User talk:Diberri/Template filler

This is a subpage of Diberri's talk page, where you can send him messages and comments.

Put new text under old text. Click here to start a new topic.
New to Wikipedia? Welcome! Learn to edit; get help.

Bug reports

Ideally, please report bugs using this link. If you're not comfortable with that, please report the bug below. The more details, the better.

Missing titles

Resolved

In cases where there is no title retrieved for a URL (like this), could the title field be populated with the domain name instead of being left blank? --Arcadian (talk) 15:19, 5 November 2008 (UTC)[reply]

Great idea. Should be fixed now. See [1]. --David Iberri (talk) 02:40, 10 November 2008 (UTC)[reply]

Thanks! --Arcadian (talk) 03:32, 10 November 2008 (UTC)[reply]

Vertical pipes in page titles

Resolved

BBC articles like http://news.bbc.co.uk/2/hi/business/7732733.stm contain vertical bars in the data that will end up in the title field, terminating it early. So something like $title =~ s/|/|/g; to quote. RDBrown (talk) 13:49, 17 November 2008 (UTC)[reply]

Thanks for the report. I've added it to the ticket tracker as ticket #41005. I've uploaded a patched version. See [2] for an example. --David Iberri (talk) 01:29, 18 November 2008 (UTC)[reply]

PMCID/PMID

A PMC doesn't imply that a PMID exists (ie PMC 1247673), also older PMC PDF image articles - a more specific error in that case would help. RDBrown (talk) 13:49, 17 November 2008 (UTC)[reply]

I'm not sure I follow. Can you explain a bit more? --David Iberri (talk) 01:29, 18 November 2008 (UTC)[reply]

Using 1247673 in a PubMedCentral search give (me at least) "Internal Server Error" which is hard to distinguish from an other problem. Indicating that no PMID can be found so the PMC can't (yet) be formatted would help.

Similarly when an ISBN can't be found. -- RDBrown (talk) 12:53, 18 November 2008 (UTC)[reply]

Ah, I see. This has been filed as ticket 41053. I've mostly fixed the PubMed Central issue, but the error messages aren't quite as user-friendly as I'd like (see [3] for example). Do you have an example of an ISBN the template filler isn't grokking? --David Iberri (talk) 03:23, 19 November 2008 (UTC)[reply]

Journal abbreviations

The journal titles are back to a terser form (applause), but they sometimes don't match the Pubmed abbreviation, if that was your aim (ie 11319089 |journal=Appl. Environ. Microbiol. Pubmed: Appl Environ Microbiol.) This seems to depend on the Journal, but Pubmed seems to suppress internal periods for abbreviation. RDBrown (talk) 13:49, 17 November 2008 (UTC)[reply]

PubMed offers several journal titles and abbreviations. Arbitrarily, the template filler prefers the ISO abbreviation ("Appl. Environ. Microbiol."), followed by the Medline abbreviation ("Appl Environ Microbiol"), and then, if all else fails, the full title ("Applied and environmental microbiology"). Since the tool's been using the ISO abbreviation, and there doesn't appear to be a standard/preference on Wikipedia, I'd just as soon continue to use the ISO abbreviation. Do you have a compelling reason to switch to the Medline abbreviation? --David Iberri (talk) 01:29, 18 November 2008 (UTC)[reply]

No compelling reasons, but I had the impression that earlier versions of the filler (say 3 months ago) used the Medline form rather than the ISO form, though it looks like a number of journals give the terser Medline form for the ISO form if that's what the filler has been requesting. For articles with a lot of refs (I've seen 200 in a few), brevity may help. -- RDBrown (talk) 12:53, 18 November 2008 (UTC)[reply]

I see, but still I'm unsure whether this necessitates a change from ISO to Medline titles. There's actually some discussion somewhere on Wikipedia (possibly on my talk page, I'm not sure) about having the template filler return full journal titles instead of either abbreviation type (WP:PAPER and all that). I can see that argument, but as for switching between types of journal title abbreviations, I'm thinking there's probably too much inertia to motivate a change. --David Iberri (talk) 03:27, 19 November 2008 (UTC)[reply]

Partial PMID data

At least twice tonight, I've had PMID data return populating only the URL field, no others. Clicking Submit again, once more than once fully populated the cite journal fields. I have seen this before infrequently but I can't identify which PMID values gave problems from my browser history. If easy could you please check that the title and journal fields have data and go around again (once) if they don't. -- RDBrown (talk) 12:53, 18 November 2008 (UTC)[reply]

This appears to be a load issue with the NCBI/PubMed servers. I've been in contact with some of the folks there, and they tell me they're unable to replicate the problem. However, it invariably results when the load level is high (more than a few requests per second during EST business hours). I've thought about the solution you mention, but I'm afraid that would only exacerbate the load problem. --David Iberri (talk) 03:37, 19 November 2008 (UTC)[reply]

Feature requests

Would you like to see the template filler do a new trick? Post it here. The more details, the better.

