7056 Skip unknown types in a dataset which cause the whole dataset to… #7057

JingMa87 · 2020-07-06T00:12:36Z

… fail from being harvested

What this PR does / why we need it: Makes datasets with unknown types harvestable using dataverse_json

Which issue(s) this PR closes: 7056

Special notes for your reviewer: Good fix

Suggestions on how to test this: Harvest the Princeton_Authored_Datasets OAI set from the harvard repo using dataverse_json. Without this fix you will get 267 successful harvest and 2 fails. With the fix you will harvest all 269 datasets. The controversial datasets are https://dataverse.harvard.edu/api/datasets/export?exporter=dataverse_json&persistentId=doi:10.7910/DVN/B6OJKG and https://dataverse.harvard.edu/api/datasets/export?exporter=dataverse_json&persistentId=doi:10.7910/DVN/I5O6OS.

Does this PR introduce a user interface change? If mockups are available, please link/include them here: No

Is there a release notes update needed for this change?: I think it's an important change for harvesting from another dataverse.

Additional documentation: N/A

… fail from being harvested

coveralls · 2020-07-06T00:19:16Z

Coverage increased (+0.003%) to 19.656% when pulling 057a8af on JingMa87:7056-skip-unknown-field-type into 6daf219 on IQSS:develop.

pdurbin

I think this is fine so I'm moving it to QA. If other developers want to talk a look, please go ahead. We've never promoted using the dataverse_json format for harvesting (I believe DDI is preferred) but this should make it more possible to use it.

(I don't think it's worth sending back to the contributor over the wildcard imports but if someone feels strongly about this we certainly can.)

pdurbin · 2020-07-09T15:31:30Z

src/main/java/edu/harvard/iq/dataverse/util/json/JsonParser.java

-import edu.harvard.iq.dataverse.FileMetadata;
-import edu.harvard.iq.dataverse.MetadataBlockServiceBean;
-import edu.harvard.iq.dataverse.TermsOfUseAndAccess;
+import edu.harvard.iq.dataverse.*;


I mentioned this in a review on another pull request but we favor using the full import statement instead of wildcards.

I'll reverse the change, my Intellij automatically does this. Super nice feature though otherwise.

Yeah, I've heard that about Intellij. Thanks for reversing the change.

pdurbin · 2020-07-09T15:37:58Z

src/main/java/edu/harvard/iq/dataverse/util/json/JsonParser.java

@@ -566,7 +553,8 @@ public DatasetField parseField(JsonObject json, Boolean testType) throws JsonPar


        if (type == null) {
-            throw new JsonParseException("Can't find type '" + json.getString("typeName", "") + "'");
+            logger.info("Can't find type '" + json.getString("typeName", "") + "'");


Is this going to happen a lot? If so, should it be reduced from "info" to "fine" so it doesn't clutter the logs?

@pdurbin This'll happen potentially multiple thousands of times when doing a harvesting run on a big OAI set for the first time. So I'll just change it to "fine".

jggautier · 2020-07-09T19:23:01Z

We've never promoted using the dataverse_json format for harvesting (I believe DDI is preferred) but this should make it more possible to use it.

Maybe not in the guides but I've promoted it to reduce information loss and most of the harvesting Harvard Dataverse does from other Dataverse installations is using dataverse_json. I would think that DDI Codebook would be preferred if all metadata could be mapped to it, but that's probably going to be less and less true with increasing support for other standards.

kcondon · 2020-07-10T16:31:03Z

@JingMa87 I am still seeing the two failures.

<record>
  <date>2020-07-10T16:28:10</date>
  <millis>1594398490946</millis>
  <sequence>8513</sequence>
  <logger>edu.harvard.iq.dataverse.harvest.client.HarvesterServiceBean.kc0605202020-07-10T16-27-22</logger>
  <level>SEVERE</level>
  <class>edu.harvard.iq.dataverse.harvest.client.HarvesterServiceBean</class>
  <method>logGetRecordException</method>
  <thread>187</thread>
  <message>Exception processing getRecord(), oaiUrl=https://dataverse.harvard.edu/oai, identifier=doi:10.7910/DVN/B6OJKG, edu.harvard.iq.dataverse.api.imports.ImportException, Failed to import harvested dataset: class edu.harvard.iq.dataverse.engine.command.exception.CommandException (Command [DatasetCreate dataset:20721] failed: Exception thrown from bean: javax.ejb.EJBTransactionRolledbackException: Exception thrown from bean: org.apache.solr.client.solrj.impl.HttpSolrClient$RemoteSolrException: Error from server at http://localhost:8983/solr/collection1: ERROR: [doc=dataset_20721] unknown field 'mraCollection')</message>
</record>

JingMa87 · 2020-07-10T23:36:37Z

@JingMa87 I am still seeing the two failures.

