PTFS-Europe / rebus-list-2

rebus:list 2+ issue tracker repository
0 stars 0 forks source link

Viewing and exporting citations missing data #354

Open hloliver opened 7 years ago

hloliver commented 7 years ago

Hi We've been trying out the view/export citations feature and are finding that hardly any of the APA or Harvard styles we've tested (around 35-40) display pagination, and more than half didn't display volume or issue number for journal articles. The 2 or 3 styles that did display the pagination only displayed the start page, not the end page. I know that there are a lot of styles available, but thought that the main ones should be fairly complete.

Here's an example where the Volume, Issue and last page are missing from an article imported into RL2 using the EBSCO connector:

gipsonrefinrl2

harvardcitethemright

There also seems to be a lot of data missing when you export references, I'm attaching two files, both for the same record, 1 exported from RL2 and 1 from EBSCO and below are screenshots of these records imported into Endnote - the RL2 has omitted almost all of the publication details

Rebus2 import: ris_import_from_rl2

Part of the EBSCO export: ris_import_from_ebsco

can you take a look? thanks Helen

GIPSONfromRL2.txt GIPSONfromEBSCO.txt

AndrewIsh commented 7 years ago

Hi Helen

Thanks for that.

As far as the citation styles are concerned, the data that is displayed is determined by both the data held in the item and the citation style definition.

Looking at the definition for harvard-cite-them-right, I can't see any definition in there that would display a page range, which would explain why one isn't being displayed. In short, if a given item contains data and the citation doesn't display it, it will be because the citation style doesn't specify that that data should be displayed.

We maintain a copy of the official repository of approx. 1,300 CSL styles. I notice that some of the styles in the official repository have been updated, so I'll arrange for us to pull in the updates as soon as possible.

As far as export is concerned, I see what you mean about the exported items from RL2 being sparse compared to the export direct from EBSCO. In some cases this will be because EBSCO simply don't pass us the data when we query their API, in other cases it will be due to our export not including that data.

To try and illustrate, here is the metadata we receive from EBSCO for the item included in your example:

Teaching child and adolescent psychiatry in the twenty-first century: A reflection on the role of technology in education.

{
    "ISSN": ["1056-4993"],
    "URL": ["http:\/\/search.ebscohost.com\/login.aspx?direct=true&site=eds-live&db=psyh&AN=2016-55931-008"],
    "author": [{
        "literal": "Gipson, Shih Yee-Marie Tan"
    }],
    "categories": ["Adolescent Psychiatry", "Child Psychiatry", "Computer Assisted Instruction", "Technology", "Medical Education"],
    "container-title": "Child and Adolescent Psychiatric Clinics of North America",
    "id": ["psyh,2016-55931-008", "psyh", "2016-55931-008"],
    "issue": "1",
    "issued": {
        "raw": "2017-01-01T00:00:01Z"
    },
    "medium": ["Academic Journal"],
    "number-of-pages": 11,
    "page": "93",
    "page-first": "93",
    "publisher": "Elsevier Science",
    "shortTitle": "Teaching child and adolescent psychiatry in the twenty-fi...",
    "source": "<highlight>Child<\/highlight> Adolesc Psychiatr Clin N Am",
    "submitted": {
        "raw": "2017-01-01T00:00:01Z"
    },
    "title": "Teaching child and adolescent psychiatry in the twenty-first century: A reflection on the role of technology in education.",
    "title-short": "Teaching child and adolescent psychiatry in the twenty-fi...",
    "type": "article",
    "volume": "26"
}

Comparing that with the data contained in your export from EBSCO, I can see that we're missing a considerable number of fields in the record we obtain from EBSCO. This would account for some of the sparseness, but not all. There is still data in our record that is not being contained in the RIS export, categories, journal title etc. This clearly should be being included.

I'll create an internal issue to investigate getting these additional fields included in the export and update this issue once this has been done.

Many thanks Andrew

hloliver commented 7 years ago

Thanks Andrew,

I can see what you mean about the page range not being defined in Cite them Right (honestly, why wouldn't they!) but Volume and Issue seem to be there but not displayed. Can you let me know when you've updated the styles and I'll check again, or find a different one.

