PDF search

giby

Hi,

I am looking for something relatively easy I think. I am going through several institution with several access but not the same. Sometime I add something in one place without access to PDF, but after I may be somewhere I could get them.

So I would like to get a feature for accessing all the doc without associated PDF and that cheak and if accessible download it…
Such a function could be either manually start, start at programed time, once a month, once a day, once every hour or been working every time as a background task.

If there already have something like that, just tell me

adamsmith

while we may get there eventually, no, that's not an easy function to add.

giby

It seems basic


for a given pub:
Simplest way
-open url
-add to zotero
-merge with previous

More sophisticate:
-open url
-cheak if PDF availlble

if availlable:
-add to zotero
-merge with previous

if not:
-skip

So in a full work:
-list pub without PDF
-go on first
-execute previous program
-go next if there is next
-come back 2 lines over
-if no next end/go to first (according to settings)

What's so hard????

adamsmith

Zotero is open source. If you think it's easy, patches are welcome.
(problems include but aren't limited to: GUI issues, identifying the right translator, using translators in a way they currently aren't coded for).

giby

I don't know how to program :(

mronkko

If you do not how to program, then how would you know that something is easy to do?

For example, the step "open url" is not as easy as it seems. What would it mean, anyway?

Option 1: "Open a tab, then load an ULR in that tab": When applied to multiple files, this would make firefox unresponsive.

Option 2: "Download the contents of an URL to a file / memory": Then what would you do with that raw HTML? Translators are design to work with loaded pages (DOM objects, as far as I know.)


It could be possible to code this as a plugin. After you have figured a way to get the PDF to the computer, then adapting Zotfile code to import the new file would be straightforward.

giby

So, If we take the option 1: as it is a procedure to look one at time it won't be unresponsive, and we could think about a close tab if it opened a new one each time. actually it could open it in curent tab, no?

option2: is that stupid to concieved to "open" a page without displaying it?

I should corect about coding, when kid, I learnt how to program in several language such as basic, logo and forth… do these languages still exist? anyway pretty useless :p

so My proposed way to import PDF (ok it looks a bad method) is
-add to zotero as new item
-merge that item with previous one (for differante way to fill a same field, choose the older one on the computer)

mronkko

Option 1: Yes, it could be current tab. If you want to do this for one item, you can already do so by right click -> View Online. Then download the file and use Zotfile to add it to the current item. This is three clicks and takes about the same that Zotero would take to do the same in the background.

Option 2: It is stupid to say that something is easy when you really have no idea if it is or is not. As far as I know, Firefox does not provide a way to construct DOM without displaying the page. I have worked with this problem and in the end found out that the easiest solution for me was to just open a large number of tabs. In this case the computer was dedicated for just this task, so unresponsiveness was not a problem.

To be useful, this would need to work also when multiple files are selected, ruling out Option 1. And Option 2 would be difficult to implement.


Also, there are other issues such as whether the URL is the same for the institution that has access to the PDF and the one that does not.

dstillman

If you want to do this for one item, you can already do so by right click -> View Online. Then download the file and use Zotfile to add it to the current item.

You don't need ZotFile. You can just drag a (direct) PDF link or its favicon onto the item in Zotero.

As far as I know, Firefox does not provide a way to construct DOM without displaying the page.

It does, via a hidden browser. We use it in translators. But that doesn't mean this is an easy problem.

giby

Ok, I think I You haven't understand what I wanted.

So I would like something that do it as a background task witout any intervention (exept turning on)

For the URL question, when I export to tex, I get these info:
@article{guimaraes_adaptive_2008,
title = {Adaptive Evolution of a {Lactose-Consuming} Saccharomyces Cerevisiae Recombinant},
volume = {74},
issn = {0099-2240, 1098-5336},
url = {http://aem.asm.org/content/74/6/1748},
doi = {10.1128/AEM.00186-08},
abstract = {The construction of Saccharomyces cerevisiae strains that ferment lactose has biotechnological interest, particularly for cheese whey fermentation. A flocculent lactose-
[…]
fermented threefold-concentrated cheese whey, providing an attractive alternative for the fermentation of lactose-based media.},
number = {6},
journal = {Applied and Environmental Microbiology},
author = {Guimarães, Pedro M. R and François, Jean and Parrou, Jean Luc and Teixeira, José A and Domingues, Lucília},
month = mar,
year = {2008},
pages = {1748--1756},
file = {Full Text PDF:/Users/benjaminleduc/Library/Application Support/Firefox/Profiles/0ek1w6cc.default/zotero/storage/7B3NXFHQ/Guimar√£es et al. - 2008 - Adaptive Evolution of a Lactose-Consuming Saccharo.pdf:application/pdf;Snapshot:/Users/benjaminleduc/Library/Application Support/Firefox/Profiles/0ek1w6cc.default/zotero/storage/HTJKHJP4/1748.html:text/html}
},

I get only one URL, It is that one to try…

For multiple files (the case that would need it) just take them one by one…

adamsmith

There is really no need to keep this going. As I say in the beginning "we might get there eventually" - retrieving PDFs (and other data) for items already in the library is something that's been planned for a long time (see towards the bottom here http://www.zotero.org/support/requested_features#zotero_interface ), but you'll just have to take our word for it that there are a pretty significant number of issues to be solved to implement that in a working and robust version.
mronkko is probably right that a quick and dirty version could be done faster as a plugin.

giby

I would like even a dirty plugin ;) If someone would like to help me with that?