it would be helpful to see a single entry with associated notes in your adjusted RDF format. We fixed a bug yesterday that would miss very long notes on import, but this sounds different.

Per aurimas' request I'll try to copy all the elements of a source and its notes into the next post (or two). I tried to put them into this post, but it's too long.

Per AdamSmith's request, umm, how would you suggest I about cutting down at 85,000+ line file to the minimum that triggers a bug in your software without guidance? Sounds like you've chased import bugs before, so some suggestions about what might trigger issues would be helpful. The best I can offer you without some help is to suggest that, like Scribe, Zotero doesn't like sources with more than 50 notes (since that's the only difference from the file that worked earlier today). I might be able to make better guesses if you could point me toward places you've seen problems in the past that are worth investigating.

Do you have a data generator for creating files with various complexities? Potentially I could write you one if you haven't done so yet, but in my past searching I've never found complete docs on the Zotero RDF format, and as best I can tell you pull from at least four name spaces, not two, which makes it a challenge to determine how Zotero handles the source files (not to mention that I know from past experience that your code doesn't validate against them so your import standard is wider than I could determine reading those standards anyway). I don't mind being part of the solution here, but some suggestions on where to look would be helpful.

There are several of these, that are near perfect matches. If you'd like, I can post them all, but the pattern continues and the post limit means it would require a long series of very similar posts. If there is some place to send a file, I'd be happy to do that instead.

<ns4:Memo rdf:about="item_973" ns5:itemType="note">
<rdf:value>[2] [B. Estimating Methodist Adherents]
"Historians conventionally multiply Methodist membership figures by between three and five to estimate adherents."
</rdf:value>
<dc:subject>member</dc:subject>
<dc:subject>none</dc:subject>
</ns4:Memo>
<ns4:Memo rdf:about="item_974" ns5:itemType="note">
<rdf:value>[4] [T. Methodist Geographic Distribution]
End of 19th Century, <10% of Methodists lived in British Isles, <75% lived in the United States. More African American Methodists than European Methodists.
</rdf:value>
<dc:subject>member</dc:subject>
<dc:subject>none</dc:subject>
</ns4:Memo>
<ns4:Memo rdf:about="item_976" ns5:itemType="note">
<rdf:value>[56] [T. Oral nature of Methodism]
Methodism was essentially oral, but has been reconstructed largely from written sources. See: the work of Leigh Eric Schmidt.
</rdf:value>
<dc:subject>hist</dc:subject>
<dc:subject>none</dc:subject>
</ns4:Memo>
<ns4:Memo rdf:about="item_977" ns5:itemType="note">
<rdf:value>[64] [T. E.P Thompson]
"Thompson's view was that the uncertainty principle in the salvation theology of Arminian Methodism combined with a psychic compulsion to adapt to new economic environments produced a manic spirituality driven by fear of backsliding and craven acquiescence in the work discipline of industrial capitalism."
</rdf:value>
<dc:subject>hist</dc:subject>
<dc:subject>none</dc:subject>
<dc:subject>theology</dc:subject>
</ns4:Memo>
<ns4:Memo rdf:about="item_978" ns5:itemType="note">
<rdf:value>[65-66] [Hannah Syng Bunting]
Counter example to E.P. Thompson's belief that Methodist deeply feared backsliding. Instead, she and other Methodists sought peace and purity.
Note: have a copy of her journal
</rdf:value>
<dc:subject>hist</dc:subject>
<dc:subject>none</dc:subject>
<dc:subject>sex</dc:subject>
<dc:subject>source</dc:subject>
</ns4:Memo>
<ns4:Memo rdf:about="item_979" ns5:itemType="note">
<rdf:value>[70] [T. Hymns]
Methodist hymns generally ignored doctrine and focused the Christian life as a pilgrimage.
</rdf:value>
<dc:subject>none</dc:subject>
<dc:subject>theology</dc:subject>
</ns4:Memo>
<ns4:Memo rdf:about="item_980" ns5:itemType="note">
<rdf:value>[77] [B. Church building]
"Mark Noll's ingenious calculations have shown that by the 1850s the Methodists had constructed almost as many churches as there were post offices in the United States and employed almost as many ministers as there were postal workers. 'Considered together,' he writes, 'the evangelical churches employed nearly double the personnel, maintained nearly twice as many facilities, and raised at least three times the money as the post office. Moreover the churches delivered their message to more people than the postal services delivered letters and newspapers.'"
</rdf:value>
<dc:subject>econ</dc:subject>
<dc:subject>hist</dc:subject>
<dc:subject>none</dc:subject>
<dc:subject>pc</dc:subject>
</ns4:Memo>

you're aware that the code for the translator is available on github, yes?
I'm not saying you have to look through this, but if you want to find a solution yourself, that'd be the place to look.

For Zotero RDF _export_, Zotero uses
https://github.com/zotero/translators/blob/master/Zotero%20RDF.js

For Bibliontology RDF Import - the richest import (and I thought the default from Scribe?):
https://github.com/zotero/translators/blob/master/Bibliontology%20RDF.js

And for general RDF:
https://github.com/zotero/translators/blob/master/RDF.js

if I read what s/he says correctly, this has worked with some similar files before though.
and this:

The best I can offer you without some help is to suggest that, like Scribe, Zotero doesn't like sources with more than 50 notes (since that's the only difference from the file that worked earlier today).

suggests it might be the same bug you fixed earlier today after all. In which case trying this with the branch xpi would probably be worth it.

Aurias,

Thank you the ID issue seems to have resolved it, I wonder how I broke that after my earlier version...oh well.

As for escaping quote marks, apostrophes, and a few other basic characters within the memo values, that does not appear to be required. It is actually present when I feed it into the python etree parser, but the values that come out after I dump it back out are no longer escaped. They import properly into Zotero, and I can't really think of a reason they shouldn't (since they are valid in HTML under most conditions these days).

I just need to build a good method for preserving line breaks and restoring formatting. But I have a sample RDF export from Zotero that helpfully provides enough of a model to make that work.

Thanks.