Citation format of the open access journals of Plos

How about a setup that gives back alternatively the template in the way that Plos journals provide their citations? Would be neat to try improving readability in citations. I think, they did a good job in that. Bollisee (German Wikipedia) —Preceding unsigned comment added by 217.227.237.248 (talk) 19:56, 16 November 2008 (UTC)[reply]

I'm not terribly familiar with PLoS, and I'm certainly not familiar with their citation style. Can you provide some additional details and/or links? Skimming the website didn't turn up much. Thanks, David Iberri (talk) 20:19, 16 November 2008 (UTC)[reply]

Would this help, formatting instruction and an example?

Author_Surname Author_Initial, further_Autors (Year) Titel. Journal. Number: First_PageMinus_SignLast_Page. DOI_or_PMID_or_Onlinelink.
Lee KB, Liu CT, Anzai Y, Kim H, Aono T, Oyaizu H (2005) The hierarchical system of the 'Alphaproteobacteria': description of Hyphomonadaceae fam. nov., Xanthobacteraceae fam. nov. and Erythrobacteraceae fam. nov. Int J Syst Evol Microbiol. 55: 1907−19. PMID 16166687.

And I almost forgot, it would be nice to have no truncation of the authors list. The last author is in newer publications in natural sciences usually the most important one. Bollisee (German Wikipedia)

Really? In most publications I'm aware of, the last author list is the PhD in whose lab the study was performed, and first author is usually the postdoc who wrote the majority of the paper, did most of the experiments, etc. --David Iberri (talk) 00:39, 25 November 2008 (UTC)[reply]

Patents

Hello again - I realise this is a bit out of the scope of your template filler wizard, but I was wondering if you would be interested in adding functionality to build the {{US patent reference}} given only a patent number? Regards—G716 <^T·_C> 05:06, 22 November 2008 (UTC)[reply]

There is some similarity. {{ref patent}} is likely the best case in point. Given a country and patent number it will link to the full patent data, which could then in principle be autoextracted to populate the several other fields, including the required "title=". I'm sure this would be a popular tool amongst patent fans.LeadSongDog (talk) 06:22, 22 November 2008 (UTC)[reply]

This would make a great addition to the template filler. As I don't use work on patent-related articles, it'd be helpful to know how one typically goes about obtaining patent information. I understand esp@cenet and uspto.gov are useful; are there others? In particular, I'm looking for a database that allows exporting; esp@cenet's CSV exporting might suffice, but XML would be much preferable. Any suggestions? --David Iberri (talk) 18:24, 22 November 2008 (UTC)[reply]

Have a look at the EPO portal. It can generate downloadable XML results, though I'm not sure just how to automate the query.LeadSongDog (talk) 20:51, 22 November 2008 (UTC)[reply]

Looks promising. Thanks for the link. --David Iberri (talk) 00:39, 25 November 2008 (UTC)[reply]

{{Chembox new}} → {{Chembox}}

Resolved

Heya

Could you code the following change into your pubchemid tool? {{Chembox new}} has been deprecated in favor of {{Chembox}}~~; the former now exists as a redirect to the latter~~. Thanks! --Rifleman 82 (talk) 11:37, 23 January 2009 (UTC)[reply]

I haven't been keeping up with these discussions. Is {{chembox}} completely backwards compatible with {{chembox new}}? If so, I'll happily update the template filler to use the former rather than the latter. Cheers, David Iberri (talk) 20:23, 25 January 2009 (UTC)[reply]

I did a run with User:Chem-awb to convert everything over to Chembox, but Physchim62 tells me that we still need a history. No article should be using Chembox new any more. --Rifleman 82 (talk) 03:35, 26 January 2009 (UTC)[reply]

Got it. I guess my real question is regarding what change(s) I need to make; is it as simple as changing "chembox new" to "chembox"? Or are some of the parameters a bit different as well? --David Iberri (talk) 23:45, 26 January 2009 (UTC)[reply]

Sorry for the late reply, yes, just change "chembox new" to "chembox"; everything else will work. --Rifleman 82 (talk) 12:54, 28 January 2009 (UTC)[reply]

Great. Should be all fixed now, for example [4]. --David Iberri (talk) 01:24, 30 January 2009 (UTC)[reply]

Great, thanks! --Rifleman 82 (talk) 03:45, 30 January 2009 (UTC)[reply]

IUPAC names

Resolved

Also, the IUPAC names from PubChem are not reliable; could your template filler not fill that field? Thanks! --Rifleman 82 (talk) 11:54, 23 January 2009 (UTC)[reply]

That's a shame. If there is a general consensus that PubChem's IUPAC names are unreliable, I'll gladly remove them from the template filler. So far, you're the first one to mention it to me; are there others who have similar experiences? Are you aware of any other sources of IUPAC names? That way, we can remove the PubChem IUPAC name and replace it with a more reliable source. Cheers, David Iberri (talk) 20:23, 25 January 2009 (UTC)[reply]

There might be something at [5] or [6] or [7] that helps. Have a boo.LeadSongDog (talk) 02:15, 26 January 2009 (UTC)[reply]

Since Rifleman was the first, let me be the second :)

IUPAC names from PubChem are often "unreliable", as are the structures themselves, though much less frequently. I usually generate names manually with ACD/ChemSketch. I've found that ChemSpider consistently provides reliable IUPAC names. Fvasconcellos (t·c) 20:51, 26 January 2009 (UTC)[reply]

