Merge pull request #130 from Open-Model-Initiative/122-create-hdr-ima…

…ge-moderation-ui

Initial HDR image upload and moderation work + setup fixes

.env.template

@@ -41,3 +41,7 @@ OAUTH2_REDIRECT_PATH=docs
 ## Hugging Face
 HF_TOKEN=your_access_token
 HF_HDR_DATASET_NAME=openmodelinitiative/hdr-submissions
+
+## Docker compose variables
+NODE_ENV=development
+UPLOAD_DIR=./uploads

.gitignore

@@ -82,3 +82,6 @@ models_cache/
 
 # Embedding models
 **/models--**
+
+# Local uploads
+uploads/

GETTING_STARTED.md

@@ -54,16 +54,19 @@ To setup your environment run `task setup`
 
 Along with installing the necessary components, this will copy the `.env.template` setup into a `.env` file.
 
-You can update this file if you'd like to change any of the default values for your development environment.
+You can update this file if you'd like to change any of the default values for your development environment. You will at minimum need to provide a valid ID and SECRET for at least one authentication method.
 
+You will also need to manually copy the `.env.template` file in `modules/odr_frontend` to a `.env` file in the same folder, and make matching updates.
 
 # Running The Local Environment
 
-To start your virtual environment, run `task activate-venv`
+To start your virtual environment, run `task activate-venv` or `source ./venv/bin/activate`
 
-Then to start all the required services, run `task start-all`
+Then to start all the required services, run `task dev`
 
-This will spin up the required front end, api, and database components.
+This will spin up the required front end, api, and database containers.
+
+Before the site will 'work' you will need to run `task data:migrate` to create the required data and tables
 
 By default, you will have the following:
 
@@ -74,14 +77,23 @@ By default, you will have the following:
     - Interactive OpenAPI documentation at: http://localhost:31100/docs
     - Redoc documentation at: http://localhost:31100/redoc
 - A PGAdmin interface at: http://localhost:35050
+    To connect to the database, open http://localhost:35050/browser/
+
+    Sign in with your PGADMIN_DEFAULT_EMAIL and PGADMIN_DEFAULT_PASSWORD from your .env file
+
+    On the left, right click servers, and click Register > Server
+
+    Name the DB something like omi-database, then switch to the connection tab
+
+    The host name/address will be 'postgres', the username will be your .env POSTGRES_USER, and the password will be your .env POSTGRES_PASSWORD
 
 # Running Tests
 
 Before committing any changes, ensure you run all tests to be sure existing behaviour continues to work. If any tests require updates to pass with your changes, ensure you update them as well.
 
-To run all tests, run `task test-all`
+To run all tests, run `task test-all`. Note this is not working at the moment and needs some work but some tests are run via the pipeline automatically.
 
-(Todo: Add instructions for how to run the diffferent types individual as well to check work under development.)
+(Todo: Add instructions for how to run the different types individual as well to check work under development.)
 
 # Committing Changes
 
@@ -93,5 +105,3 @@ Before committing any changes, remember DCO and sign offs! If any commits are no
 To stop the environment, start by using ctrl+c (or cmd+c) to stop the current processes.
 
 Then use `task stop-all` to full stop the environment.
-
-If you want to remove the docker volumes used by the database, run `task db:teardown`

Taskfile.yml

@@ -4,7 +4,6 @@ version: '3'
 
 includes:
   api: ./modules/odr_api/Taskfile.api.yml
-  db: ./modules/odr_database/Taskfile.yml
   core: ./modules/odr_core/Taskfile.yml
   data: ./modules/odr_datamodel/Taskfile.yml
   frontend: ./modules/odr_frontend/Taskfile.yml
@@ -34,7 +33,7 @@ tasks:
         platforms: [linux, darwin]
       - cmd: docker network inspect omi-network >/dev/null 2>&1 || docker network create omi-network
       - cmd: task setup-venv
-      - cmd: task frontend:install
+      - cmd: task frontend:install-dev
 
