Interlinear Import Failed (from ELAN)

61 views
Skip to first unread message

Daniel Krauße

unread,
Feb 4, 2018, 6:33:49 AM2/4/18
to FLEx list
Hello,

I have tried to get old time-aligned ELAN files into Flex using the workflow by Tim Gaved and Sophie Salffner. I cannot use their template because I would need to start from scratch, so I tried to adjust all setting in ELAN to fit their template for Flex. I exported the file to Flex, opened it, and the following error message appeared:

"Interlinear Import Failed. Are you sure you are importing a file in FLExText Interlinear Format? Word, Open Office, and Libre Office docs cannot yet be imported."

Of course, I am importing a *.flextext file, but there must be a problem with the tiers and types, which look exactly as in the Tim Gaved workflow description. Has anyone experienced this before and found a solution?

Thanks in advance!

Best,
Daniel
VVKHVDP21Jul0701-MC-FLEX1.eaf
VVKHVDP21Jul0701-MC.flextext

Ann Bush

unread,
Feb 5, 2018, 3:01:47 PM2/5/18
to flex...@googlegroups.com

This doesn’t work.  There is an issue open to fix the problem, but it is not currently being worked on.

 

Ann

--
You are subscribed to the publicly accessible group "FLEx list".
Only members can post but anyone can view messages on the website.
To change your status, please write to flex_d...@sil.org.
You can join this group by going to http://groups.google.com/group/flex-list.
---
You received this message because you are subscribed to the Google Groups "FLEx list" group.
To unsubscribe from this group and stop receiving emails from it, send an email to flex-list+...@googlegroups.com.
To post to this group, send email to flex...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/flex-list/71e2a742-141b-4192-8630-681e1c41bc64%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Tim Gaved

unread,
Feb 6, 2018, 5:02:56 AM2/6/18
to flex...@googlegroups.com
Daniel,
The problem is that at the end of the ELAN file you have two empty annotations. This generates some xml in the FLExtext file that FLEx doesn't like. Either delete the empty annotations from the .eaf file or delete the section below from the .flextext file. (I'm not certain whether it's the empty phrase that's the problem or the empty word)
TIm

            <paragraph guid="6fd64708-35de-4073-ae7c-ec18afe8ff44">
                <phrases>
                    <phrase begin-time-offset="1185450"
                        end-time-offset="1187310"
                        guid="4b0f2444-ce98-4273-8e5e-4c8862deba1f"
                        media-file="995d1aab-d1e8-4778-aa8d-b86839c7ec6a" speaker="A">
                        <item lang="msn" type="txt"/>
                        <words>
                            <word guid="aa350e2c-5638-4229-81bf-db377135f81a">
                                <item lang="msn" type="txt"/>
                            </word>
                        </words>
                        <item lang="en" type="gls"/>
                        <item lang="en" type="note">A</item>
                    </phrase>
                </phrases>
            </paragraph>
            <paragraph guid="460c5b1a-6d6a-4447-8b79-0a17080716f2">
                <phrases>
                    <phrase begin-time-offset="1187700"
                        end-time-offset="1189660"
                        guid="0cb14a67-2706-46ab-8c63-83bcf5792225"
                        media-file="995d1aab-d1e8-4778-aa8d-b86839c7ec6a" speaker="A">
                        <item lang="msn" type="txt"/>
                        <words>
                            <word guid="aa350e2c-5638-4229-81bf-db377135f81a">
                                <item lang="msn" type="txt"/>
                            </word>
                        </words>
                        <item lang="en" type="gls"/>
                        <item lang="en" type="note">A</item>
                    </phrase>
                </phrases>
            </paragraph>

Daniel Krauße

unread,
Feb 7, 2018, 3:16:50 AM2/7/18
to FLEx list
Hi Tim,

Thanks for the hint! I just deleted the empty annotations in the ELAN file and was able to import it into Flex, everything worked well, but then there's another issue. As you can see in the image, nothing happens with the glosses:



The lexicon has some 3600 entries, but the text does not link to the dictionary at all. The parser doesn't even start as if the dictionary was empty. I tried it with a new project and imported the lexicon, but again, the parser does not "see" the dictionary. What's wrong with the exported ELAN file?

