Im trying to get Gramps to import Ancestry’s export with as few errors as possible. But my use of “Other sources” or “Web links” is generating meny errors.
I’ve identified the following problem but don’t know what is the next step to solve it.
Ancestry.com is not following the GEDCOM-format 5.5.1 when using “Other sources” or “Web links”, even if they say the file is that format and standard.
These are the validation errors from GED-inline
Tag WWW is not allowed under DATA
Tag PLAC is not allowed under PUBL
Tag CALN is not allowed under SOUR
Gramps import errors:
“Line ignored as not understood” for WWW and TITL
" Skipped subordinate line" for PLAC
Those errors are my main problem with geting going with gramps. Is there a known solution to this errors? Any help would be appreciated.
Tag recognized but not supported REFN and CALN. But that is another issue and those are few and could be handel manually in gramps.
I know a bit of python but noting about Gramps so would need to be pointed in the right direction if im part of a solution to this problem.
Keep in mind that Gramps data model doesn’t exactly match GEDCOM, so it might not be possible to get this into a reasonable place in Gramps.
If Ancestry is ‘enhancing’ GEDCOM, as often occurs with various programs, it might be even more difficult to put data into Gramps.
We can always drop the data into a note. Actually it is almost certainly already in a note, the one indicating imort errors attached to the primary object (INDI-person or SOUR-source) . But it would look better if it did not include the GEDCOM lines and just the data.
If I can figure this out, I might be able to create a patch that would help you out.
I think im on to something, have been studying the code in libgedcom and I see that there are other special cases so. I will add this one too. Just needs more testing, then I will push PR.
Essentially this is my plan:
libgedcom.py
citation_data_tbl add TOKEN_WWW
__citation_data_www probably like __citation_link
Im looking in to NoteType now trying to understan the solution in __citation_link.
In particular we use the SOURCE_TEXT type to store the contents of the Gedcom SOUR.DATA.TEXT tag.
The CITATION type can be used to store formatted citations.
The LINK type may be appropriate for the WWW tag. I’m not really sure. Perhaps this would be better stored in a source attribute? A web address together with the accessed date are citation elements.
A PLAC tag under the PUBL tag is actually quite a good idea. Publication information often consists of a publisher’s name, place and year of publication. We may decide to store this information separately in Gramps when we enhance the citation formatters further.
Interesting. I was thinking about notetype I settled on SOURCE_TEXT after reading the comments in notetype.py.
LINK was the obvious choice but function __citation_link dosnt use it but CITATION. And I dont understand the comment for LINK in notetype.py “# indicate a note used as link in another note”. And nothing in libgedcom is using LINK as notetype. So SOURCE_TEXT seems like the fallback.
I’m asking, because they’re often not recognized when you do a direct import from a downloaded GEDCOM, and there are a couple of workarounds for that.
One such workaround is to use RootsMagic Essentials next to Gramps. It has the added advantage that it will also download media that you have on Ancestry, and put that in a separate folder. If you then create a GEDCOM in RootsMagic it will have proper media references.
RootsMagic Essentials is free. I use it often to synchronize data with FamilySearch.
About the dates: I found that Ancestry often exports long names, like January, and AFAIK the GEDCOM importer only accepts English abbreviations, and possibly also long names if they are written in the users local language.
And that part gives a problem, because in my language, the long name is Januari, so it is not recognized. Same for October, what we write with a K.