Hey David, would you mind removing the IUPAC field? Thanks. --Rifleman 82 (talk) 11:20, 19 February 2009 (UTC)[reply]

No problem. Should be removed now. Cheers, David Iberri (talk) 20:56, 20 February 2009 (UTC)[reply]

That's great, thanks! --Rifleman 82 (talk) 05:39, 21 February 2009 (UTC)[reply]

Older books

From a URL like http://books.google.com/books?id=ZLtXAAAAMAAJ, could you take "ZLtXAAAAMAAJ" as a parameter, and populate the cite book template (including the "url" paramter)? --Arcadian (talk) 23:52, 9 March 2009 (UTC)[reply]

Seems reasonable enough, but there doesn't appear to be a way to do this that's currently documented in the Google Book Search API. I've posted a question to their API help forum^[8] and will keep you posted. --David Iberri (talk) 03:03, 10 March 2009 (UTC)[reply]

As I suspected, it seems there's no way to fetch book information using just a Google ID like that. Bummer. I've left another post with the Google Book Search folks, essentially saying Really? Are you sure? and I'll let you know if I hear anything back. :-) --David Iberri (talk) 22:51, 16 March 2009 (UTC)[reply]

Not sure about the API, but on the interactive format there's an "Other editions" section shown. The "more" at the bottom of that section has an OCLC number embedded in the associated URL. Perhaps the API has something similar?LeadSongDog (talk) 05:13, 17 March 2009 (UTC)[reply]

And even if it isn't possible to do that (or if it would take some time to determine the best way to use the OCLC data), could you populate the cite book title field from the HTML title field, as is done in when using the template filler in URL mode? Even if this new feature wouldn't seem to provide much additional functionality right now, it would encourage editors to use cite-book instead of cite-web, and the increased structuring and data granularity of the citation information could lead to long-term benefits. --Arcadian (talk) 18:32, 17 March 2009 (UTC)[reply]

Another tool doing this is now available at http://reftag.appspot.com/ . Would it be possible to integrate this functionality into your template filler? --Arcadian (talk) 13:06, 10 April 2010 (UTC)[reply]

DOIs and URLs

(Copied from User talk:Diberri#DOIs and URLs.)
Hi, I've noticed that your very effective, widely-used template builder does not automatically pull up DOIs. It also seems to automatically pull up URLs regardless of whether they are freely-accessible or not.

I think it's important for Wikipedia to start telegraphing to readers whether the referenced articles are freely-accessible or not. As you've probably noticed, PubMed now does this. I started a discussion on this over at the citation template talk page and Orangemarlin suggested that urls be excluded when the article isn't freely available. A replacement for URLs on these articles would be DOIs, which will always redirect to the for-pay URL. Have you considered automatically adding DOIs? II | (t - c) 22:02, 16 March 2009 (UTC)[reply]

Actually, the tool pulls up both DOIs and URLs. Here is an example of it populating the DOI field. It will also pull up the URL if you ask it to; for example, here. It's definitely possible that PubMed may be offering a DOI but the template filler isn't picking up on it, however. I'd have to investigate that on an article-by-article basis. As for determining whether a given DOI or URL points to a free full-text article, I haven't seen anything in PubMed's eUtils documentation [9] to suggest that this is practical. (Technically, it's possible to do this, but it requires an extra call to their databases, and that may well exhaust our usage rights -- we're limited to a few requests per second. Even if it wouldn't push us over our quota, this method would entail a clunky piece of code that I'm not terrible comfortable with. I suppose the bottom line, at least from my vantage point, is that it's not feasible to use PubMed's eUtils interface to determine whether an article's full text is provided for free. But there may be other methods that don't use the eUtils interface.) --David Iberri (talk) 22:40, 16 March 2009 (UTC)[reply]

I guess I'd just used it for something without a DOI. Also, my wording wasn't all that great, but my second sentence said your tool automatically pull up URLs. Anyway, hopefully we can agree that telegraphing which articles are freely-available is a good goal, and if you have any ideas on how to make it more common, I'd love to hear them. II | (t - c) 22:53, 18 March 2009 (UTC)[reply]

Yes, we definitely agree that free full-text is better than a URL that points to a sign-in page. The problem is that there's no documented way to do this, making me a bit hesitant. I'll certainly keep investigating ways of accomplishing this. --David Iberri (talk) 02:00, 19 March 2009 (UTC)[reply]

Full Journal Titles

Resolved

Hi Diberri. How difficult and/or controversial would it be to change the cite journal template to display full journal names rather than the abbreviated version? Personally I prefer the full journal name and this seems to be more in line with wikipedia guidelines, was discussing this issue with Boghog2 here and he suggested I talk to you about it...Meodipt (talk) 06:19, 15 April 2009 (UTC)[reply]

Should probably be discussed at WT:MEDMOS first. LeadSongDog come howl 07:34, 15 April 2009 (UTC)[reply]

I also support adding at least the option of returning full journal names. I also wanted to point out as mentioned here that the use of full journal names is already recommended by WP:Scientific_citation_guidelines#Annotations. Boghog2 (talk) 17:12, 15 April 2009 (UTC)[reply]

