621: use default connection

[al-niessner]

🗒️ Summary

Use the app://connections/cognito/registry.xml as the default index since it is an alias that connects all of the indices to one name.

⚙️ Test Data and/or Report

NA - user interface change not a procedural change

♻️ Related Issues

Closes #621
Closes #1053

[nutjob4life]

I must be doing something wrong. 

I tried this but got two test failures; related to the "Depends on part 1 of 2 from NASA-PDS/registry-common@121" I assume.

$ mvn package
…
[ERROR] Tests run: 251, Failures: 2, Errors: 0, Skipped: 0, Time elapsed: 346.4 s <<< FAILURE! -- in cucumber.CucumberTest
[ERROR] < 3.6.NASA-PDS/validate#690-1 -- Time elapsed: 19.37 s <<< FAILURE!
org.opentest4j.AssertionFailedError: summary:totalWarnings ==> expected: <888> but was: <900>
    at cucumber.StepDefs.compare_to_the(StepDefs.java:182)
    at ✽.compare to the expected outcome "summary:totalWarnings=888,summary:messageTypes:warning.table.field_value_out_of_special_constant_min_max_range=888".(file:///Users/kelly/Documents/Clients/JPL/PDS/Development/nasa-pds/validate/src/test/resources/features/pre.3.6.x.feature:5)

[ERROR] < 3.6.NASA-PDS/validate#427-1 -- Time elapsed: 17.79 s <<< FAILURE!
org.opentest4j.AssertionFailedError: summary:totalWarnings ==> expected: <2> but was: <14>
    at cucumber.StepDefs.compare_to_the(StepDefs.java:182)
    at ✽.compare to the expected outcome "summary:totalWarnings=2,summary:messageTypes:warning.table.field_value_out_of_special_constant_min_max_range=2".(file:///Users/kelly/Documents/Clients/JPL/PDS/Development/nasa-pds/validate/src/test/resources/features/pre.3.6.x.feature:5)

[INFO] 
[INFO] Results:
[INFO] 
[ERROR] Failures: 
[ERROR]   summary:totalWarnings ==> expected: <2> but was: <14>
[ERROR]   summary:totalWarnings ==> expected: <888> but was: <900>
[INFO] 
[ERROR] Tests run: 274, Failures: 2, Errors: 0, Skipped: 23
…

So then I tried building the distribution and skipping the tests:

$ mvn -DskipTests=true package
…
$ tar xzf target/validate-3.7.0-SNAPSHOT-bin.tar.gz
$ validate-3.7.0-SNAPSHOT/bin/validate --help | egrep -e -r
 -r,--report-file <file name>             Specify the report file name. Default
 -R,--rule <validation rule name>         Specifies the validation rules to
 -s,--report-style <full|json|xml>        Specify the level of detail for the
    --skip-context-reference-check        Disable context reference check. The
                                          to include in the human-readable

and I don't see the new -r option.

I think I better just let someone more experienced review this! 😁

[al-niessner]

@nutjob4life

No. This is because the JAXA website vanishes for long periods of time and we cannot get the data we need for testing. There is a ticket to download said data and put it into validate test data to remove this problem.

Oh, this is for validate-refs not validate. Tool does not belong here software logically. However it is here because those that run validate will want to run validate-refs and cannot(?) download registry-mgr. I suspect it is keep the bar low (low effort to adopt).

[nutjob4life]

@al-niessner thanks for the explanation!

[sonarqubecloud]

Quality Gate failed

Failed conditions
1 Security Hotspot

See analysis details on SonarQube Cloud

[al-niessner]

@jordanpadams

Okay. Using my computer at JPL on JPLNet this is what I see with changes on this branch and the index of en-registry (showing it overrides what I told it):

10:28:10.398 [Thread-0] ERROR gov.nasa.pds.validate.ri.CommandLineInterface - Found duplicate file: /data/pds4/documents/xml_schema/sp/v1/PDS4_SP_1700.sch
10:28:10.542 [Thread-0] ERROR gov.nasa.pds.validate.ri.CommandLineInterface - Found duplicate file: /data/pds4/documents/xml_schema/sp/v1/PDS4_SP_1700.xsd

Hopefully that is what you expected. If you review #1146 I will merge those changes to here and make it so the unit tests pass. Or I guess you can do it.

[al-niessner]

@jordanpadams

Just doing clean up of issue and saw that it worked for you with en-registry but not registry. What are all of the non-alias registry names? Just one other valid one will do. I tried en-registry-fred and it got the timeout error you did. It seems something is afoot but I need another valid index name to start the chase.

[al-niessner]

@jordanpadams

There is something else amiss beyond our code. I went back to en-registry and it fails too with a timeout. The timeout is big like minutes. I am going to let it idle for a bit then try again to see what happens if I do the same thing twice in a row.

[jordanpadams]

@al-niessner your account will only work with en-registry (I think). Each user gets access to 1 index. We will have to get @sjoshi-jpl @viviant100 or @tloubrieu-jpl to help us add you to other indexes if we need to test that. They should all work the same.

[jordanpadams]

@al-niessner I think I mentioned this before, but as an extra sanity check for your access, the registry-client is super simple, and really helpful IMO. https://nasa-pds.github.io/registry-client/

Once I configured my login and the appropriate info (same as my validate-refs config), I get this when trying to ping the registry alias:

$ pds-registry-client '/registry/_search' --data '{"fields":["lidvid"],"query":{"match":{"lidvid":"urn:nasa:pds:context:telescope:parkes.dss49_64m::1.1"}},"_source": false}' --pretty
{
  "took": 72,
  "timed_out": false,
  "_shards": {
    "total": 0,
    "successful": 0,
    "skipped": 0,
    "failed": 0
  },
  "hits": {
    "total": {
      "value": 1,
      "relation": "eq"
    },
    "max_score": 5.99446,
    "hits": [
      {
        "_index": "en-registry",
        "_id": "urn:nasa:pds:context:telescope:parkes.dss49_64m::1.1",
        "_score": 5.99446,
        "fields": {
          "lidvid": [
            "urn:nasa:pds:context:telescope:parkes.dss49_64m::1.1"
          ]
        }
      }
    ]
  }
}

[al-niessner]

@jordanpadams

validate-refs is Java SDK2. pds-registry-client is python and presumably some version of boto. The failure is in the Java SDK2, which claims they seamlessly handle aliases, not in validate-refs our code. Are you trying to tell me that you want me to rewrite validate-refs in python (I am not opposed) because we know that the python SDK works and is better supported? Are you saying that pds-registry-client already figured out the alias issues and I should be looking at their code?

side note: I have isolate the problem to be the alias and not any of the real nodes

[al-niessner]

@jordanpadams

Asking for additional help/guidance

[jordanpadams]

@al-niessner I was not asking to rewrite this in Python, I was just saying it may be good to confirm you have sufficient access to the alias outside of the SDK to check that off as a possible reason for the timeout. If you are confident that is not the case, then that works for me.