Merge pull request #7 from bioimage-io/web-ui

Web UI

.github/workflows/docker-publish.yml

@@ -4,9 +4,6 @@ on:
   push:
     branches:
       - main  # Trigger the workflow on pushes to the main branch
-  pull_request:
-    branches:
-      - main  # Trigger the workflow on pull requests to the main branch
 
 jobs:
   build-and-push:

Dockerfile

@@ -29,6 +29,9 @@ COPY ./bioimageio_colab/register_sam_service.py /app/register_sam_service.py
 # Change ownership of the application directory to the non-root user
 RUN chown -R bioimageio_colab:bioimageio_colab /app/
 
+# Add a build argument for cache invalidation
+ARG CACHEBUST=1
+
 # Fetch the Hypha server version and reinstall or upgrade hypha-rpc to the matching version
 RUN HYPHA_VERSION=$(curl -s https://hypha.aicell.io/assets/config.json | jq -r '.hypha_version') && \
     pip install --upgrade "hypha-rpc<=$HYPHA_VERSION"

build_and_push.sh

@@ -22,7 +22,10 @@ IMAGE_NAME=ghcr.io/${GITHUB_REPOSITORY}:latest
 # Log in to GHCR
 echo "$GHCR_PAT" | docker login ghcr.io -u "$GITHUB_ACTOR" --password-stdin
 
-# Build the Docker image using Docker Compose
+# Generate a dynamic CACHEBUST value (timestamp)
+export CACHEBUST=$(date +%Y%m%d%H%M%S)
+
+# Build the Docker image using Docker Compose with the CACHEBUST argument
 docker-compose build
 
 # Push the Docker image to GHCR

docker-compose.yml

@@ -6,7 +6,7 @@ services:
       context: .
       dockerfile: Dockerfile
       args:
-        SOURCE_LABEL: "https://github.com/bioimage-io/bioimageio-colab"
+        CACHEBUST: ${CACHEBUST}
     image: ghcr.io/bioimage-io/bioimageio-colab:latest
     env_file:
       - .env

docs/data-providing-service.py

@@ -0,0 +1,102 @@
+import os
+import json
+from typing import Tuple
+import time
+from functools import partial
+
+import numpy as np
+from hypha_rpc import connect_to_server
+from kaibu_utils import features_to_mask
+from tifffile import imread, imwrite
+
+
+def list_image_files(image_folder: str, supported_file_types: Tuple[str]):
+    return [f for f in os.listdir(image_folder) if f.endswith(supported_file_types)]
+
+
+def read_image(file_path: str):
+    image = imread(file_path)
+    if len(image.shape) == 3 and image.shape[0] == 3:
+        image = np.transpose(image, [1, 2, 0])
+    return image
+
+
+def get_random_image(image_folder: str, supported_file_types: Tuple[str]):
+    filenames = list_image_files(image_folder, supported_file_types)
+    r = np.random.randint(len(filenames) - 1)
+    file_name = filenames[r]
+    image = read_image(os.path.join(image_folder, file_name))
+    return (image, file_name.split(".")[0])
+
+
+def save_annotation(annotations_folder: str, image_name: str, features, image_shape):
+    mask = features_to_mask(features, image_shape)
+    n_image_masks = len(
+        [f for f in os.listdir(annotations_folder) if f.startswith(image_name)]
+    )
+    mask_name = os.path.join(
+        annotations_folder, f"{image_name}_mask_{n_image_masks + 1}.tif"
+    )
+    imwrite(mask_name, mask)
+
+
+def upload_image_to_s3():
+    """
+    Steps:
+    - Create a user prefix on S3
+    - Create a data and annotation prefix
+    - For every image:
+        - Load the image from the data folder into a numpy array
+        - Upload the image to the data prefix
+
+    Return:
+    - The user prefix
+
+    # TODO: register a data providing service on K8S cluster that uses the user prefix (get_random_image_s3, save_annotation_s3)
+    """
+    raise NotImplementedError
+
+async def register_service(
+    server_url: str,
+    token: str,
+    supported_file_types_json: str,
+    name: str,
+    description: str,
+):
+    # Define path to images and annotations
+    images_path = "/mnt"
+    annotations_path = "/mnt/annotations"
+
+    # Check if the images folder exists
+    if not os.path.isdir(images_path):
+        raise FileNotFoundError("Mounted images folder not found")
+
+    # Decode the JSON string to a Python tuple
+    supported_file_types = tuple(json.loads(supported_file_types_json))
+
+    # Connect to the server link
+    server = await connect_to_server({"server_url": server_url, "token": token})
+
+    # Register the service
+    svc = await server.register_service(
+        {
+            "name": name,
+            "description": description,
+            "id": "data-provider-" + str(int(time.time()*100)),
+            "type": "annotation-data-provider",
+            "config": {
+                "visibility": "public",  # TODO: make protected
+                "run_in_executor": True,
+            },
+            # Exposed functions:
+            # get a random image from the dataset
+            # returns the image as a numpy image
+            "get_random_image": partial(
+                get_random_image, images_path, supported_file_types
+            ),
+            # save the annotation mask
+            # pass the filename of the image, the new filename, the features and the image shape
+            "save_annotation": partial(save_annotation, annotations_path),
+        }
+    )
+    print(f"Service registered with ID: {svc['id']}")