(Outdent)
Thanks to all of you for bringing up this discussion again. My recollection was that there was no consensus about abbreviated or full-length journal titles, but I'm happy to hear that this has changed (or alternatively, that my recollection was wrong). I've started some work on adding an option for full-length titles, and in so doing, I've noticed that some of the journal titles supplied by PubMed are in what I consider incorrect case (eg, "The Journal of pharmacy and pharmacology" instead of "The Journal of Pharmacy and Pharmacology"). And I haven't encountered an easy way of translating these titles into the canonical ones. I'm sure there's a solution to this, but I haven't yet found it (nor have I spent much time looking for one). In the meantime, I can add an option to provide full-length journal titles, with the knowledge that there will be a risk that the title is inappropriately cased, and that users of the option will have to manually change the case of the journal title if appropriate. --David Iberri (talk) 15:54, 16 April 2009 (UTC)[reply]

Thanks for your positive response! Concerning capitalization of journal names, Lingua::EN::Titlecase would seem to provide a straight forward perl solution. For example:

#!/opt/local/bin/perl5.10

use Lingua::EN::Titlecase;

$journal = "Journal of medicinal chemistry"; $tc = Lingua::EN::Titlecase->new($journal); print "$journal -> $tc\n";

produces:

Journal of medicinal chemistry -> Journal of Medicinal Chemistry

Cheers. Boghog2 (talk) 17:48, 16 April 2009 (UTC)[reply]

It might also be worth revisiting Vancouver system and the List of Serials Indexed for Online Users : Terms and Conditions. page at the NLM. Cheers.LeadSongDog come howl 19:24, 16 April 2009 (UTC)[reply]

(Outdent)
I thought about L::EN::Titlecase, but wasn't sure whether it generalized well enough for journal titles. But you're right; it probably does, and at the very least, it'll provide a more sensibly cased version of journal names than that occasionally reported by PubMed. I'll add it tonight. --David Iberri (talk) 00:33, 17 April 2009 (UTC)[reply]

Thanks for adding the optional "Use full journal title" It works great! Cheers. Boghog2 (talk) 15:35, 17 April 2009 (UTC)[reply]

Thank you! :-) Meodipt (talk) 21:42, 17 April 2009 (UTC)[reply]

Tool down or PubMed itself ?

http://toolserver.org/~diberri/cgi-bin/templatefiller/index.cgi?ddb=&type=pubmed_id&id=8601117&add_ref_tag=1 gives after protracted thinking <ref name="pmid">{{cite journal |author= |title= |journal= |volume= |issue= |pages= |year= |pmid= |pmc=2350106 |doi= |url=}}</ref>

Trying http://www.ncbi.nlm.nih.gov/pubmed/8601117 gives "backend-exception Exception from Backend: bePfetch (%23PmXmlSrv): Error+111+(Connection+refused) " report from PubMed

Whilst http://www.ncbi.nlm.nih.gov/pubmed/8601117?ordinalpos=39 manages to locate and display abstract. Seems PubMed having a very bad day. David Ruben ^Talk 02:04, 4 June 2009 (UTC)[reply]

The two PubMed links you provided are working fine for me, not surprising given this is some 17 hours after your initial post here. I'm still having some trouble loading the URL to the template filler [10], however. Even the basic form itself [11], which does no querying, is taking a while to load. I suspect the initial problems you noticed were due to an internal PubMed error, and that the problems with lag I'm experiencing are unrelated. Hopefully the toolserver will pick up the pace soon. Thanks for the heads up. --David Iberri (talk) 19:30, 4 June 2009 (UTC)[reply]

PubMed links back working, toolserver bradyphrenic but gets there eventually :-) David Ruben ^Talk 20:39, 4 June 2009 (UTC)[reply]

Nothing's changed with the template filler code, so I'd bet there's just general sluggishness with the toolserver. We'll probably just have to wait this one out. --David Iberri (talk) 00:29, 5 June 2009 (UTC)[reply]

Tool not working at all

When I visit http://toolserver.org/~diberri/cgi-bin/templatefiller/ and enter "19697444" (a valid PubMed ID), it responds:

Error:
Package for 'pubmed_id' source could not be loaded (tried WWW::Wikipedia::TemplateFiller::Source::pubmed_id WWW::Wikipedia::TemplateFiller::Source::PubmedId WWW::Wikipedia::TemplateFiller::Source::PUBMED_ID WWW::Wikipedia::TemplateFiller::Source::Pubmed_id). Error: Can't locate WWW/Wikipedia/TemplateFiller/Source/Pubmed_id.pm in @INC (@INC contains: /home/diberri/lib/lib/site_perl/5.10 /home/diberri/lib/perl5 /home/diberri/lib/share/perl/5.10.0 /home/diberri/lib/lib/perl/5.10.0 /opt/ts/perl/5.10/lib/5.10 /opt/ts/perl/5.10/lib/site_perl/5.10 /opt/ts/perl/5.10/lib/vendor_perl/5.10 .) at (eval 19) line 1. BEGIN failed--compilation aborted at (eval 19) line 1. at /home/diberri/lib/share/perl/5.10.0/WWW/Wikipedia/TemplateFiller/WebApp.pm line 83

It doesn't matter which PMID it uses; I get a similar message. Eubulides (talk) 23:30, 24 August 2009 (UTC)[reply]

I get an "Error 500" from the server. I really like your excellent tool and use it often. Pleeease make it work again!--84.137.80.19 (talk) 10:25, 29 September 2009 (UTC)[reply]

update on template filler?

it's been more than a month since the template filler stopped working. how are things going with getting it back up and running? —Chris Capoccia ^T⁄_C 05:11, 29 September 2009 (UTC)[reply]

Not done NOT WORKING!!!! ► RATEL ◄ 23:13, 6 October 2009 (UTC)[reply]

Here's an alternative: http://toolserver.org/~holek/cite-gen/index.php --Arcadian (talk) 02:13, 7 October 2009 (UTC)[reply]

Done WORKING!!!