Re the export - in the sample record I've just seen that EBSCO isn't even releasing all of the authors! Shall I open a separate issue for this?

Best wishes Helen

AndrewIsh commented 7 years ago

Hi Helen

An updated set of CSL styles will be included in next Monday night's update.

Regarding the single authors. I've discussed this with Alex and he noted there was a modification to the Connect configuration required in order to pass all authors through the Connect response. He has now done this, so any items added to lists in future will contain all authors. In order to update your current list items, a data resync will be required. I'll tag @mrenvoize in this issue so he can initiate this.

I hope that helps.

Many thanks Andrew

hloliver commented 7 years ago

Hi Andrew

I’ve had a look today and can see that some of the styles have been updated but we’re still getting some discrepancies

• Citations include the start page of an article if the record has been pulled through with the EBSCO connector, though not if it’s been added manually (see the first two Klein articles on this list - https://tavi.rebuslist.co.uk/#/list/3688, the first was added with the EBSCO connector, the second added manually)

• I’ve been able to get the complete page range to display in the citation by adding it in the Misc tab – see https://www.screencast.com/t/xT6rgCBAAE0n - this only works for manually added items though, we can’t get to this if we’ve added a record through the EBSCO connector, also this field isn’t available in the Misc tab for book chapters

• Harvard-cite-them-right and APA styles aren’t displaying Volume/Issue though it looks as though they are included in the citation style definition

• Exported records are still missing a lot of data (is that something you’re still working on?)

• New records imported with the EBSCO connector are now pulling through all authors – thanks

Best wishes Helen

AndrewIsh commented 7 years ago

Hi Helen

Citations include the start page of an article if the record has been pulled through with the EBSCO connector, though not if it’s been added manually (see the first two Klein articles on this list - https://tavi.rebuslist.co.uk/#/list/3688, the first was added with the EBSCO connector, the second added manually)

I’ve been able to get the complete page range to display in the citation by adding it in the Misc tab – see https://www.screencast.com/t/xT6rgCBAAE0n - this only works for manually added items though, we can’t get to this if we’ve added a record through the EBSCO connector, also this field isn’t available in the Misc tab for book chapters

There are three "fields" in CSL that can hold page data, page, page-first and number-of-pages. It seems that the citation styles you're using use the page field for page information, if that's not populated then page information won't display.

In the "Article pages" tab on the material form, the "First page" field populates the page-first CSL field and the "Number of pages" field populates the number-of-pages CSL field, on the Misc tab, the "Page Number(s)" field populate the page CSL field. This is why you're seeing the page data when you add the value in the Misc tab.

You're right that you can't modify this for EBSCO material, but I'm seeing page data for EBSCO material, are you not?

The "Page Number(s)" field not being available for the Book Chapter material type form is an oversight, I'll add this.

Harvard-cite-them-right and APA styles aren’t displaying Volume/Issue though it looks as though they are included in the citation style definition

This appears to be because your articles are coming into rebus:list with a CSL type of article whereas the APA style definition will only display volume and issue information for article-journal, article-magazine, article-newspaper:

<if type="article-journal article-magazine article-newspaper" match="any">

I'm not sure why articles are coming through with this type, I'd expect them to come through as article-journal (assuming they're journal articles). I'll investigate why this is.

Exported records are still missing a lot of data (is that something you’re still working on?)

We're using Zotero import / export filters for our material export so I'd expect exports to match what Zotero would export for the same material. If you could let me know some specific examples I can try and dig into this further.

New records imported with the EBSCO connector are now pulling through all authors – thanks

Hurray

Many thanks Andrew

AndrewIsh commented 7 years ago

I'm not sure why articles are coming through with this type, I'd expect them to come through as article-journal (assuming they're journal articles). I'll investigate why this is.

Discussing this with Alex, it seems that it's actually very hard for us to identify journal article material as such when it comes in from EBSCO. There's currently no way in Connect for us to know whether a search result is a journal or a journal article, this is due to EBSCO not specifying in it's API response. In
order to correctly identify the type we'd need to do some Connect development work to try and identify the type based on included material metadata. I will add this as an enhancement request.