Page 2 of 3

Re: ParseResume Issue on API 2.0 (URGENT)

Posted: Fri Apr 13, 2012 5:07 pm
by rkarthy
Hi K.Pham,

I'm getting the same error while parsing the resume on our another business unit. Can you please check and let me know?

FYI. I'm using the same code what you suggested in this post. Also i'm using the real resume for parsing.

Api Username: api.c1

Error Info:
HttpResponseCode=200
InitStatus=0
InitStatusStr=
StellentStatus=0
StellentStatusStr=
RexStatus=-1
RexStatusStr=Rex has not been initialized


Regards,
Karthik

Re: ParseResume Issue on API 2.0 (URGENT)

Posted: Wed Apr 18, 2012 10:54 am
by sourov.chatterjee
Is there any update to this? I am getting the same issue. Please help.

Api Username: api.c1

Error Info:
HttpResponseCode=200
InitStatus=0
InitStatusStr=
StellentStatus=0
StellentStatusStr=
RexStatus=-1
RexStatusStr=Rex has not been initialized

If this is being worked on, please let me know.

Re: ParseResume Issue on API 2.0 (URGENT)

Posted: Mon Apr 23, 2012 5:48 pm
by jroorda
Hi sourov.chatterjee,
This iRex error is coming from our third party parsing service.
Unfortunately form time to time you will get a refused connection error: RexStatusStr=Rex has not been initialized
There is no real way to prevent this error. It will happen from time to time.
The only thing I can recommend is to build around it. For example having the parse function run again or try to parse the record later once the error was received.
Once you have a base64ChunkedResume in binary, with the content of the candidate’s resume in 76 character blocks, you have done everything to Bullhorn specifications.
The irex API should accept this information and parse it returning the xml however occasionally there can be connection errors between the Bullhorn API and the iRex API.


Sincerely,
Justin with Bullhorn Support

Re: ParseResume Issue on API 2.0 (URGENT)

Posted: Mon Apr 30, 2012 2:54 pm
by mguterl
Hi Justin,

I have run into the same error during our implementation and I have also seen a few other failures that I did not make note of. Can you provide some insight into which RexStatusStr values we should be retrying on and which are outright failures that are not recoverable? Adding this to your documentation would also be very helpful.

Also, this service seems to fail very frequently. In a small experiment 6 out of 10 calls to the service failed. We are going to be pushing over 200,000 resumes to your service and this is somewhat disconcerting.

Re: ParseResume Issue on API 2.0 (URGENT)

Posted: Tue May 01, 2012 8:08 am
by r.knight
Hi mguterl ,

Apologies for the inconvenience. We are actually noticing an increasing number of iRex initialization errors internally and through other various applications using the API's.

At this time, we (Bullhorn Support) have raised an escalation ticket to our engineering team for further analysis into the problem. As soon as I learn more about the status of the issue and our next steps, I will post them to this forum.

Thanks in advance.

Re: ParseResume Issue on API 2.0 (URGENT)

Posted: Tue May 01, 2012 8:49 am
by mguterl
Hi Ryan,

Thanks for the update concerning the failed requests. I still need a list of which errors should be retried and which are outright failures to parse. Without this list, it is hard to know which resumes the service is unable to parse and which are intermittent failures. If you could provide this, it would be very helpful.

Re: ParseResume Issue on API 2.0 (URGENT)

Posted: Tue May 01, 2012 1:48 pm
by jroorda
Hi mguterl,
We do not have a list of all the errors.
The errors are coming from working with our third party partner that parses the resume and returns the hrxml.
I would suggest working off of the "Success" part of the of the object.
It will be false if the resume did not parse and true if the resume did parse.

Here are some samples to see which part of the object the success field is:

Failed Request

Code: Select all

<S:Envelope xmlns:S="http://schemas.xmlsoap.org/soap/envelope/">
   <S:Body>
      <ns2:parseResumeResponse xmlns:ns2="http://apiservice.bullhorn.com/">
         <return>
            <session>rO0ABXVyAAJbQqzzF/gGCFTgAgAAeHAAAAIAZ0XJoWgo9Jnd5xnLWs5IU7scmIXI6yX7FhNaGKBy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</session>
            <diagnostics>HttpResponseCode=200
InitStatus=0
InitStatusStr=
StellentStatus=0
StellentStatusStr=
RexStatus=-1
RexStatusStr=Rex has not been initialized</diagnostics>
            <success>false</success>
Sucessful request:

Code: Select all

<S:Envelope xmlns:S="http://schemas.xmlsoap.org/soap/envelope/">
   <S:Body>
      <ns2:parseResumeResponse xmlns:ns2="http://apiservice.bullhorn.com/">
         <return>
            <session>rO0ABXVyAAJbQqzzF/gGCFTgAgAAeHAAAAIAyZes+YCj1B+Da2pDAxWl2NIMZyHCiLUYjUElwlLQ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</session>
            <diagnostics>HttpResponseCode=200
InitStatus=0
InitStatusStr=Initialization OK
StellentStatus=0
StellentStatusStr=KVERR_Success
RexStatus=0
RexStatusStr=No error</diagnostics>
            <hrXml><![CDATA[<?xml version="1.0" encoding="UTF-8"?>
<Resume xmlns="http://ns.hr-xml.org" xmlns:hr="http://ns.hr-xml.org/PersonDescriptors" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xmlns:bh="bullhorn.com" xml:lang="ENGLISH">
  <ResumeId idOwner="Resume Mirror">
    <IdValue name="RESUME ID"/>
  </ResumeId>