Thank you! Very much appreciated!! Boghog (talk) 14:15, 11 October 2009 (UTC)[reply]

It's still not working on the toolserver, unfortunately. Right now the tool is served from a separate server. --David Iberri (talk) 12:06, 12 October 2009 (UTC)[reply]

diberri.dyndns.org not reachable

Hello, the server isn't available, for several hours now, and possibly days, since I'm just back from a week-long tour without using it. Just FYI. ----Ayacop (talk) 18:52, 24 November 2009 (UTC)[reply]

I haven't had any trouble with it, and it appears to be working now. [12] --David Iberri (talk) 00:22, 25 November 2009 (UTC)[reply]

Indeed. Today no problems. I'll observe this. Thanks. --Ayacop (talk) 11:32, 25 November 2009 (UTC)[reply]

Google Books

Hello. I just wanted to let you know that I use Google Books to source my articles quite a bit and an option for it in your template filler would be wonderful if you're up for it. el3ctr0nika (Talk | Contribs) 10:29, 15 January 2010 (UTC)[reply]

In the case of recent books, that can be done via the ISBN option, however an LCCN or OCLC option would be a nice addition to cover books that predate the use of ISBNs (about 1970 iirc). Caveat: an OCLC number often encompasses multiple editions of the same book or serial, so assigning a date would have to be manual when that applies. Citation bot can handle OCLC, but is presently blocked for other reasons. LeadSongDog come howl 15:31, 15 January 2010 (UTC)[reply]

Bug common to Citation bot

Hi, David. At User:Citation_bot/bugs#Incorrect splitting of author into other there's an apparent bug identified common to both Citation bot and Template filler. When parsing PMID 17967920 for example, both tools seem to get the personal authors correct but omit the collective author. Any idea about the cause? LeadSongDog come howl 04:07, 3 February 2010 (UTC)[reply]

Thanks for the report. I've replied over there. --David Iberri (talk) 00:24, 6 February 2010 (UTC)[reply]

Translated titles and brackets and doublequotes, oh my!

Please see Wikipedia_talk:WikiProject_Medicine#Ref_formatting_on_non-English_journals for a report on malformed anglicized titles. It's basically a template bug, but has a bearing on the filler too. It's an opportunity to improve the filler. LeadSongDog come howl 14:10, 28 April 2010 (UTC)[reply]

Filler down?

For a few days now I have been unable to connect to the template filler. Is this (extremely useful) tool down? Hope this can be corrected easily, because I have become very dependent on it! --Crusio (talk) 16:58, 2 October 2010 (UTC)[reply]

Ditto. LeadSongDog come howl! 18:12, 5 October 2010 (UTC)[reply]

IT'S BACK!!! THANKS!!! --Crusio (talk) 13:35, 10 October 2010 (UTC)[reply]

Donate?

I clicked on the "donate" button here but nothing happened. I looked to see if I could email you, but I don't see that you have email enabled. If you take Paypal, I would like to make a very modest donation. (Fair warning, if we figure out how to get a donation to you, it will be followed by a request.)--SPhilbrick T 13:22, 10 October 2010 (UTC)[reply]

Apostrophe encoding error on template filler XML format

Hi, I recently came across your template filler tool. Looks great, I'll hopefully be able to use it for serveral thousand edits to add missing PMIDs etc. I have noticed a minor issue with the XML format of the output e.g. this one in that the apostrophes are incorrectly escaped somehow. Thanks Rjwilmsi 22:56, 7 December 2010 (UTC)[reply]

Ignore PubMed dummy DOI

Hi, as reported to the Citation Bot, please ignore the dummy DOI that's crept into some PubMed articles, example. Thanks Rjwilmsi 22:08, 9 December 2010 (UTC)[reply]

If checking DOIs, you may wish to warn if they appear to be doubled

{{cite journal |author=Adamantidis AR, Zhang F, Aravanis AM, Deisseroth K, de Lecea L |title=Neural substrates of awakening probed with optogenetic control of hypocretin neurons |journal=Nature |volume=450 |issue=7168 |pages=420–4 |year=2007 |month=November |pmid=17943086 |doi=10.1038/nature0631010.1038/nature06310}}

$len_doi = length($doi);

   if ($len_doi && !($len_doi & 1) && substr($doi, 0, $len_doi >> 1) eq substr($doi, $len_doi >> 1)) {

	printf "DOI doubled %s %s\n", substr($doi, $len_doi >> 1), $doi;

The version just generated is fine, but shows another peeve, the URL just repeats the DOI with the DOI resolution prefix.

