First version of mask metadata #55

joshmoore · 2020-07-07T11:58:05Z

see also:

spec.md

Co-authored-by: Simon Li <[email protected]>

joshmoore · 2020-07-07T14:53:02Z

Changes pushed, @manics. Thanks.

manics

Added an image reference as a suggestion
https://github.com/ome/omero-cli-zarr/blob/184ddf4b6d7848895a281d7e997c808b76d1b97d/src/omero_zarr/masks.py#L172-L173
Also happy to open this as a separate PR after this is merged.

spec.md

manics · 2020-07-07T16:35:17Z

spec.md

+    "1": 8388736,
+    ...
+```
+


Suggested change

### "image"

The `image` key is an optional dictionary which contains information on the image the mask is associated with.

If included it must include a key `array` whose value that is either:

- A relative path to a Zarr image array, for example:

```

{

"image": {

"array": "../../0"

}

}

```

- A URL to a Zarr image array (use this if the mask is stored seperately from the image Zarr), for example:

```

{

"image": {

"array": "https://s3.embassy.ebi.ac.uk/idr/zarr/v0.1/6001240.zarr/0"

}

}

```

See also open discussion at ome/omero-cli-zarr#19 (comment) about the specification of the image key

spec.md

will-moore · 2020-07-09T09:38:20Z

spec.md

+
+### "color"
+
+The `color` key defines an image that is "labeled", i.e. every unique value in the image


We also export color with style=--split. The mask only has 1 values (not labelled) but it does have a color.

Hmmm.... reading it, I'm not sure just having only one label makes something not labeled. The example also only shows one value ;)

Hmm, so both --style=labelled and --style=split produce masks that are labelled. Maybe this is a bit less confusing if we ignore the --style=labelled option (since that is the default). If we also consider removing the non-compliant 6d option then all masks are "labelled". Are any masks NOT "labelled"? So I think you can remove that whole sentence "The color key defines an image that is labeled, i.e. every unique value in the image represents a unique, non-overlapping object within the image." since even without any 'color', that statement is still true.

jburel · 2020-07-09T09:55:13Z

Discussed this morning with @will-moore and @manics, we need to indicate the resolution the mask corresponds to.
I have been looking at large images (bdv from IDR), and did some segmentation on resolution 4 then saving the segmented planes. The description for mask should probably indicate the source image/resolution so when we load the mask only, we know what to load next if required.

joshmoore · 2020-07-10T07:33:32Z

@manics Also happy to open this as a separate PR after this is merged.

If you don't mind, let's go for that. You can take the reins on a next round of PRs without all mine to get in your way while I'm gone.

joshmoore · 2020-07-10T07:37:26Z

the need to indicate the resolution...

see c15b68b

Co-authored-by: Simon Li <[email protected]>

jburel · 2020-07-10T07:40:07Z

spec.md

-                ├── .zattrs   # "color".
-                │
-                └── t.c.z.y.x # Chunks as above.
+                ├── .zarray   # Each mask itself is a 5D array matching the highest resolution


that means that we won't be able to store masks generated at a different resolution

Not yet, no. But then ome-zarr-py code likely wouldn't support that either, nor omero-cli-zarr that output. Let's file that as an issue on one or several of the repositories and then we can work through the implications. I'm trying to get this PR into a state that it can be merged before I leave so no one is blocked on me.

understood.
mainly a problem I am facing with the large DBV files that I have been segmenting at a resolution that is not the highest one but we can have that in another round

@jburel can you file an issue about the mask resolutions?

spec.md

mtbc · 2020-07-15T12:25:58Z

Staged at https://github.com/joshmoore/omero-ms-zarr/blob/masks-etc/spec.md.

Co-authored-by: Simon Li <[email protected]>

sbesson

Thanks all for the contributions. Merging as soon as Travis is green to have a reference point for the 0.1.3 version of the spec

sbesson · 2020-07-17T12:01:43Z

spec.md

-                ├── .zattrs   # "color".
-                │
-                └── t.c.z.y.x # Chunks as above.
+                ├── .zarray   # Each mask itself is a 5D array matching the highest resolution


@jburel can you file an issue about the mask resolutions?

joshmoore · 2020-08-18T19:30:14Z

#55 (comment) Discussed this morning with @will-moore and @manics, we need to indicate the resolution the mask corresponds to.
I have been looking at large images (bdv from IDR)

hmm... @jburel: compare https://gist.github.com/jni/62e07ddd135dbb107278bc04c0f9a8e7 and let me know what you think.

