search

elifesciences / enhanced-preprints-import

Enhanced Preprints import system
1 stars 0 forks source link

MSID: 89337 Version: 2 DOI: 10.48550/arXiv.2301.00759 #2421

Open nlisgo opened 9 months ago

nlisgo commented 9 months ago

MSID: 89337

Version: 2

Preprint DOI: https://doi.org/10.48550/arXiv.2301.00759

Step 1. Awaiting reviews

Editorial to post reviews via hypothesis

Useful links:

For trouble shooting (e.g. no Docmaps available):

Step 2. Preview reviewed preprint

Production QC content ahead of publication

Instructions:

Useful links:

Step 3: Awaiting search reindex

This step adds the reviewed preprint to the homepage: https://elifesciences.org

The search reindex is triggered once an hour. We need the reviewed preprint to be indexed as the search application serves the journal homepage.

Useful links:

Step 4: Published! PDF requested

Waiting for PDF to be generated

Useful links:

Step 5: Introduce PDF to data folder and git repo

Upload PDF to relevent folder in git repo https://github.com/elifesciences/enhanced-preprints-data/

Step 6: Done!

JGilbert-eLife commented 9 months ago

I'm not sure the correct version of the article has been loaded here. I'm seeing discrepancies between this and what's in the meca package http://prod-elife-epp-meca.s3.amazonaws.com/89337-v2-meca.zip.

E.g. there should be four affiliations, extra text in the legend for Figure 1 etc.

Image

Also, the reviews don't include an author response, but that is present on the V1. I think this might be another case of them being muddled between versions.

fred-atherden commented 9 months ago

Thanks James. On Temporal, I see that the workflow has failed. The message is:

"cause": {
    "message": "Preprint doesn't have a published date",
    "source": "TypeScriptSDK",
    "stackTrace": "NonRetryableError: Preprint doesn't have a published date\n    at Activity.generateVersionJson [as fn] (/app/src/activities/generate-version-json.ts:32:11)",
    "encodedAttributes": null,
    "cause": null,
    "applicationFailureInfo": {
      "type": "NonRetryableError",
      "nonRetryable": false,
      "details": null
    }

In the docmap we have a missing publication date for the preprint it seems:

"inputs": [
    {
        "type": "preprint",
        "doi": "10.48550/arXiv.2301.00759",
        "url": "https://arxiv.org/abs/2301.00759v3",
        "versionIdentifier": "3",
        "content": [
            {
                "type": "computer-file",
                "url": "s3://prod-elife-epp-meca/89337-v2-meca.zip"
            }
        ]
    }
],

So I'm adding the preprint v3 pub date to this spreadsheet in order for the preprint v3 publication date to be introduced into the DocMap.

I hope that once this is recognised by DataHub (hopefully within the next few hours), EPP will recognise the DocMap has updated, and will import the content at s3://prod-elife-epp-meca/89337-v2-meca.zip.

I think the preview must be showing the old content for now (which isn't particularly helpful behaviour).

fred-atherden commented 9 months ago

@acollings do you know if there should be an author response on the v2 of this one?

JGilbert-eLife commented 9 months ago

@fred-atherden Just to check - this is a v2, not a v3. Is the discrepancy in the steps you took deliberate?

fred-atherden commented 9 months ago

@JGilbert-eLife, sorry I was referring to the preprint version (reviewed preprint v1 = preprint v2; reviewed preprint v2 = preprint v3). Let me know if that's not what you were expecting.

JGilbert-eLife commented 9 months ago

Ah, of course! Thanks for clarifying!

acollings commented 9 months ago

Hi @fred-atherden, yes! A couple of things. We're missing the author response but also the review here, https://prod--epp.elifesciences.org/previews/89337v2/reviews#tab-content (and assessment), is the original one, not the updated one.

So I think something has gone horribly wrong with the posting of the assessment/reviews/response. I've got a query out to Sue.

fred-atherden commented 9 months ago

Thanks Andy!

Apologies, I think the behaviour of EPP is really quite bizarre if a preprint match can't be found. I believe the v2 preview is actually just showing the preview for the v1, which would explain the mismatch for the reviews. I think EPP first pulls in the preprint content and then the reviews. If it fails to pull in the preprint content it will fail to pull in the latest reviews, and by default it will just show previous content (when ideally it would show an error page or similar).

For the v2 in the DocMap we have the following in the DocMap:

Aside from the missing author response, if these two look correct to you then we're likely OK for the reviews.

fred-atherden commented 9 months ago

Doesn't look like the preprint date has made it's way to the DocMap yet. I have asked Hazal in slack here.

acollings commented 9 months ago

Hi @fred-atherden cool, thanks! That's good. The author response was also posted on Jan 22:

https://hyp.is/3bp4hrl4Ee6jLV85xpEAvA/arxiv.org/abs/2301.00759

Can you let me know if we should delete them all and repost? Not really sure what to do next! :-)

fred-atherden commented 9 months ago

Hi @acollings. It looks like the author response has been added to https://arxiv.org/abs/2301.00759 instead of https://arxiv.org/abs/2301.00759v3, meaning it has now appeared on the RPv1.

If you don't mind, I think it would be worth deleting them all and reposting them to the v3 link.

fred-atherden commented 9 months ago

Image

fred-atherden commented 9 months ago

@JGilbert-eLife the preprint content should now be fixed I think. Might be worth waiting for the reviews before QCing again in any case.

acollings commented 9 months ago

(The reviews have been reposted)

JGilbert-eLife commented 9 months ago

Article QC is OK but reviews not updated yet -- and I see there's an issue with the pipelines this morning so I'll check again later.

fred-atherden commented 8 months ago

Hi @acollings, this is what we have now:

Image

They're all on the same unversioned link.

Any ideas on how to fix? Is this an issue with Kotahi?

Links in case they're useful:

acollings commented 8 months ago

Thanks @fred-atherden Sue and I have tried deleting/re-posting several times to no avail, and I've tried using hypothes.is directly. I think we probably need some help (from Hazal or Daniel)? I don't mind deleting/re-posting, but at this point I don't know what we've done wrong in previous attempts (or what to try next)!

fred-atherden commented 8 months ago

Thanks Andy! No worries.

@de-code, @hazalCiplak. Any suggestions on how to proceed here? We can't seem to be able to get the hypothesis comments attached to the right preprint versions, so that dataHub can associate them with the right RP version

de-code commented 8 months ago

We experienced the issue previously in #1863 The solution was to re-post the Hypothes.is annotation via Kotahi rather than manually via the browser. The latter won't post for the specific version. That will be an issue in any case when trying to add evaluations to multiple versions.

I would like to explore the option of updating the URL of a Hypothes.is annotation via an API though. That is because deleting Hypothes.is annotation has some undesired side effects:

I don't know whether the Hypothes.is URL can be updated via the API. I could try that tomorrow.

acollings commented 8 months ago

Thanks! Both sets re-posted using Kotahi.

de-code commented 8 months ago

I've also investigated updating the preprint URL via the API and I can confirm that it works (I tested it with a private annotation using my own account).

I would propose to do that going forward. Ideally Kotahi had the option to do, but we could also think about alternatives. In the short term we could do that manually.

fred-atherden commented 8 months ago

Many thanks @de-code!

@JGilbert-eLife, do you think we're good to go now?

JGilbert-eLife commented 8 months ago

Looks good to me!