Hi
I have used Zotero only for a few days. Have used EndNote for years, as my university has a site license. I was delighted at first trying out Zotero, but then soon very disappointed as I discovered none of the PDFs had been imported from the EndNote RIS.

I soon found the wiki page with a solution, doing a search and replace on the RIS file first:
http://www.zotero.org/support/kb/exporting_from_endnote_with_pdfs

It worked. However I recommended Zotero to a colleague having trouble with EndNote. She gave up as soon as she saw the above mentioned wiki page contained some strange code...

So tonight I took a look at the code in the translator. Getting it to recognize the EndNote specific 'internal-pdf' was of course simple, but I still had only the relative path to the PDF EndNote supplied. The Zotero import function needs the full path.

On export, by default EndNote puts the exported RIS file in the same folder as the endnote database, with the same name but another extension (at least on my system, OSX, EndNote X5).

If I only could get the path of the RIS file I could figure out the path to the Endnote PDF folder (of course supposing the RIS file had not been moved or renamed relative to the EndNote PDF folder).

I found no way to retrieve the path from within the translator. Unfortunately this forced me to make changes also to the Zotero client core (besides the translator). As
Zotero.Translate.IO.Read.prototype in translate_firefox.js hosted the read function I used that to also expose nslFile.path.

On my system it now reads:
Zotero.Translate.IO.Read.prototype = {
"__exposedProps__":{
"_getXML":"r",
"RDF":"r",
"read":"r",
"setCharacterSet":"r",
"filePath":"r"
},

...code omitted for clarity...

"filePath":function() {
return this.file.path;
}
}

If there is a way to get the above path without this change, that I have not found, no core change would be needed.

In the translator I have changed how the L1 tag is handled in processTag:

...
else if(tag == "UR" || tag == "L1" || tag == "L2" || tag == "L4") {
// URL
if(!item.url) {
item.url = value;
}
if(tag == "UR") {
item.attachments.push({url:value});
} else if(tag == "L1") {

var importFile = Zotero.filePath();
var endnotePdfPath = importFile.slice(0,importFile.lastIndexOf(".") + 1) + 'Data/PDF/';
Zotero.debug('debug: endnotePdfPath= ' + endnotePdfPath);
var filename = value.slice(value.lastIndexOf("/") + 1);

if(value.substr(0, 15) == 'internal-pdf://'){
value = value.replace('internal-pdf://','file://' + endnotePdfPath);
Zotero.debug('debug: attachment url value= ' + value);
}
item.attachments.push({url:value, mimeType:"application/pdf",
title:filename, downloadable:true});
} else if(tag == "L2") {
item.attachments.push({url:value, mimeType:"text/html",
title:"Full Text (HTML)", downloadable:true});
} else if(tag == "L4") {
item.attachments.push({url:value,
title:"Image", downloadable:true});
}
....

I can now import _with_ pdfs.

The downside is of course that the RIS file must be named _exactly_ like the library but with another extension (not .enl) and that it must be place in the same folder as the .enl file - or if moved that the '[libraryname].data/PDF/' folder gets moved along with it. However in most cases this will not be a problem as this naming and file location seems to be default behavior for EndNote? In any other case it should behave just as it does now, ignoring the attachments. And anyway, explaining the importance of the naming and the placing of the RIS file should be less intimidating to most users than the script code currently on the above mentioned wiki page.

As you can see I also made a small unrelated change: The attachent title is no longer the generic 'Full text (PDF)' but rather the proper file name. If more than one attached file I found identical non descriptive titles very annoying.

And also, I have only tested it on OSX in firefox!