{{cite journal |author=Adamantidis AR, Zhang F, Aravanis AM, Deisseroth K, de Lecea L |title=Neural substrates of awakening probed with optogenetic control of hypocretin neurons |journal=Nature |volume=450 |issue=7168 |pages=420–4 |year=2007 |month=November |pmid=17943086 |doi=10.1038/nature06310 |url=http://dx.doi.org/10.1038/nature06310}}

RDBrown (talk) 22:58, 10 December 2010 (UTC)[reply]

Lay summary

Could the "extended fields" be extended to include the |laysummary= |laydate= and |laysource= fields? WhatamIdoing (talk) 06:49, 18 January 2011 (UTC)[reply]

Reftoolbar autofill

Hi David, auto template filling for PMID/DOI/ISBN has recently been added to the Reftoolbar, integrating really well into the editing interface. However it doesn't seem to be as accurate or complete as your tool. Perhaps you could work together with its author to get it to be as good as yours? --WS (talk) 22:20, 18 January 2011 (UTC)[reply]

Incorrect DOI returned

For PMID 12122039 an incorrect DOI of 20026621 is returned. Would it be possible to add more validation to the DOIs, maybe that they start with 10. or contain a slash? Thanks Rjwilmsi 19:26, 7 February 2011 (UTC)[reply]

Another example is PMID 12023862 where a DOI of 10.1042/ is returned. Rjwilmsi 22:21, 11 February 2011 (UTC)[reply]

Tool down. No DNS entry?

Firefox can't find the server at diberri.dyndns.org.
nslookup gives
> server ns5.dyndns.org.
Default server: ns5.dyndns.org.
Address: 203.62.195.75#53
> diberri.dyndns.org.
Server:		ns5.dyndns.org.
Address:	203.62.195.75#53

** server can't find diberri.dyndns.org.: NXDOMAIN

Is it possible to host at toolserver.org? I vaguely remember you having tried before. RDBrown (talk) 23:40, 8 May 2011 (UTC)[reply]

Still down. I've gotten lazy about citations for medical articles. May have to do this by hand. :( OrangeMarlin ^{Talk• Contributions} 01:13, 14 May 2011 (UTC)[reply]

{{cite pmid|}} or {{cite doi|}} will work, and the tool Arcadian mentioned. My scripts use the templatefiller though. RDBrown (talk) 05:55, 14 May 2011 (UTC)[reply]

I've used the cite pmid, but when I clicked on the link to speed it up, it seemed to be disabled too. I'll check to see if one of my edits from earlier today got filled in. OrangeMarlin ^{Talk• Contributions} 06:33, 14 May 2011 (UTC)[reply]

Nope, cite pmid is also not working right. Bummer. OrangeMarlin ^{Talk• Contributions} 06:38, 14 May 2011 (UTC)[reply]

cite pmid should be an easier fix, [updating a toolserver account] by the sound. I think David's tool requires either paying for the DNS entry or porting to toolserver again, though by IP address might work. In the meantime, I've been leaving PMID droppings to fix later. The perl source is available if a LAMP stack holds no fears — I should fix my script to call the PMID perl directly. RDBrown (talk) 07:50, 14 May 2011 (UTC)[reply]

Wow. You wrote way over my head. I'm at the level of typing in the cite template, and hope that everything behind it just works. :) OrangeMarlin ^{Talk• Contributions} 19:05, 14 May 2011 (UTC)[reply]

Can't say I understand anything of the above either, but am sorely missing this wonderful tool!! --Crusio (talk) 07:26, 24 May 2011 (UTC)[reply]

{{cite pmid|}} or {{cite doi|}} are working now (The Wikipedia citation bot fills them in — I think).

I think the tool Arcardian mentioned is the Universal reference formatter which takes DOIs and PMIDs. RDBrown (talk) 11:43, 24 May 2011 (UTC)[reply]

Missing you. ♥ --Anthonyhcole (talk) 04:28, 25 May 2011 (UTC)[reply]

Tragedy! So tremendously sorry for letting this slip. For some unholy reason, DynDNS.org has forgotten about my account and the former diberri.dyndns.org domain no longer works. I've attempted to reestablish my account to no avail, and in the meantime have established a secondary dynamic DNS domain over at http://diberri.crabdance.com. For now, the toolserver page (http://toolserver.org/~diberri/cgi-bin/templatefiller) will redirect to the crabdance page (http://diberri.crabdance.com/cgi-bin/templatefiller). Hopefully I'll soon have my tools running directly on the toolserver again to avoid this majority of this hassle in the future. Again, my apologies for this hiccup. --David Iberri (talk) 06:33, 2 June 2011 (UTC)[reply]

Page labeled dangerous

Trend Smart Surfing labels http://diberri.crabdance.com/cgi-bin/templatefiller as dangerous. Andreas ^(T) 16:42, 25 February 2012 (UTC)[reply]

diberri.crabdance.com.: NXDOMAIN

DNS entry gone again. What was the problem with the toolserver installation, if someone here can help? RDBrown (talk) 00:15, 22 September 2012 (UTC)[reply]

Yup, it's dead for me too. I think he doesn't watch over it much any more unfortunately, since it was volunteer activity. I wish it was open source. I could host it, for example. Oh well. SkepticalRaptor (talk) 21:22, 27 September 2012 (UTC)[reply]

