eScriptorium issues
https://gitlab.inria.fr/scripta/escriptorium/-/issues
2021-02-02T11:21:24+01:00
https://gitlab.inria.fr/scripta/escriptorium/-/issues/239
Old models
2021-02-02T11:21:24+01:00
daniel stoekl
Old models
Neither legacy segmentation models (I knew that) nor transcription models (this is new to me) are working in the new instance. Is there a workaround?
Neither legacy segmentation models (I knew that) nor transcription models (this is new to me) are working in the new instance. Is there a workaround?
https://gitlab.inria.fr/scripta/escriptorium/-/issues/186
Catch old models error and display it nicely
2021-02-02T11:21:24+01:00
Robin Tissot
Catch old models error and display it nicely
```
Traceback:
File "/usr/src/app/apps/core/tasks.py" in transcribe
525. part.transcribe(model=model)
File "/usr/src/app/apps/core/models.py" in transcribe
590. model_ = kraken_models.load_any(model.file.path)
...
```
Traceback:
File "/usr/src/app/apps/core/tasks.py" in transcribe
525. part.transcribe(model=model)
File "/usr/src/app/apps/core/models.py" in transcribe
590. model_ = kraken_models.load_any(model.file.path)
File "/usr/local/lib/python3.7/site-packages/kraken/lib/models.py" in load_any
157. seq = TorchSeqRecognizer(nn, train=train, device=device)
File "/usr/local/lib/python3.7/site-packages/kraken/lib/models.py" in __init__
52. if nn.model_type not in [None, 'recognition']:
File "/usr/local/lib/python3.7/site-packages/kraken/lib/vgsl.py" in model_type
471. return self.user_metadata['model_type']
Exception Type: KeyError
Exception Value: 'model_type'
Request data not supplied
```
https://gitlab.inria.fr/scripta/escriptorium/-/issues/324
freeze possibility to rotate when segmenting (or transcribing?)
2021-02-01T16:22:50+01:00
daniel stoekl
freeze possibility to rotate when segmenting (or transcribing?)
I rotated an image in the middle of segmenting it automatically. Sent you the reference via email. Do need to freeze this possibility when item is locked.
I rotated an image in the middle of segmenting it automatically. Sent you the reference via email. Do need to freeze this possibility when item is locked.
https://gitlab.inria.fr/scripta/escriptorium/-/issues/321
deleting splitting regions needs automatic line de-reassociation
2021-01-27T16:16:24+01:00
daniel stoekl
deleting splitting regions needs automatic line de-reassociation
If one deletes a region, the lines associated to it should be automatically disassociated from it.
If one divides a region into two (or more) all lines in it should automatically be reassociated to the line in which their central point i...
If one deletes a region, the lines associated to it should be automatically disassociated from it.
If one divides a region into two (or more) all lines in it should automatically be reassociated to the line in which their central point is. Any other behavior is counterintuitive and will cause great confusion with users also because it is not shown in the UI that these lines are still associated to a ghost or to the wrong region. I call this a bug and not a feature request as it is unwanted behavior.
https://gitlab.inria.fr/scripta/escriptorium/-/issues/322
browse through regions
2021-01-27T15:01:31+01:00
daniel stoekl
browse through regions
We do need a UI tool to browse through the regions of a part. I have seen many examples where regions are so small or hidden under other points of another region that it is impossible or really hard to attain them individually. This coul...
We do need a UI tool to browse through the regions of a part. I have seen many examples where regions are so small or hidden under other points of another region that it is impossible or really hard to attain them individually. This could be a column with region numbers next to the line numbers in the textannotation panel, which are associated with these regions. This would also allow to check whether there are empty regions. Clicking on one of these region numbers would select that specific region that could then be deleted or modified or whatever.
https://gitlab.inria.fr/scripta/escriptorium/-/issues/320
Onboarding updates needed
2021-01-17T20:52:45+01:00
daniel stoekl
Onboarding updates needed
Please add "pdf" to the import options in onboarding.
