Systems Frontend issueshttps://gitlab.rlp.net/motionbank/applications/systems-frontend/-/issues2022-11-07T09:56:09+01:00https://gitlab.rlp.net/motionbank/applications/systems-frontend/-/issues/1694Set agent in data model for automatically generated annotations2022-11-07T09:56:09+01:00Rittershaus, DavidSet agent in data model for automatically generated annotationsIf we get automatically generated annotations from the AI pipeline (whisper speech to text or later kinesphere labels) the "generator" of the annnotations should be specified in our data model: https://www.w3.org/TR/annotation-model/#age...If we get automatically generated annotations from the AI pipeline (whisper speech to text or later kinesphere labels) the "generator" of the annnotations should be specified in our data model: https://www.w3.org/TR/annotation-model/#agents
Will also be helpful to easily display and hide them.Feature: Data Viewshttps://gitlab.rlp.net/motionbank/applications/systems-frontend/-/issues/1730Show analysis jobs in the frontend2022-11-07T09:55:08+01:00Koch, AntonShow analysis jobs in the frontendThe jobs `bounding_box`, `tracking` and `transcript` now show up in the `$jobs` array.The jobs `bounding_box`, `tracking` and `transcript` now show up in the `$jobs` array.Feature: Data Viewshttps://gitlab.rlp.net/motionbank/applications/systems-frontend/-/issues/1717"people detection" tooltip: add text "visible when video is paused"?2022-11-07T09:54:37+01:00Hansen, Christian"people detection" tooltip: add text "visible when video is paused"?I just had the situation that I wanted to activate the detection and wondered why no boxes appeared. This was because the video was playing. This situation could also occur with other users, so a little hint would be useful in my opinion...I just had the situation that I wanted to activate the detection and wondered why no boxes appeared. This was because the video was playing. This situation could also occur with other users, so a little hint would be useful in my opinion.
<br /><br />
![Bildschirmfoto_2022-10-21_um_16.31.13](/uploads/9d5c74d787a12b2bbe5c7638138450f2/Bildschirmfoto_2022-10-21_um_16.31.13.png)Feature: Data ViewsMilan GressMilan Gresshttps://gitlab.rlp.net/motionbank/applications/systems-frontend/-/issues/1743Allow the annotation of audio files?2022-11-07T09:53:01+01:00Rittershaus, DavidAllow the annotation of audio files?I think this has already been a feature request in the past and you are sort of able to do it if you turn your audio file into a video file. With the transcript option this seems now to be even more interesting. Would be kind of cool if ...I think this has already been a feature request in the past and you are sort of able to do it if you turn your audio file into a video file. With the transcript option this seems now to be even more interesting. Would be kind of cool if you can add an audio file to your timeline and then annotate it like a video.
Video frame in post annotator could simply show the waveform image. Image related features (bouding boxes etc.) should probably be hidden. Maybe the whisper container is set to process only video files at the moment?
@anton.kochMilan GressMilan Gresshttps://gitlab.rlp.net/motionbank/applications/systems-frontend/-/issues/1698Wire speech-to-text request (whisper container)2022-11-07T09:52:47+01:00Rittershaus, DavidWire speech-to-text request (whisper container)When #1697 is implemented, the UI needs to be wired to the whisper container (requesting transcription, transmitting selected language).
This should come after the bounding boxes are properly implemented.When #1697 is implemented, the UI needs to be wired to the whisper container (requesting transcription, transmitting selected language).
This should come after the bounding boxes are properly implemented.Feature: Data Viewshttps://gitlab.rlp.net/motionbank/applications/systems-frontend/-/issues/1699Save speech-to-text transcription as annotations2022-11-07T09:52:40+01:00Rittershaus, DavidSave speech-to-text transcription as annotations#1698 should return segments of automatically transscripted speech with timestamps. They have to be mapped to our annotation model and should be saved. It would make sense to set an agent to identify how they have been created (and also ...#1698 should return segments of automatically transscripted speech with timestamps. They have to be mapped to our annotation model and should be saved. It would make sense to set an agent to identify how they have been created (and also to filter them). See issue #1694.