It is opensource, it's up on CPAN. RDBrown (talk) 00:25, 28 September 2012 (UTC)[reply]

Shame, because its really useful. If someone manages to host it successfully elsewhere, please post here

Thanks for these messages. The free DNS service disabled my account for unclear reasons without notice. I've restored the account today. I'll work on getting the tool back on toolserver.org. --David Iberri (talk) 17:43, 4 October 2012 (UTC)[reply]

Thank you David. The perils of being useful, I suppose. :-) LeadSongDog come howl! 17:53, 4 October 2012 (UTC)[reply]

Back up! Yeah!!! Much appreciated. Boghog (talk) 18:49, 4 October 2012 (UTC)[reply]

Opps. Slightly premature. Site is back up, but it cannot yet connect to PubMed. Boghog (talk) 19:28, 4 October 2012 (UTC)[reply]

Bleh, what a mess. I'm on call today but will figure this out when I get home. --David Iberri (talk) 22:39, 4 October 2012 (UTC)[reply]

Ok team, it's back up; the tool is still over at http://diberri.crabdance.com/cgi-bin/templatefiller/. I can't manage to get it setup on toolserver.org because that website is down. The problem was that for some ungodly reason, the CPAN library XML::LibXML stopped accepting the doctype tag returned in PubMed's XML. Ugh. Thanks for all the messages and the kind donations -- it's really my goal to keep the template filler living for as long as I can and as long as it's useful. --David Iberri (talk) 06:07, 5 October 2012 (UTC)[reply]

Thanks Dave! PubMed ID, URL, and ISBN searches now work, however PubMed Central ID searches do not (http error : unreachable network). The later is not critical since pmc normally cross references pmid. But I would appreciate ifyou would look into that if you have a chance. Boghog (talk) 06:33, 5 October 2012 (UTC)[reply]

Similar bug over there -- fixed it as well. Thanks for the tip. --David Iberri (talk) 07:20, 5 October 2012 (UTC)[reply]

Current source, Aim for Wikimedia Labs rather than toolserver?

The source on CPAN dates to 2009. Is the current source already available from somewhere like sourceforge or github? Or just a 7z or tar.gz on diberri.crabdance.com?

Wikimedia Tool Labs may work for hosting the template filler and it's source.

Rich Farmbrough might help if asked. RDBrown (talk) 10:42, 23 December 2012 (UTC)[reply]

Internal server error

When I tried this link, which I got from yUser:Diberri/Template_filler. Klortho (talk) 01:49, 22 May 2013 (UTC)[reply]

Ditto. I suspect that David's toolserver account has lapsed again. In the meanwhile, you could try out http://diberri.crabdance.com/cgi-bin/templatefiller/ or failing that https://toolserver.org/~holek/cite-gen/index.php

LeadSongDog come howl! 05:16, 22 May 2013 (UTC)[reply]

The http://diberri.crabdance.com is down :-( Doc James (talk · contribs · email) (if I write on your page reply on mine) 17:01, 9 September 2013 (UTC)[reply]

It is! OMG, what am I to do?!?! Please somebody do something... --Randykitty (talk) 16:58, 19 September 2013 (UTC)[reply]

No, account is not lapsed, the account home page still works. Klortho (talk) 16:07, 2 October 2013 (UTC)[reply]

The server was moved from http://toolserver.org/ to http://diberri.crabdance.com/. The later does appear to be down. Sigh. Boghog (talk) 16:20, 2 October 2013 (UTC)[reply]

Oh, that's interesting. I thought the move was in the other direction. Klortho (talk) 18:34, 2 October 2013 (UTC)[reply]

ISBN error

When using the tool to cite an ISBN, I have received this error several times over a period of months. "Error: There is no form named "frmconvert" at /data/project/citation-template-filling/perl5/lib/perl5/WWW/Mechanize.pm line 1930."

Should ISBN entries still work as they did when Diberri maintained this tool? Gzuufy (talk) 18:48, 30 June 2014 (UTC)[reply]

I just received the same error trying to use the tool. Ward20 (talk) 22:57, 2 May 2015 (UTC)[reply]

Code modification request: `|month=` deprecated in CS1 citations

This template filler is still generating cite templates that use the deprecated parameter |month=, which places articles into Category:Pages containing cite templates with deprecated parameters. A few of us have just finished getting thousands of month parameters merged with years to reduce the article count in this category, and I've been poking around looking for tools that editors might be using to create more of these errors.

Is there someone here who can adjust the code so that it merges the month and the year into |date= in order to avoid this error? Thanks.

P.S. For a bonus, the code could also be adjusted to add |accessdate= only if a URL is present. Adding |accessdate= if no URL is present causes a citation error. – Jonesey95 (talk) 15:48, 19 September 2014 (UTC)[reply]

There is no real error. Where was the consensus to deprecate the month parameter and why was it necessary in the first place? The only answer that I have received to date is that is no longer needed. It is equally valid to argue that it is not causing any problem. A better solution is just to omit the month entirely and just use the year parameter. Finally the accessdate parameter is not added by default. It is only added if specifically requested. Boghog (talk) 19:13, 19 September 2014 (UTC)[reply]

Omitting the month parameter from the tool would be fine as well. I do not know when the month parameter was deprecated, but it's been a while, and the now-hidden error messages will probably be unhidden in the next few months. The month parameter was useful when Harvard-style references worked only with |year=, but now that they can extract a year from |date=, it is no longer useful.

