Display of Annotation Type #1192
Conversation
Priya4607
force-pushed
the
auto_vs_manual
branch
from
5ada9a5 to
2712d54
Compare
Priya4607
changed the title
|
First of all please fix codacy and build (probably you forgot to add migrations). Please check that the patch works in case of undo/redo, propagation, ... When you ready please remove WIP from the patch and we will review and provide feedback. |
Priya4607
force-pushed
the
auto_vs_manual
branch
from
2712d54 to
0ae7df8
Compare
Priya4607
force-pushed
the
auto_vs_manual
branch
from
0ae7df8 to
f0303cf
Compare
|
@azhavoro , could you please review and give your opinion. |
There was a problem hiding this comment.
Thanks for the contribution.
I have question:
What do you think about support of this field for export/import annotations functionality (CVAT format only). For now if I dump and upload anno again I will completely lose auto/manual field for all objects, right?
There was a problem hiding this comment.
@Priya4607 , we have a conflict. Could you please resolve it?
@azhavoro , could you please summarise all your comments into a check list? Thus it will be easy to understand where we are with the PR.
Priya4607
force-pushed
the
auto_vs_manual
branch
from
e985dc5 to
da8726b
Compare
|
@Priya4607 thank you for the update
Could you please fix the build and squash all migrations into one file? |
Priya4607
force-pushed
the
auto_vs_manual
branch
from
da8726b to
5c170c8
Compare
|
@Priya4607 , what is your plans regarding the PR? |
|
@nmanovic |
|
I'm ok with the back-end changes (2 minor comments above). |
*added field type for tracks
|
@Priya4607 , could you please look at #1767. Do you think you can adapt your changes to the PR? It looks like I will merge it before your PR. I hope that together we can merge your PR till middle of July. Really appreciate your time and efforts! |
|
@azhavoro , please work with @Priya4607 to accept the PR. What should be done for that? Could you please preview it again and provide your comments? |
| migrations.AlterField( | ||
| model_name='labeledshape', | ||
| name='type', | ||
| field=models.CharField(choices=[('rectangle', 'RECTANGLE'), ('polygon', 'POLYGON'), ('polyline', 'POLYLINE'), ('points', 'POINTS'), ('cuboid', 'CUBOID')], max_length=16), |
There was a problem hiding this comment.
Why do we have the migration for "type" field here?
There was a problem hiding this comment.
Had a migration conflict. So I merged migrations to resolve it.
There was a problem hiding this comment.
@Priya4607 , we should have the old migration file and plus new one with only your changes.
There was a problem hiding this comment.
When I create a new migration, I get the labeledshape model changes added. I'm not sure how to include my changes alone. Could you provide some suggestions on how to proceed?
There was a problem hiding this comment.
@Priya4607 , it was a problem with develp branch. It missed some migrations. I have fixed that. Could you please update and create a new migration?
cvat/apps/tf_annotation/views.py
Outdated
| @@ -174,6 +174,7 @@ def convert_to_cvat_format(data): | |||
| "group": None, | |||
| "occluded": False, | |||
| "attributes": [], | |||
| "annotation_type": "Auto", | |||
There was a problem hiding this comment.
@Priya4607 , can we rename "annotation_type" into "source". Also I will vote for "auto", "manual" (fist letter should be lower case). In the future we can store in the field even username (or its id).
There was a problem hiding this comment.
Okay, should it be applied to both frontend and backend to maintain uniformity?
|
@Priya4607 , we are ready to merge the PR. Let's do the final minor refactoring, update CHANGELOG.md and I will merge. Thanks for the great idea and contribution. |
@zhiltsov-max Updated test cases |
CHANGELOG.md
Outdated
| @@ -61,6 +61,8 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0 | |||
| - [Datumaro] Added `convert` command to convert datasets directly (<https://github.com/opencv/cvat/pull/1837>) | |||
| - [Datumaro] Added an option to specify image extension when exporting datasets (<https://github.com/opencv/cvat/pull/1799>) | |||
| - [Datumaro] Added image copying when exporting datasets, if possible (<https://github.com/opencv/cvat/pull/1799>) | |||
| - Source type support for tags, shapes and tracks (<https://github.com/opencv/cvat/pull/1192>) | |||
There was a problem hiding this comment.
Wrong place. Could you please move these lines into the section for next release?
|
@Priya4607 , I'm fine with the PR. Need to fix CHANGELOG.md and I will merge. Thanks! |
|
Complexity increasing per file
==============================
- cvat/apps/engine/migrations/0027_auto_20200719_1552.py 1
- cvat-core/src/annotations-objects.js 1
Clones removed
==============
+ cvat-core/src/annotations-saver.js -1
See the complete overview on Codacy |
CHANGELOG.md
Outdated
| @@ -467,6 +468,12 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0 | |||
| ### Added | |||
| - Initial version | |||
|
|
|||
| ## [0.6.1] - 2020-04-14 | |||
| ### Added | |||
| - Annotation type to determine whether a drawn shape is auto annotated using Tensorflow Object detection API or manually annotated by the user | |||
There was a problem hiding this comment.
@Priya4607 , some problems with merge. These lines are wrong.
There was a problem hiding this comment.
Removed those lines.
|
@Priya4607 , thanks again for the PR. If you can add the information into statistics about the task (aka info), it will be awesome. Thus we can understand number of manual, interpolated and automatic annotated objects. |
This PR adds a feature which can help the user distinguish which of the boxes are auto-annotated. When a user annotates an image it is marked as
(MANUAL)in each of the boxes created and in case of auto-annotation it is marked as(AUTO). When the user edits a auto-generated box it is changed from(AUTO)to(MANUAL).