Best,
Daniel
Auto Generated Inline Image 1

Tim Gaved

unread,
Feb 7, 2018, 4:12:50 AM2/7/18
to flex...@googlegroups.com
Daniel,
Unfortunately that's the limit of what FLEx can import. The help file says this:
"Currently, the Baseline tab words (the text) and Free translations, are imported, but not include Lit or Note lines, or text charts. It does not change the lexicon. The Info tab data is imported, except any Genres field selections."
Doing a full import of all the parsing is a non-trivial task, for which there are not enough resources to write the software for at the moment (and it is probably not a very high priority either). This was why the workflow that Sophie and I originally wrote up was to annotate in ELAN, parse in FLEx, and then reimport to ELAN.
Sorry to disappoint you.
Tim

Daniel Krauße

unread,
Feb 7, 2018, 8:18:40 AM2/7/18
to FLEx list
Hi Tim,

I think I didn't express myself well.

I am not trying import any interlinear glossing from ELAN, there is none. The ELAN files only contain the original text and the translation, no glossing. But what I want to do is using an existing LIFT file as the dictionary for the recently imported ELAN file.

My workflow is like this:
1. Having an existing time-aligned, non-glossed, but translated ELAN file. (done)
2. Adjusting all tiers and types to fit the Flex template to prepare the export. (done)
3. Export to Flex file. (done)
4a. Import into an existing Flex database with the same language settings and an existing dictionary. (import working, but parsing no working because the interlinearization is not taken from the dictionary entries)
4b. Import into a new project and importing an existing LIFT dictionary (again, import working, but I would have to add everything again to the dictionary even though those words already exist)
5. Parse everything and re-import into ELAN. (don't even get there)

So, I think the imported file somehow cannot link to the existing dictionary. What am I missing? Is there any way to tell Flex that the baseline should link to the lexicon? I don't even know how to rephrase my thoughts because this case seems very strange...

Best,
Daniel

Tim Gaved

unread,
Feb 7, 2018, 9:12:15 AM2/7/18
to FLEx list
Daniel,
It's possible that there is a mismatch between writing systems. If this is same project as in your discussion with Andreas, then I notice that the lexemes there are labelled with the abbreviation "Kay", and not "Vur" the abbreviation for the language with the code "msn".
Goto  - Tools/Configure/Set Up Writing Systems.
In that dialogue do you have more than one vernacular writing system? If so that could be the problem, with all the texts in one writing system, and all the dictionary lexemes in another.
If the lexemes are in the wrong writing system, then you will need to do a bulk edit entries, add a column for the lexeme in the correct writing system, and bulk copy the lexemes into that.
If the text is in the wrong writing system then you can check it by going to the Baseline tab, selecting the whole text and then changing the writing system from the list box just below the menu bar. For any future exports from ELAN rename your word tiers so that they end with the code of the correct writing system.
Now go back to the Set Up Writing Systems dialogue and hide the incorrect writing system.
Hope that helps

Tim


On 07/02/2018 08:16, Daniel Krauße wrote:

Daniel Krauße

unread,
Feb 7, 2018, 8:55:02 PM2/7/18
to FLEx list
Hi Tim,

The other discussion with Andreas was about another project, for Kaytetye (gbb), which has different issues. For the Vurës database, I only have two writing systems (Vurës and Bislama), but I needed to delete the Bislama tier in ELAN because it wouldn't let me import two translation tiers.

But your last comment was absolutely perfect! Somehow, the writing system for the imported text was blank, so I could adjust it in the baseline tab, which worked perfectly! Great help, I am happy. :-)

Best,
Daniel

Tim Gaved

unread,
Feb 8, 2018, 3:40:41 AM2/8/18
to flex...@googlegroups.com
Daniel,
I can't work out why that should have happened. Your file imports ok for me, with Vurës correctly selected as the language of the baseline, so the FLExtext file is fine I think.
However, glad you've managed to work around the problem.
Tim
Reply all
Reply to author
Forward
0 new messages