I understand that the accessdate parameter is not added by default, but its addition in the absence of the URL parameter results in an error and is invalid because accessdate refers to the date a URL was accessed. – Jonesey95 (talk) 16:19, 21 September 2014 (UTC)[reply]

Fixed @Jonesey95: I apologize for taking so long to fix this, but I finally got around to removing the |month= parameter. I am not very experienced with perl modules and I was afraid this would be complicated, but it actually turned out to be quite simple. Cheers. Boghog (talk) 21:38, 6 January 2015 (UTC)[reply]

Thanks for getting to it! I appreciate it. – Jonesey95 (talk) 22:04, 6 January 2015 (UTC)[reply]

Modifications

In order to take advantage of the new {{vcite2 journal}} template, I have changed the template filler tool output as follows:

{{cite journal}} → {{vcite2 journal}}
|author= → |vauthors=

Please note that {{vcite2 journal}} is identical to {{cite journal}} except that former will now output clean author metadata if the optional |vauthors= parameter is used. Boghog (talk) 21:54, 6 January 2015 (UTC)[reply]

Since {{cite journal}} now directly supports |vauthors=, the output of the template filler tool has been changed back to {{cite journal}} while retaining |vauthors= parameter. Boghog (talk) 16:40, 29 August 2015 (UTC)[reply]

Server Error

The webtool seems to be down. I am getting a "503 Service Temporarily unavailable". Can this tool be run locally? Julialturner (talk) 23:07, 18 February 2016 (UTC)[reply]

Back up and working. Julialturner (talk) 06:36, 19 February 2016 (UTC)[reply]

Error

For script developer. There have been errors for a few days. I tried PMID, PMC, ISBN. --122.24.105.38 (talk) 12:12, 18 June 2019 (UTC)[reply]

Tool not functioning

The template filling tool returns an error when a PMID is entered: "Error: Can't call method "findnodes" on an undefined value at /data/project/citation-template-filling/perl/ActivePerl-5.26/site/lib/WWW/Search/PubMedLite.pm line 117." Can this be fixed? Thanks, PeaBrainC (talk) 15:59, 1 July 2019 (UTC)[reply]

Errors using Drugbank ID (can't find module)

I get an error for every drugbank ID request. Here's what it looks like:

Error: can not load backend DrugBank (Can't locate WWW/Search/DrugBank.pm in @INC (you may need to install the WWW::Search::DrugBank module) (@INC contains: /data/project/citation-template-filling/perl/ActivePerl-5.26/site/lib /data/project/citation-template-filling/perl/ActivePerl-5.26/lib) at (eval 19) line 1. BEGIN failed--compilation aborted at (eval 19) line 1. ) at /data/project/citation-template-filling/perl/ActivePerl-5.26/site/lib/WWW/Wikipedia/TemplateFiller/Source.pm line 42.

I hope you'll be able to take a look and fix this. Thanks! proton donor^H⁺ 16:48, 2 April 2021 (UTC)[reply]

(talk page stalker) @Proton donor: (SMARTS: [!$([#6,H0,-,-2,-3)] A H-bond donor is a non-negatively charged heteroatom with at least one H ;-) Diberri wrote the tool, but I have been maintaining it. I have been able to migrate the tool to the new Toolforge server and get the pmid and pmc searches to work, but not the other searches such as DrugBank. The tool is written in Perl scripting language which I am not an expert. If anyone else can update the tool so that it can again search for DrugBank entries, I would be very happy to deploy it. Cheers. Boghog (talk) 20:51, 2 April 2021 (UTC)[reply]

Thanks! I'm no Perl hacker, but perhaps I can dig up a friend who is to take a look at it. Or I can learn to love the meditative art of filling out drugbox templates by hand ;) proton donor^H⁺ 19:29, 7 April 2021 (UTC)[reply]

Bug reports

Missing titles

Vertical pipes in page titles

PMCID/PMID

Journal abbreviations

Partial PMID data

Feature requests

Citation format of the open access journals of Plos

Patents

{{Chembox new}} → {{Chembox}}

IUPAC names

Older books

DOIs and URLs

Full Journal Titles

Tool down or PubMed itself ?

Tool not working at all

update on template filler?

diberri.dyndns.org not reachable

Google Books

Bug common to Citation bot

Translated titles and brackets and doublequotes, oh my!

Filler down?

Donate?

Apostrophe encoding error on template filler XML format

Ignore PubMed dummy DOI

Lay summary

Reftoolbar autofill

Incorrect DOI returned

Tool down. No DNS entry?

Page labeled dangerous

diberri.crabdance.com.: NXDOMAIN

Current source, Aim for Wikimedia Labs rather than toolserver?

Internal server error

ISBN error

Code modification request: |month= deprecated in CS1 citations

Modifications

Server Error

Error

Tool not functioning

Errors using Drugbank ID (can't find module)

Code modification request: `|month=` deprecated in CS1 citations