Initial draft of mask metadata

b24c65b

manics reviewed Jul 7, 2020

View reviewed changes

spec.md Show resolved Hide resolved

spec.md Show resolved Hide resolved

spec.md Outdated Show resolved Hide resolved

spec.md Show resolved Hide resolved

spec.md Outdated Show resolved Hide resolved

spec.md Outdated Show resolved Hide resolved

joshmoore and others added 4 commits July 7, 2020 16:51

Clarify RGBA color

bc54cce

Co-authored-by: Simon Li <[email protected]>

Clarify "masks" metadata

cad78f6

Co-authored-by: Simon Li <[email protected]>

Clarify array names

a4f61cd

Remove extraneous text

60ee51a

manics reviewed Jul 7, 2020

View reviewed changes

manics reviewed Jul 8, 2020

View reviewed changes

spec.md Show resolved Hide resolved

will-moore reviewed Jul 9, 2020

View reviewed changes

spec.md Outdated Show resolved Hide resolved

will-moore reviewed Jul 9, 2020

View reviewed changes

joshmoore mentioned this pull request Jul 10, 2020

Support opening of masks ome/ome-zarr-py#30

Merged

joshmoore changed the title ~~Initial draft of mask metadata~~ First version of mask metadata Jul 10, 2020

Drop groups and specify highest-resolution

c15b68b

Clarify mask dimensions

228fbff

Co-authored-by: Simon Li <[email protected]>

jburel reviewed Jul 10, 2020

View reviewed changes

mtbc reviewed Jul 13, 2020

View reviewed changes

spec.md Show resolved Hide resolved

mtbc mentioned this pull request Jul 16, 2020

Mask spec proposal #49

Closed

Add optional image key

6add680

Co-authored-by: Simon Li <[email protected]>

sbesson approved these changes Jul 17, 2020

View reviewed changes

sbesson merged commit e6ccb16 into ome:master Jul 17, 2020

joshmoore deleted the masks-etc branch July 23, 2020 16:23

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

First version of mask metadata #55

First version of mask metadata #55

joshmoore commented Jul 7, 2020 •

edited by sbesson

Loading

joshmoore commented Jul 7, 2020

manics left a comment

manics Jul 7, 2020 •

edited

Loading

sbesson Jul 23, 2020

will-moore Jul 9, 2020

joshmoore Jul 10, 2020

will-moore Jul 14, 2020

jburel commented Jul 9, 2020 •

edited

Loading

joshmoore commented Jul 10, 2020

joshmoore commented Jul 10, 2020

jburel Jul 10, 2020 •

edited

Loading

joshmoore Jul 10, 2020

jburel Jul 10, 2020

sbesson Jul 17, 2020

mtbc commented Jul 15, 2020

sbesson left a comment

sbesson Jul 17, 2020

joshmoore commented Aug 18, 2020

+### "image"
+The `image` key is an optional dictionary which contains information on the image the mask is associated with.
+If included it must include a key `array` whose value that is either:
+- A relative path to a Zarr image array, for example:
+    ```
+    {
+      "image": {
+        "array": "../../0"
+      }
+    }
+    ```
+- A URL to a Zarr image array (use this if the mask is stored seperately from the image Zarr), for example:
+    ```
+    {
+      "image": {
+        "array": "https://s3.embassy.ebi.ac.uk/idr/zarr/v0.1/6001240.zarr/0"
+      }
+    }
+    ```


		### "color"

		The `color` key defines an image that is "labeled", i.e. every unique value in the image

First version of mask metadata #55

First version of mask metadata #55

Conversation

joshmoore commented Jul 7, 2020 • edited by sbesson Loading

joshmoore commented Jul 7, 2020

manics left a comment

Choose a reason for hiding this comment

manics Jul 7, 2020 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

jburel commented Jul 9, 2020 • edited Loading

joshmoore commented Jul 10, 2020

joshmoore commented Jul 10, 2020

jburel Jul 10, 2020 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

mtbc commented Jul 15, 2020

sbesson left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

joshmoore commented Aug 18, 2020

joshmoore commented Jul 7, 2020 •

edited by sbesson

Loading

manics Jul 7, 2020 •

edited

Loading

jburel commented Jul 9, 2020 •

edited

Loading

jburel Jul 10, 2020 •

edited

Loading