Upgrade Storage"Retrieve Metadata for PDF" not working as expected
Hi all,
I usually add new items to my Zotero collection by dragging a PDF I have on my hard-drive into Zotero and then choosing the right-click menu option "Retrieve Metadata for PDF" to have Zotero automatically create an item that has the PDF as an attachment and that fills in automatically all the tags (Title, Author, etc).
The problem is, the metadata retrieval isn't always working, even for PDFs that you would expect it to work for, such as recent papers. If I look at the Document Properties of one of those PDFs, for example, I see that the Title is "doi:10.1016/j.neubiorev.2009.03.005", which should presumable enable Zotero to retrieve all the metadata from the DOI database!
An example of a PDF that did *not* have problems retrieving metadata was one that had Title = "PII: 0013-4694(93)90006-H", so presumably also a link to an online database, only this time a "working" one, as opposed to the previous example!
Is there any way to make this feature more useful, for example by having Zotero search for metadata in more places? Many thanks in advance for any help!
I usually add new items to my Zotero collection by dragging a PDF I have on my hard-drive into Zotero and then choosing the right-click menu option "Retrieve Metadata for PDF" to have Zotero automatically create an item that has the PDF as an attachment and that fills in automatically all the tags (Title, Author, etc).
The problem is, the metadata retrieval isn't always working, even for PDFs that you would expect it to work for, such as recent papers. If I look at the Document Properties of one of those PDFs, for example, I see that the Title is "doi:10.1016/j.neubiorev.2009.03.005", which should presumable enable Zotero to retrieve all the metadata from the DOI database!
An example of a PDF that did *not* have problems retrieving metadata was one that had Title = "PII: 0013-4694(93)90006-H", so presumably also a link to an online database, only this time a "working" one, as opposed to the previous example!
Is there any way to make this feature more useful, for example by having Zotero search for metadata in more places? Many thanks in advance for any help!
I can't tell you why it failed for the document in question, but usually it's doing well for recent pdfs - there is also always the question of whether DOI lookup works for a given doi - have you tried "add by identifier for this one?
But why don't you use Zotero's own import features - i.e. the URL bar item? I understand that initially, to get your pdfs into Zotero, you need the retrieve function - but then?
Am I doing things the hard way here?.. :-)
If it doesn't, your data is still likely to be better and more reliable, as it's right from the source and Zotero doesn't have to guess - and you can just attach the pdf afterwards (download and drag from the file system).
Google scholar data, which you'll often get for "retrieve" is pretty good, but not super reliable - and, for example you never get full first names, only initials.
However I just tried applying the method you suggested on a new article, and this time it worked, so maybe some bugs were fixed in the meantime (well, I'm guessing they are all the time).
Many thanks once again for your help adam!
Just a few moments ago I was importing metadata from several PDFs of articles published in the Journal of the Indian Academy of Forensic Medicine. This journal doesn't use the DOI system. Thus, Zotero uses Google Scholar. For some articles, a similarly titled article is imported instead of the one I have requested. This seems to be a problem when the article I want isn't included in Google Scholar. This doesn't happen frequently but even if it happens only every now and then; having the wrong metadata attached to a pdf can be frustrating. For articles without DOIs I always bring in metadata one article at a time and check them carefully.
For example: grabbing metadata from the following article:
A study of homicidal deaths by mechanical injuries in Surat, Gujarat. J Indian Acad. Forensic Med. 2010; 32(2):134-138.
http://medind.nic.in/jal/t10/i2/jalt10i2p134.pdf
will instead bring metadata for the article:
Pattern of head injury in homicidal deaths. Indian J Forensic Med. Technol. 2009; 3(2)18-21.
This is the wrong article, wrong authors, wrong year, and the wrong journal. Sometimes the article error is much less obvious.
About LyZ I have tried to get it to work on three different systems but only get error messages of the type
SERVER ERROR:
[Exception... "Component returned failure code: 0x80520001 (NS_ERROR_FILE_UNRECOGNIZED_PATH) [nsILocalFile.initWithPath]" nsresult: "0x80520001 (NS_ERROR_FILE_UNRECOGNIZED_PATH)" location: "JS frame :: chrome://lyz/content/lyz.js :: anonymous :: line 85" data: no]
and
Could not contact server at: \\.\pipe\lyxpipe
To judge from comments from more or less satisfied LyZ users on the net it seems that it should indeed be possible to run LyZ from inside zotero. Wonder what tricks these users have used to get it work...?
You're probably best off making LyZ-related inquiries at the LyZ launchpad page.
So I guess for the foreseeable future you'll just have to deal.
1. Too liberal a use of {} which limits the effects of bibstyle, possibly producing wrong
results in terms of lower/upper case in TITLE.
2. No way to use @§TRING constructs and journal abbreviations.
3. No way of modifying the key to include, e.g., the .bib file name as an element.
4. No way of changing the layout of the bib entries themselves ("" vs {}, lower vs upper case, etc)
(1) I believe that {} is only triggered when an uppercase letter appears in a non-initial position in a word (though there have been examples where we should probably break on punctuation in addition to spaces). Do you have other specific examples of it being over-zealous?
(2) No program I know handles this well. Are you aware of any? This might be possible when/if there is hierarchical support, so that journal names can be better normalized.
(3) I don't think I've ever seen that notation. Examples?
(4) How does this impact you? Zotero data should be stored in title-case as much as possible & your .bst should handle case changes (as you said in your first point). There are various tools to perform these sorts of transformations, though.
Finding a journal name is hard. You can look in the text for journal names already in your zotero database, but you'd have to avoid having the bibliography provide false positives & I have no idea how you'd identify some journal names that are likely very common (like 'Nature' or 'Science'). Using heuristics to find vol/issue would also be hard (page & year seem easier, but not without issues).
Zotero has reasonable success by querying google scholar with phrases in the pdf if a unique identifier (doi) is not found. Going back to the publisher after such a process might make for better data, but would not help with false identification.