<record>
  <date>2020-07-10T16:28:10</date>
  <millis>1594398490946</millis>
  <sequence>8513</sequence>
  <logger>edu.harvard.iq.dataverse.harvest.client.HarvesterServiceBean.kc0605202020-07-10T16-27-22</logger>
  <level>SEVERE</level>
  <class>edu.harvard.iq.dataverse.harvest.client.HarvesterServiceBean</class>
  <method>logGetRecordException</method>
  <thread>187</thread>
  <message>Exception processing getRecord(), oaiUrl=https://dataverse.harvard.edu/oai, identifier=doi:10.7910/DVN/B6OJKG, edu.harvard.iq.dataverse.api.imports.ImportException, Failed to import harvested dataset: class edu.harvard.iq.dataverse.engine.command.exception.CommandException (Command [DatasetCreate dataset:20721] failed: Exception thrown from bean: javax.ejb.EJBTransactionRolledbackException: Exception thrown from bean: org.apache.solr.client.solrj.impl.HttpSolrClient$RemoteSolrException: Error from server at http://localhost:8983/solr/collection1: ERROR: [doc=dataset_20721] unknown field 'mraCollection')</message>
</record>

@kcondon I can't reproduce the error. I've looked into the DatasetCreate Command but without the actual error happening it's like looking for a needle in a haystack. Are you on the latest version with the latest changes? What happens if you re-run the harvest, does it work the second time?

kcondon · 2020-07-11T15:00:46Z

@JingMa87 I have run it again and still see the failures. Running it again just results in 0 harvested, 0 failed. Looks like failures do are not harvested again.

Is it possible you have some unchecked in code on your local build env?

JingMa87 · 2020-07-11T17:57:23Z

@kcondon I pushed all my changes to the branch already. What if you delete the harvesting client, add it again and re-run, do you still get the issue? Sometimes I get weird intermittent errors that disappear if I retry.

kcondon · 2020-07-11T22:05:49Z

Yes, I have done that several times, every time I test. You could add debug code and I could run a harvest to see what it thinks it is doing.

…

________________________________ From: Jing Ma <notifications@github.com> Sent: Saturday, July 11, 2020 1:57 PM To: IQSS/dataverse <dataverse@noreply.github.com> Cc: Condon, Kevin M <kcondon@hmdc.harvard.edu>; Mention <mention@noreply.github.com> Subject: Re: [IQSS/dataverse] 7056 Skip unknown types in a dataset which cause the whole dataset to… (#7057) @kcondon<https://urldefense.proofpoint.com/v2/url?u=https-3A__github.com_kcondon&d=DwMCaQ&c=WO-RGvefibhHBZq3fL85hQ&r=TUpjWt9sVfaAC8ETCY_cDPtqJKl7s242PLg6-Wx6UpM&m=XGIPKdwAYwbhchG7w37k3GokfK48iJOWQDNIySxBG34&s=lzLQT3-iK0_oeScaJwjtKPwczT-TPyvMwCgOX_lpDRU&e=> I pushed all my changes to the branch already. What if you delete the harvesting client, add it again and re-run, do you still get the issue? — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub<https://urldefense.proofpoint.com/v2/url?u=https-3A__github.com_IQSS_dataverse_pull_7057-23issuecomment-2D657103689&d=DwMCaQ&c=WO-RGvefibhHBZq3fL85hQ&r=TUpjWt9sVfaAC8ETCY_cDPtqJKl7s242PLg6-Wx6UpM&m=XGIPKdwAYwbhchG7w37k3GokfK48iJOWQDNIySxBG34&s=h3h8pBuL7S1lNA6OI3FPhjBxeWtdp84WPyr4-zwA3sc&e=>, or unsubscribe<https://urldefense.proofpoint.com/v2/url?u=https-3A__github.com_notifications_unsubscribe-2Dauth_ABGAYCFNPTSDUF2GU5C5VKDR3CRZBANCNFSM4ORDOAZQ&d=DwMCaQ&c=WO-RGvefibhHBZq3fL85hQ&r=TUpjWt9sVfaAC8ETCY_cDPtqJKl7s242PLg6-Wx6UpM&m=XGIPKdwAYwbhchG7w37k3GokfK48iJOWQDNIySxBG34&s=uF_8KKQMqrFnXeuUu3Rkp_kIb-eqLYGIvBi2dv4QDME&e=>.

JingMa87 · 2020-07-12T22:10:40Z

Does this query in the database return something for you? If so, this row is not supposed to be there and stops the field type mraCollection from being filtered out by my code change. Then, it gets passed to Solr and the error happens.
SELECT * FROM public.datasetfieldtype where name = 'mraCollection';

I've looked into adding debugging code, but the class where the error happens (CreateHarvestedDatasetCommand) is a whole big mess of polymorphism, generics and interfaces gone wrong. When I use the debugger, I can't inspect the execute() function of that Command where the error happens so I'm relegating to reading the code. Hopefully this helps.

kcondon · 2020-07-13T13:17:10Z

@JingMa87 Yes, that sql query returns a record. Why do you say it should not be there? Also, wouldn't you have seen the same record in your own test? I understand about the code complexity. I was thinking about a lower tech debugging of adding debug logging if that helps.