Again: bounding boxes have a higher priority.Feature: Data Viewshttps://gitlab.rlp.net/motionbank/applications/systems-frontend/-/issues/1732Use transcript information in the front end2022-11-07T09:52:33+01:00Koch, AntonUse transcript information in the front endYou can query the transcript info (after the job is finished) by calling the `/analysis` endpoint as for a bounding box, but without the `time` param and instead specifying `type=transcript`. If data exists, it will return an array conta...You can query the transcript info (after the job is finished) by calling the `/analysis` endpoint as for a bounding box, but without the `time` param and instead specifying `type=transcript`. If data exists, it will return an array containing transcription items with a local timecode for the media file.Feature: Data ViewsMilan GressMilan Gresshttps://gitlab.rlp.net/motionbank/applications/systems-frontend/-/issues/1679Show AI processes (running and ready) in /Files2022-11-07T09:52:17+01:00Hansen, ChristianShow AI processes (running and ready) in /FilesNo analysis starting there
<br /><br />
![Bildschirmfoto_2022-10-13_um_10.53.09](/uploads/1afcf7e8558568bbe82ef565050eaddf/Bildschirmfoto_2022-10-13_um_10.53.09.png)
<br /><br />
![Bildschirmfoto_2022-10-13_um_11.36.42](/uploads/d8367f53...No analysis starting there
<br /><br />
![Bildschirmfoto_2022-10-13_um_10.53.09](/uploads/1afcf7e8558568bbe82ef565050eaddf/Bildschirmfoto_2022-10-13_um_10.53.09.png)
<br /><br />
![Bildschirmfoto_2022-10-13_um_11.36.42](/uploads/d8367f53e930da6216e6ce8f2acfe652/Bildschirmfoto_2022-10-13_um_11.36.42.png)Feature: Data ViewsMilan GressMilan Gresshttps://gitlab.rlp.net/motionbank/applications/systems-frontend/-/issues/1697Add "Speech to Text" to Video Analysis menu2022-11-04T13:57:47+01:00Rittershaus, DavidAdd "Speech to Text" to Video Analysis menuThis has a lower priority compared to finishing the implementation of the bounding boxes and should be addressed afterwards.
Since they are preparing the deplopyment of container for whisper (speech to text) at wizai, we should prepare ...This has a lower priority compared to finishing the implementation of the bounding boxes and should be addressed afterwards.
Since they are preparing the deplopyment of container for whisper (speech to text) at wizai, we should prepare the UI to implement it. When you select this feature, you should be requested to define the language of your video, because whisper can only automatically identify the language in the first 30 seconds. If there's nothing to hear in these 30s or another language it might fail with the transcription or perform badly.
Here are the languages that are supported with the respective error rate:
![language-breakdown.svg](/uploads/1b6b360e47394d93b91bd4291efcb44f/language-breakdown.svg)
Maybe we could offer everything in a drop down up to a score of 21 (Hungarian) but I would offer English and German at the first place, because those might be the most common ones for our users.Feature: Data ViewsMilan GressMilan Gresshttps://gitlab.rlp.net/motionbank/applications/systems-frontend/-/issues/1738Remove unnecessary reloading of Annotations from Server on changes2022-10-27T14:58:21+02:00Mittermeier, MaximilianRemove unnecessary reloading of Annotations from Server on changesstore/postAnnotator/loadAnnotations() is only necessary once when opening postAnnotator to initially populate internal vuex-store with Annotations from server.
the .annotations computed property on post-annotate.vue automatically update...store/postAnnotator/loadAnnotations() is only necessary once when opening postAnnotator to initially populate internal vuex-store with Annotations from server.
the .annotations computed property on post-annotate.vue automatically updates when Annotations are added/edited/removed.Mittermeier, MaximilianMittermeier, Maximilianhttps://gitlab.rlp.net/motionbank/applications/systems-frontend/-/issues/1721Cache labeled Ticks in RulerTrack2022-10-27T10:19:30+02:00Mittermeier, MaximilianCache labeled Ticks in RulerTrackMittermeier, MaximilianMittermeier, Maximilianhttps://gitlab.rlp.net/motionbank/applications/systems-frontend/-/issues/1687Swimlane: default size depending on annotations?2022-10-27T09:07:11+02:00Rittershaus, DavidSwimlane: default size depending on annotations?I still think it's a bit weird that I always have to adjust the the swimlane size if I have no annotations with duration (and no other lanes). The video is way too small to work with it while there is a lot of black/white space in the sw...I still think it's a bit weird that I always have to adjust the the swimlane size if I have no annotations with duration (and no other lanes). The video is way too small to work with it while there is a lot of black/white space in the swimlane by default.
![Bildschirmfoto_2022-10-17_um_09.52.56](/uploads/104e90f114be5fc6c6fab2facd544a12/Bildschirmfoto_2022-10-17_um_09.52.56.png).
