Indexing failures in a few studies #34
Comments
Indexing is failing for some studies. |
The cURL call that fails to return the studies for the list is:
(via copying the call form the chrome dev tools that is loading the curator). |
@chinchliff @blackrim Any idea what is going on here? Is this the same (or similar) problem as discussed here? Currently, not all of the studies in the synthetic tree are showing up in the curator app. |
I am not sure what is going on. What errors are returned when indexing is On Mon, Jun 15, 2015 at 12:15 PM Karen Cranston notifications@github.com
|
@chinchliff , here (see below) are the three errors described above with nicer formatting. But that was back in January, so it's probably best to attempt a fresh re-indexing and monitor the output, using cURL or Study ot_232 reports multiple records(?) found, but there's just one JSON with this id in phylesystem-1. This is very similar to a problem discussed in #26:
Two others failed with no explanation:
|
This is a major problem, top priority. I'll take a look at it; any help is welcome. |
There are multiple issues here. For study ot_232, I have no idea why it The other errors can be retrieved using curl against the devapi OTI. They I am not sure why those errors are not showing up in the output of the curl -X POST http://devapi.opentreeoflife.org/v2/studies/index_studies -H { "message" : "The node Tn295230 is identified as a leaf but has not been "exception" : "NexsonParseException", "fullname" : "org.opentree.nexson.io.NexsonParseException", "stacktrace" : [ } curl -X POST http://devapi.opentreeoflife.org/v2/studies/index_studies -H { "message" : "The node Tn295230 is identified as a leaf but has not been "exception" : "NexsonParseException", "fullname" : "org.opentree.nexson.io.NexsonParseException", "stacktrace" : [ } On Wed, Jul 15, 2015 at 12:09 PM Jonathan A Rees notifications@github.com
|
Thanks. ot_31 seems to fail the validator. This comes from the treebase converter, so the converter probably needs to change. If it fails to validate I don't know how it got into the repo. I'll copy this to a new peyotl issue. |
The indexing script is using the ext/studies/graphdb/index_study call, On Wed, Jul 15, 2015 at 3:33 PM, Cody Hinchliff notifications@github.com
|
I wonder if this can be done with a browser proxy configuration instead of On Wed, Jul 15, 2015 at 9:09 PM, Jonathan A Rees rees@mumble.net wrote:
|
Last message "I wonder ..." pertains to a different issue. Please ignore.
|
For ot_31 and 32, it looks to me as if we can just replace the throw with a return null, to get out of our immediate predicament, pending resolution of OpenTreeOfLife/peyotl#126. Sound right? |
Re ot_232, here's what I get: curl -X POST http://localhost:7478/db/data/ext/studies/graphdb/index_studies -H 'content-type:application/json' -d '{"urls": ["http://api.opentreeoflife.org/phylesystem/v1/../default/v1/study/ot_232.json"]}' |
I indexed production phylesystem a few days ago, and there were no failures. So this issue is no longer critical. |
Today I used the deployment script to re-index oti, and a few studies failed (details below).
Study ot_232 reports multiple records(?) found, but there's just one JSON with this id in phylesystem-1. This is very similar to a problem discussed in #26:
Two others failed with no explanation:
The text was updated successfully, but these errors were encountered: