Add COG processing for large images

isic/ingest/models/accession.py

@@ -3,6 +3,8 @@
 import io
 import logging
 from mimetypes import guess_type
+from pathlib import Path
+import subprocess
 import tempfile
 from uuid import uuid4
 
@@ -33,6 +35,10 @@
 
 logger = logging.getLogger(__name__)
 
+# The number of square pixels at which an image is stored as a
+# cloud optimized geotiff.
+IMAGE_COG_THRESHOLD: int = 100_000_000
+
 
 class Approx(Transform):
     lookup_name = "approx"
@@ -307,6 +313,11 @@ def __str__(self) -> str:
         RemappedField("rcm_case_id", "rcm_case", "private_rcm_case_id", RcmCase),
     ]
 
+    @property
+    def requires_cog(self):
+        if self.width and self.height:
+            return self.width * self.height > IMAGE_COG_THRESHOLD
+
     @property
     def published(self):
         return hasattr(self, "image")
@@ -333,6 +344,80 @@ def metadata_keys():
             field.name for field in Accession._meta.fields if hasattr(AccessionMetadata, field.name)
         ]
 
+    def _save_blob_from_pil_image(self, img: PIL.Image.Image):
+        if img.height * img.width < IMAGE_COG_THRESHOLD:
+            with tempfile.SpooledTemporaryFile() as stripped_blob_stream:
+                img.save(stripped_blob_stream, format="JPEG")
+
+                stripped_blob_stream.seek(0, io.SEEK_END)
+                stripped_blob_size = stripped_blob_stream.tell()
+                stripped_blob_stream.seek(0)
+
+                self.blob_name = f"{uuid4()}.jpg"
+                self.blob = InMemoryUploadedFile(
+                    file=stripped_blob_stream,
+                    field_name=None,
+                    name=self.blob_name,
+                    content_type="image/jpeg",
+                    size=stripped_blob_size,
+                    charset=None,
+                )
+                self.blob_size = stripped_blob_size
+                self.height = img.height
+                self.width = img.width
+
+                self.save(update_fields=["blob_name", "blob", "blob_size", "height", "width"])
+        else:
+            from rio_cogeo import cog_translate
+
+            with (
+                tempfile.NamedTemporaryFile() as stripped_blob_stream,
+                tempfile.NamedTemporaryFile(delete=False) as cog_stream,
+            ):
+                img.save(stripped_blob_stream, format="PNG")
+
+                cog_translate(
+                    source=stripped_blob_stream.name,
+                    dst_path=cog_stream.name,
+                    dst_kwargs={
+                        "driver": "GTiff",
+                        "interleave": "pixel",
+                        "tiled": True,
+                        "blockxsize": 512,
+                        "blockysize": 512,
+                        "compress": "DEFLATE",
+                        "BIGTIFF": "IF_SAFER",
+                    },
+                    # to do
+                    quiet=False,
+                )
+
+            # run exiftool -ModelTransform= on the file via subprocess
+            # subprocess.run(
+            #     [  # noqa: S607, S603
+            #         "exiftool",
+            #         "-ModelTransform=",
+            #         cog_stream.name,
+            #     ],
+            #     check=True,
+            # )
+
+            self.blob_size = Path(cog_stream.name).stat().st_size
+            with Path(cog_stream.name).open("rb") as cog_stream:
+                self.blob_name = f"{uuid4()}.tif"
+                self.blob = InMemoryUploadedFile(
+                    file=cog_stream,
+                    field_name=None,
+                    name=self.blob_name,
+                    content_type="image/tiff",
+                    size=self.blob_size,
+                    charset=None,
+                )
+                self.height = img.height
+                self.width = img.width
+
+                self.save(update_fields=["blob_name", "blob", "blob_size", "height", "width"])
+
     def generate_blob(self):
         """
         Generate `blob` and set `blob_size`, `height`, `width`.
@@ -368,29 +453,10 @@ def generate_blob(self):
                 raise
 
             # Strip any alpha channel
-            img = img.convert("RGB")
-
-            with tempfile.SpooledTemporaryFile() as stripped_blob_stream:
-                img.save(stripped_blob_stream, format="JPEG")
-
-                stripped_blob_stream.seek(0, io.SEEK_END)
-                stripped_blob_size = stripped_blob_stream.tell()
-                stripped_blob_stream.seek(0)
+            if blob_mime_type == "image/jpeg":
+                img = img.convert("RGB")
 
-                self.blob_name = f"{uuid4()}.jpg"
-                self.blob = InMemoryUploadedFile(
-                    file=stripped_blob_stream,
-                    field_name=None,
-                    name=self.blob_name,
-                    content_type="image/jpeg",
-                    size=stripped_blob_size,
-                    charset=None,
-                )
-                self.blob_size = stripped_blob_size
-                self.height = img.height
-                self.width = img.width
-
-                self.save(update_fields=["blob_name", "blob", "blob_size", "height", "width"])
+            self._save_blob_from_pil_image(img)
 
             self.generate_thumbnail()
 
@@ -409,8 +475,9 @@ def generate_blob(self):
             self.status = AccessionStatus.SUCCEEDED
             self.save(update_fields=["status"])
 
+    # TODO
     def generate_thumbnail(self) -> None:
-        with self.blob.open() as blob_stream:
+        with Accession.objects.order_by("created").first().blob.open() as blob_stream:
             img: PIL.Image.Image = PIL.Image.open(blob_stream)
             # Load the image so the stream can be closed
             img.load()

isic/ingest/templates/ingest/accession_cog_viewer.html

@@ -1,15 +1,15 @@
 {% extends 'core/base.html' %}
-{% load static %}
 
+{% load static %}
 
 {% block content %}
+    <div id="image" style="width:100%; height:800px;"></div>
 
-<div id="image" style="width:100%; height:400px;"></div>
-
-<script type="module" src="{% static 'core/dist/cog.js' %}"></script>
-
-<script type="text/javascript">
-    initializeCogViewer('{{ accession.blob.url|safe }}');
-</script>
+    <script src="{% static 'core/dist/cog.js' %}"></script>
 
+    <script type="text/javascript">
+        initializeCogViewer('{{ accession.blob.url|safe }}',
+            {{ accession.width }},
+            {{ accession.height }});
+    </script>
 {% endblock %}

isic/ingest/templates/ingest/partials/accession.html

@@ -6,7 +6,14 @@
       {% if PLACEHOLDER_IMAGES %}
         <img @mouseenter="hovered = true" @click="open = true" src="https://picsum.photos/seed/{{accession.id}}/256">
       {% else %}
-        <img @mouseenter="hovered = true" @click="open = true" src="{{ accession.thumbnail_256.url }}" />
+        {% if accession.requires_cog %}
+          <a href="{% url 'accession-cog-viewer' accession.id %}">
+            <img src="{{ accession.thumbnail_256.url }}" />
+          </a>
+        {% else %}
+          <img @mouseenter="hovered = true" @click="open = true" src="{{ accession.thumbnail_256.url }}" />
+        {% endif %}
+
       {% endif %}
     </a>
 

isic/ingest/views/accession.py

@@ -1,23 +1,7 @@
-from collections import defaultdict
-import itertools
-
-from django.contrib import messages
 from django.contrib.admin.views.decorators import staff_member_required
-from django.contrib.humanize.templatetags.humanize import intcomma
-from django.core.exceptions import ValidationError
-from django.core.paginator import Paginator
-from django.db.models import Count, Prefetch, Q
-from django.http import HttpResponseRedirect
 from django.shortcuts import get_object_or_404, render
-from django.urls.base import reverse
 
-from isic.core.permissions import needs_object_permission
-from isic.ingest.forms import MergeCohortForm
-from isic.ingest.models import Cohort
-from isic.ingest.models.accession import Accession, AccessionStatus
-from isic.ingest.models.contributor import Contributor
-from isic.ingest.services.cohort import cohort_merge, cohort_publish_initialize
-from isic.ingest.views import make_breadcrumbs
+from isic.ingest.models.accession import Accession
 
 
 @staff_member_required

node-src/cog.mjs

@@ -0,0 +1,46 @@
+import GeoTIFF from 'ol/source/GeoTIFF.js';
+import Map from 'ol/Map.js';
+import Projection from 'ol/proj/Projection.js';
+import TileLayer from 'ol/layer/WebGLTile.js';
+import View from 'ol/View.js';
+import { getCenter } from 'ol/extent.js';
+
+// TODO: figure out how to override getOrigin, see
+// https://github.com/geotiffjs/geotiff.js/blob/da936684e30ef994f1d0d1e2da844b0e9a6c3cd0/src/geotiffimage.js#L825.
+// This will obviate the need for exif stripping the model transformation tag.
+
+function initializeCogViewer(url, width, height) {
+    const extent = [0, 0, width, height];
+
+    const projection = new Projection({
+        code: 'custom',
+        units: 'pixels',
+        extent: extent,
+    });
+
+    const geotiff = new GeoTIFF({
+        sources: [
+            {
+                url: url,
+                nodata: 0,
+            },
+        ],
+    });
+
+    const map = new Map({
+        target: 'image',
+        layers: [
+            new TileLayer({
+                source: geotiff,
+            }),
+        ],
+        view: new View({
+            projection: projection,
+            center: getCenter(extent),
+            extent: extent,
+            zoom: 1,
+        }),
+    });
+}
+
+window.initializeCogViewer = initializeCogViewer;

package.json

@@ -31,9 +31,7 @@
     },
     "cog": {
       "source": "./node-src/cog.mjs",
-      "distDir": "./isic/core/static/core/dist",
-      "context": "browser",
-      "outputFormat": "global"
+      "distDir": "./isic/core/static/core/dist"
     }
   },
   "browserslist": "> 0.5%",
@@ -45,8 +43,5 @@
   },
   "dependencies": {
     "ol": "^9.2.4"
-  },
-  "resolutions": {
-    "ol/**/@petamoriken/float16": "3.4.7"
   }
-}
+}

setup.py

@@ -75,6 +75,7 @@
         "pyparsing",
         "python-magic",
         "requests",
+        "rio-cogeo",
         "sentry-sdk[pure_eval]",
         "tenacity",
         "zipfile-deflate64",

yarn.lock

@@ -790,7 +790,7 @@
     "@parcel/utils" "2.12.0"
     nullthrows "^1.1.1"
 
-"@petamoriken/float16@3.4.7", "@petamoriken/float16@^3.4.7":
+"@petamoriken/float16@^3.4.7":
   version "3.4.7"
   resolved "https://registry.yarnpkg.com/@petamoriken/float16/-/float16-3.4.7.tgz#63ce6cb698881bca0fc272807196d497615c95ea"
   integrity sha512-Mir0MAKxg5v6BUIg9SI5VAyrIa/3uptf7aPyvPhHNh0RMYMevrWbaLrsVSZ1f92C39hWd8v7GrjvvOT+WG5VUQ==