Ok, I think I understand what you were getting at: the purpose of the fix is so that harvest clients who do not have this field in their db can skip a field that is present on the harvest server/source. I did not consider this when I looked at the test case. However, it should actually work normally in the case where it is present? I will test with and without it on the client so at least we have information and compare to the code currently on develop branch.

A quick test using another branch shows json format harvest also fails when the field is present on both client and server. Although this should work, I believe it is outside the scope of this ticket. Will confirm the original intent, masking the fields that are not present on the client.

JingMa87 · 2020-07-14T04:47:55Z

@kcondon I made a new ticket for clients that do have the field in the DB #7076 and will try to fix that in the coming days.

7056 Skip unknown types in a dataset which cause the whole dataset to…

e4fef1d

… fail from being harvested

pdurbin approved these changes Jul 9, 2020

View reviewed changes

JingMa87 added 2 commits July 10, 2020 07:00

4597 Change log level to fine.

1e0ea12

4597 Change back to full import statements.

057a8af

kcondon self-assigned this Jul 10, 2020

kcondon mentioned this pull request Jul 11, 2020

Harvest: Harvesting appears to just die silently intermittently, using Princeton set in prod #7074

Closed

kcondon merged commit 865c57b into IQSS:develop Jul 13, 2020

kcondon mentioned this pull request Jul 13, 2020

Harvest: Json format harvest fails with unknown field when field exists on client. #7075

Closed

JingMa87 deleted the 7056-skip-unknown-field-type branch July 14, 2020 14:16

djbrooke added this to the Dataverse 5 milestone Jul 14, 2020

pdurbin added the Feature: Harvesting label Apr 13, 2022

pdurbin mentioned this pull request Apr 13, 2022

Spike: Inventory and prioritize all existing Harvesting related issues IQSS/dataverse-pm#24

Closed

3 tasks

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

7056 Skip unknown types in a dataset which cause the whole dataset to… #7057

7056 Skip unknown types in a dataset which cause the whole dataset to… #7057

JingMa87 commented Jul 6, 2020

coveralls commented Jul 6, 2020 •

edited

Loading

pdurbin left a comment

pdurbin Jul 9, 2020

JingMa87 Jul 10, 2020

pdurbin Jul 10, 2020

pdurbin Jul 9, 2020

JingMa87 Jul 10, 2020

jggautier commented Jul 9, 2020

kcondon commented Jul 10, 2020 •

edited

Loading

JingMa87 commented Jul 10, 2020

kcondon commented Jul 11, 2020 •

edited

Loading

JingMa87 commented Jul 11, 2020 •

edited

Loading

kcondon commented Jul 11, 2020 via email

JingMa87 commented Jul 12, 2020

kcondon commented Jul 13, 2020 •

edited

Loading

JingMa87 commented Jul 14, 2020 •

edited

Loading

7056 Skip unknown types in a dataset which cause the whole dataset to… #7057

7056 Skip unknown types in a dataset which cause the whole dataset to… #7057

Conversation

JingMa87 commented Jul 6, 2020

coveralls commented Jul 6, 2020 • edited Loading

pdurbin left a comment

Choose a reason for hiding this comment

pdurbin Jul 9, 2020

Choose a reason for hiding this comment

JingMa87 Jul 10, 2020

Choose a reason for hiding this comment

pdurbin Jul 10, 2020

Choose a reason for hiding this comment

pdurbin Jul 9, 2020

Choose a reason for hiding this comment

JingMa87 Jul 10, 2020

Choose a reason for hiding this comment

jggautier commented Jul 9, 2020

kcondon commented Jul 10, 2020 • edited Loading

JingMa87 commented Jul 10, 2020

kcondon commented Jul 11, 2020 • edited Loading

JingMa87 commented Jul 11, 2020 • edited Loading

kcondon commented Jul 11, 2020 via email

JingMa87 commented Jul 12, 2020

kcondon commented Jul 13, 2020 • edited Loading

JingMa87 commented Jul 14, 2020 • edited Loading

coveralls commented Jul 6, 2020 •

edited

Loading

kcondon commented Jul 10, 2020 •

edited

Loading

kcondon commented Jul 11, 2020 •

edited

Loading

JingMa87 commented Jul 11, 2020 •

edited

Loading

kcondon commented Jul 13, 2020 •

edited

Loading

JingMa87 commented Jul 14, 2020 •

edited

Loading