Please add "pdf" to the import options in onboarding.
https://gitlab.inria.fr/scripta/escriptorium/-/issues/318
button to create a new transcription level
2021-01-15T18:09:53+01:00
daniel stoekl
button to create a new transcription level
We need a button to create new transcription levels or rename existing ones. Best here: ![image](/uploads/012173b0d23f9322fa6757add36d2140/image.png)
We need a button to create new transcription levels or rename existing ones. Best here: ![image](/uploads/012173b0d23f9322fa6757add36d2140/image.png)
https://gitlab.inria.fr/scripta/escriptorium/-/issues/317
toggle transcription comparison in transcription modal
2021-01-15T16:51:26+01:00
daniel stoekl
toggle transcription comparison in transcription modal
Pressing the "toggle transcription comparison" button in the transcription modal should activate or not the DIFF presentation in the transcription modal, i.e. either show the green and red boxes around characters different from "current"...
Pressing the "toggle transcription comparison" button in the transcription modal should activate or not the DIFF presentation in the transcription modal, i.e. either show the green and red boxes around characters different from "current" text or not (the text of each transcription level without DIFF comparison). However, the best solution would be a characterwise aligned display of the texts as asked for in another issue.
https://gitlab.inria.fr/scripta/escriptorium/-/issues/314
Incomplete importation?
2021-01-08T18:14:54+01:00
Alix Chagué
Incomplete importation?
Hi,
I'm seeing a weird behavior when I try to import a transcription. Not all segments are created, as you can see on the images below:
> document as it should be imported:
![image](/uploads/809bcd53da227a497cba243606b743d4/image.png)
...
Hi,
I'm seeing a weird behavior when I try to import a transcription. Not all segments are created, as you can see on the images below:
> document as it should be imported:
![image](/uploads/809bcd53da227a497cba243606b743d4/image.png)
> Document after import:
![image](/uploads/4eb11fe5b9e4c56e2e75fdae81031782/image.png)
It seems that the problem happens during the importation task because the transcriptions are in the XML files.
![image](/uploads/d9459ba05980c40553ca93a6524ae707/image.png)
Note that the bug happens whether I use ALTO or PAGE. It's a major problem for me at the moment because we were trying to finetune a segmenter using data exported from several eScriptorium documents.
Did you notice this issue before? We are on version 0.7d.
https://gitlab.inria.fr/scripta/escriptorium/-/issues/94
Importing from one instance of escriptorium to another fails
2021-01-04T10:51:49+01:00
Robin Tissot
Importing from one instance of escriptorium to another fails
Because the system tries to match the IDs, need to try/except that somehow?
Because the system tries to match the IDs, need to try/except that somehow?
https://gitlab.inria.fr/scripta/escriptorium/-/issues/312
keep region and line IDs in export
2021-01-04T10:51:48+01:00
daniel stoekl
keep region and line IDs in export
Currently, the export assigns cryptic region and line IDs in the export instead of using the eScriptorium identifiers. I think it would be much more convenient to keep them for external treatment and reimport because the identification w...
Currently, the export assigns cryptic region and line IDs in the export instead of using the eScriptorium identifiers. I think it would be much more convenient to keep them for external treatment and reimport because the identification would be straightforward as long as you stay on one platform. I see no advantage in the currently used cryptic IDs that break these links.
https://gitlab.inria.fr/scripta/escriptorium/-/issues/75
pdf upload for manuscripts and print
2021-01-04T10:47:30+01:00
STOKL BEN EZRA Daniel
pdf upload for manuscripts and print
Some libraries only permit pdf download. It would be extremely practical to be able to import them directly into eScriptorium. Similarly for prints which are almost always in pdf rather than single images.
Some libraries only permit pdf download. It would be extremely practical to be able to import them directly into eScriptorium. Similarly for prints which are almost always in pdf rather than single images.
https://gitlab.inria.fr/scripta/escriptorium/-/issues/311
Invitation to group doesn't make it clear the user needs to exist already
2020-12-18T15:48:26+01:00
Robin Tissot
Invitation to group doesn't make it clear the user needs to exist already
People trying to invite new users to the service through this form, need to change the wording.