I also see the issue that users might not understand where they can find their annotations with duration if you can't see that lane by default or if it's so small, so that you can only see one or two annotations with duration. This might be tricky coding wise, but I'm wondering if the default size of the swimlane could depend on amount of annotations with duration? If there are no annotations with duration the second lane could be quite small, only offering enough space to drop an annotation from the lane above etc.Feature: SwimlaneMittermeier, MaximilianMittermeier, Maximilianhttps://gitlab.rlp.net/motionbank/applications/systems-frontend/-/issues/1737Post Annotator does not load more than 50 Annotations2022-10-27T09:05:41+02:00Mittermeier, MaximilianPost Annotator does not load more than 50 AnnotationsMittermeier, MaximilianMittermeier, Maximilianhttps://gitlab.rlp.net/motionbank/applications/systems-frontend/-/issues/1688Swimlane: add instruction for annotations with duration?2022-10-26T17:16:40+02:00Rittershaus, DavidSwimlane: add instruction for annotations with duration?Could we add something like in the mock up below to the lane with the annotations with duration (as long as there is no annotation with duration)?
![Bildschirmfoto_2022-10-17_um_10.01.41](/uploads/78e990196bc30a7efa015211b2957e80/Bildsc...Could we add something like in the mock up below to the lane with the annotations with duration (as long as there is no annotation with duration)?
![Bildschirmfoto_2022-10-17_um_10.01.41](/uploads/78e990196bc30a7efa015211b2957e80/Bildschirmfoto_2022-10-17_um_10.01.41.png)Feature: Swimlanehttps://gitlab.rlp.net/motionbank/applications/systems-frontend/-/issues/1729Analysis jobs should not block the annotation option2022-10-26T16:41:07+02:00Koch, AntonAnalysis jobs should not block the annotation optionAs the bounding_box, tracking and transcript jobs now run and are reported in the `$jobs` array, the media file cannot be annotated until all jobs are finished. This needs to be changed to only block annotation until `preview` and `stand...As the bounding_box, tracking and transcript jobs now run and are reported in the `$jobs` array, the media file cannot be annotated until all jobs are finished. This needs to be changed to only block annotation until `preview` and `standard-version` are done.Feature: Data ViewsHansen, ChristianHansen, Christianhttps://gitlab.rlp.net/motionbank/applications/systems-frontend/-/issues/1649Adjustments to UI element sizes in Post-Annotator2022-10-26T15:21:41+02:00Rittershaus, DavidAdjustments to UI element sizes in Post-AnnotatorThere is a major issue regarding the size of the video in the Post-Annotator interface. It's linked to the swimlane (#1644) but even before we took more space for UI elemets (like the bigger annotation input field) compared to previous v...There is a major issue regarding the size of the video in the Post-Annotator interface. It's linked to the swimlane (#1644) but even before we took more space for UI elemets (like the bigger annotation input field) compared to previous versions. We should check which elements can be adjusted to get a bigger video frame. This is really important since our users often have to focus on small details in the video which are anyway hard to see.
I think in this screenshot the issue is clearly visible:
![Bildschirmfoto_2022-09-27_um_12.50.03](/uploads/4d69d50054b77a55731d3952f14ccf23/Bildschirmfoto_2022-09-27_um_12.50.03.png)Hansen, ChristianHansen, Christianhttps://gitlab.rlp.net/motionbank/applications/systems-frontend/-/issues/1703Swimlane allows you to change start- and endtime of annotations that are not ...2022-10-25T18:34:10+02:00Rittershaus, DavidSwimlane allows you to change start- and endtime of annotations that are not your ownWhile I don't have an "edit" option I can nevertheless change the position of annotations created by others on the swimlane (change start- and endtime via simwlane interface).
Not critical but should be fixed at some point.While I don't have an "edit" option I can nevertheless change the position of annotations created by others on the swimlane (change start- and endtime via simwlane interface).
Not critical but should be fixed at some point.Feature: Swimlanehttps://gitlab.rlp.net/motionbank/applications/systems-frontend/-/issues/1724Offer download option for files2022-10-25T17:22:45+02:00Rittershaus, DavidOffer download option for filesActual usecase: draw on video has been used and the drawing is supposed to be used outside of the software.
Might also be relevant if you don't have your video files anymore that you uploaded (e.g. from your phone).Actual usecase: draw on video has been used and the drawing is supposed to be used outside of the software.
Might also be relevant if you don't have your video files anymore that you uploaded (e.g. from your phone).Hansen, ChristianHansen, Christianhttps://gitlab.rlp.net/motionbank/applications/systems-frontend/-/issues/1718Swimlane should have indication of possible Drag'n'Drop interactions2022-10-25T13:21:35+02:00Mittermeier, MaximilianSwimlane should have indication of possible Drag'n'Drop interactionsFeature: SwimlaneMittermeier, MaximilianMittermeier, Maximilianhttps://gitlab.rlp.net/motionbank/applications/systems-frontend/-/issues/1719Grouped Annotations on Swimlane should have a tooltip2022-10-25T13:21:22+02:00Mittermeier, MaximilianGrouped Annotations on Swimlane should have a tooltipFeature: SwimlaneMittermeier, MaximilianMittermeier, Maximilian