task should clearly out a) what file was being processed and b) provide a little more context.

melaasar commented 1 year ago

This is indeed a bug.. I plan to fix.