People trying to invite new users to the service through this form, need to change the wording.
https://gitlab.inria.fr/scripta/escriptorium/-/issues/161
line rotation algo for the transcription panel
2020-12-17T15:43:45+01:00
daniel stoekl
line rotation algo for the transcription panel
For vertical or very oblique lines, the transcription panel becomes impractical without a rotation of the line. This should be easy to solve with an image rotation. However, for curved lines it will be more complex. You could use the lin...
For vertical or very oblique lines, the transcription panel becomes impractical without a rotation of the line. This should be easy to solve with an image rotation. However, for curved lines it will be more complex. You could use the line rectification algo which is I think in the contrib folder
![image](/uploads/7457de73319172f376d9be24ea561e08/image.png)
https://gitlab.inria.fr/scripta/escriptorium/-/issues/229
transcription of rotated lines does not work as well as the others
2020-12-17T15:43:29+01:00
daniel stoekl
transcription of rotated lines does not work as well as the others
There is a bug somewhere in the pipeline that sends lines that are not in the usual horizontal rotation to kraken. Try transcribing the same page with the same model but rotate it 3 times by 90 degrees. The rotations will always give muc...
There is a bug somewhere in the pipeline that sends lines that are not in the usual horizontal rotation to kraken. Try transcribing the same page with the same model but rotate it 3 times by 90 degrees. The rotations will always give much worse results. I sent this already by email but never got a response so I prefer to file it here so that it does not get lost.
https://gitlab.inria.fr/scripta/escriptorium/-/issues/100
image rotation
2020-12-17T15:42:53+01:00
STOKL BEN EZRA Daniel
image rotation
I have just imported a IIIF manifest. Three quarters of the manuscript are top down. So I think we do need a button to bulk rotate selected images.
I have just imported a IIIF manifest. Three quarters of the manuscript are top down. So I think we do need a button to bulk rotate selected images.
https://gitlab.inria.fr/scripta/escriptorium/-/issues/308
Line gets bound to a ghost region (?)
2020-12-17T15:42:36+01:00
Robin Tissot
Line gets bound to a ghost region (?)
Gives a js error when trying to delete or update it.
> Uncaught TypeError: this.region.get is not a function
Gives a js error when trying to delete or update it.
> Uncaught TypeError: this.region.get is not a function
https://gitlab.inria.fr/scripta/escriptorium/-/issues/309
wrong codepoint ordering
2020-12-17T15:27:46+01:00
daniel stoekl
wrong codepoint ordering
I don't know whether it is on the level of kraken or of eScriptorium but sometimes lines in RTL texts that begin with a chain of LTR words seem ordered in the wrong sens.
![image](/uploads/55a0eade44451531e906250158c8587b/image.png)
I don't know whether it is on the level of kraken or of eScriptorium but sometimes lines in RTL texts that begin with a chain of LTR words seem ordered in the wrong sens.
![image](/uploads/55a0eade44451531e906250158c8587b/image.png)
https://gitlab.inria.fr/scripta/escriptorium/-/issues/310
need transcription modal max height
2020-12-17T15:26:47+01:00
daniel stoekl
need transcription modal max height
For short lines the modal takes up the whole page so that e.g. the comparison lines are not displayed, i.e. https://www.escriptorium.fr/document/301/part/55455/edit/
bottom line (the only one with manual entry). One can get more real es...
For short lines the modal takes up the whole page so that e.g. the comparison lines are not displayed, i.e. https://www.escriptorium.fr/document/301/part/55455/edit/
bottom line (the only one with manual entry). One can get more real estate with F11 but for some lines
even that does not suffice.
![image](/uploads/dc6ae8c5eb6d9bb1b2aaec72dca85950/image.png)
https://gitlab.inria.fr/scripta/escriptorium/-/issues/162
full image rotation
2020-12-09T11:13:26+01:00
daniel stoekl
full image rotation
Sometimes images imported via iiif etc come in the wrong rotation. We need a possibility to rotate by 90/180 degrees
![image](/uploads/f1a0ceb0bada23db1fdfddf83770e1b9/image.png)
Sometimes images imported via iiif etc come in the wrong rotation. We need a possibility to rotate by 90/180 degrees
![image](/uploads/f1a0ceb0bada23db1fdfddf83770e1b9/image.png)