Re: ParseResume Issue on API 2.0 (URGENT)

Posted: Tue May 01, 2012 3:39 pm
by mguterl
I know where the success field is, thanks. So you're telling me that I should retry any time success is false? What happens when it fails to parse the resume altogether? Shouldn't this failure be indicated in a different way than an intermittent failure?

Re: ParseResume Issue on API 2.0 (URGENT)

Posted: Tue May 01, 2012 3:54 pm
by jroorda
Hi mguterl,
I do recommend that you should retry any time success is false.
However I would put a limit on it so that it does not loop forever.
Unfortunately this is intermittent failure we are working to decrease the rate of failure however it will always be there.
For other work around you can take a look at the Diagnostics.

The connect error will always say:
RexStatusStr=Rex has not been initialized

Code: Select all

<diagnostics>HttpResponseCode=200
InitStatus=0
InitStatusStr=
StellentStatus=0
StellentStatusStr=
RexStatus=-1
RexStatusStr=Rex has not been initialized</diagnostics>
            <success>false</success>
However there are a significant variety of errors that can occur and the RexStatusStr will be different based upon the error.
For example:

Code: Select all

<diagnostics>HttpResponseCode=200
InitStatus=0
InitStatusStr=Initialization OK
StellentStatus=4
StellentStatusStr=KVERR_badInputStream
RexStatus=7
RexStatusStr=Html conversion error</diagnostics>
            <success>false</success>

Re: ParseResume Issue on API 2.0 (URGENT)

Posted: Wed May 02, 2012 12:36 pm
by mguterl
Thanks that is what I was looking for. We're going to log all of these diagnostics and I'll report back for clarification if I find anything else.

Re: ParseResume Issue on API 2.0 (URGENT)

Posted: Tue Aug 07, 2012 5:22 pm
by Snookums Dread
Some questions for Bullhorn Support.

While working today with parseResume() in API 2.0, I have been able to successfully retrieve results only once.
I have received repeated "RexStatus=-1" responses, to the extent that this should be regarded as the norm.
I do not know if this reflects the general situation with communications between Bullhorn and your third-party parsing service.

If the failure rate is generally this high, has there been any discussion or attempted resolution of this issue with your third-party parsing service? Should this service no longer be relied upon?

If the failure rate is generally much lower, is there any way to account for (or avert) the high number of "RexStatus=-1" responses that I am receiving?
Is it possible, for instance, that this diagnostic is being returned due to Bullhorn's handling of some other error in my code?
[edit: for instance, initiating a connection, and then aborting the process due to a Dto -> ??? conversion error, et cetera.]

A status update on the parseResume() situation would be appreciated.

Re: ParseResume Issue on API 2.0 (URGENT)

Posted: Fri Aug 17, 2012 3:12 pm
by jroorda
Hi Snookums Dread,
This is Justin with Bullhorn Support.
Our parsing service is highly integrated and there is no better alternative currently.
I am afraid that the repeated "RexStatus=-1" responses are a reflection the general situation.
We are in the process of improving our API and the connection with our third-party parsing service however these improvements are months away.

As too the failure reasons, the most common will be a connection error. However conversion error and documents that are unable to be parsed, like images, will always throw Rex errors and never parse.
Where with connection errors if you try the operation several times it will eventually return the XML.


Unfortunately currently the only course of action is to develop around these errors and account for the parsing problems.
I apologize a do understand this is and hurdle and an annoyance to the parsing process however currently it is a reality we are gong to have to work with.

Re: ParseResume Issue on API 2.0 (URGENT)

Posted: Mon Nov 05, 2012 12:03 pm
by kkatusic
Is any update about this?

Re: ParseResume Issue on API 2.0 (URGENT)

Posted: Mon Nov 05, 2012 12:29 pm
by c.choy
Hello,

There was a deploy that went out a while back to reduce the number of failures. That being said, the depencency is still present and you should still loop around the call to parse the resume. How many times is up to your discrection. I've done loops of 5 and still encounter times where another instance will make it work, however my experience would not be exactly equal to your own.

Sorry I can't provide a good hard number, you will want to try and see what works and you're comfortable using.

Re: ParseResume Issue on API 2.0 (URGENT)

Posted: Tue Nov 06, 2012 8:20 am
by kkatusic
Thx @c.choy, I worked around and make it work with PHP:

Code: Select all

$resultcontent = file_get_contents($uri);
$contents64 =  chunk_split(base64_encode($resultcontent), 64, "\n");
and now I am getting no error, can I attach that content to Resume field in Bullhorn?