   setup-venv:
     desc: Setup the virtual environment
@@ -74,7 +73,6 @@ tasks:
       - |
         task core:test &
         task core:test-db &
-        task db:test-postgres &
         task api-test:test-all &
         task monitoring:test &
         wait
@@ -87,10 +85,9 @@ tasks:
       - docker compose logs -f
 
   smoke_tests:
-    desc: Run smoke tests for both API and PostgreSQL
+    desc: Run smoke tests for the API
     cmds:
       - task api-test:health-check
-      - task db:test-postgres
 
   build-all:
     desc: Build all services

docker-compose.yml

@@ -70,11 +70,13 @@ services:
       - ./modules/odr_frontend/docker/entrypoint.sh:/app/docker/entrypoint.sh
       - odr_frontend_node_modules:/app/node_modules
       - odr_frontend_pnpm_store:/app/.pnpm-store
+      - ${UPLOAD_DIR:-./uploads}:/app/uploads
     ports:
       - "5173:5173"
     environment:
-      NODE_ENV: development
+      NODE_ENV: ${NODE_ENV:-development}
       API_SERVICE_URL: http://odr-api:31100/api/v1
+      UPLOAD_DIR: /app/uploads
     env_file:
       - ./modules/odr_frontend/.env
     networks:

modules/odr_api/odr_api/api/endpoints/image.py

@@ -14,6 +14,7 @@
 from typing import Any
 import base64
 import traceback
+import json
 import subprocess
 import tempfile
 import os
@@ -72,21 +73,6 @@ def calculate_entropy(tensor: torch.Tensor):
 
 
 # Helper functions for HDR metadata and preview conversion
-def extract_metadata(image_bytes: bytes) -> Dict:
-    metadata = {}
-    try:
-        with Image.open(BytesIO(image_bytes)) as img:
-            exif_data = img.getexif()
-            if exif_data:
-                for tag_id, value in exif_data.items():
-                    tag = TAGS.get(tag_id, tag_id)
-                    metadata[tag] = value
-        print('Metadata extracted from image file')
-    except UnidentifiedImageError:
-        print('Could not extract metadata from image')
-    return metadata
-
-
 def convert_ifd_rational(value):
     if isinstance(value, IFDRational):
         return float(value)
@@ -95,16 +81,40 @@ def convert_ifd_rational(value):
     return value
 
 
-def get_desired_metadata(metadata: Dict) -> Dict[str, Any]:
-    important_keys = ['Make', 'Model', 'BitsPerSample', 'BaselineExposure', 'LinearResponseLimit', 'ImageWidth', 'ImageLength', 'DateTime']
-    result = {key: convert_ifd_rational(metadata.get(key)) for key in important_keys if key in metadata}
+def get_metadata_with_exiftool(image_bytes: bytes) -> Dict[str, Any]:
+    with tempfile.NamedTemporaryFile(delete=False, suffix='.dng') as temp_input_file:
+        temp_input_file.write(image_bytes)
+        temp_input_filename = temp_input_file.name
+
+    try:
+        tags = [
+            '-Make', '-Model', '-BitsPerSample', '-BaselineExposure',
+            '-LinearResponseLimit', '-ImageWidth', '-ImageHeight',
+            '-DNGVersion'
+        ]
+
+        result = subprocess.run(
+            ['exiftool', '-j'] + tags + [temp_input_filename],
+            capture_output=True, text=True, check=True
+        )
+
+        metadata = json.loads(result.stdout)[0]
 
-    if 'DNGVersion' in metadata:
-        dng_version = metadata['DNGVersion']
-        version_string = '.'.join(str(b) for b in dng_version)
-        result['DNGVersion'] = version_string
+        if 'BitsPerSample' in metadata:
+            metadata['BitsPerSample'] = int(metadata['BitsPerSample'])
+        if 'BaselineExposure' in metadata:
+            metadata['BaselineExposure'] = float(metadata['BaselineExposure'])
+        if 'LinearResponseLimit' in metadata:
+            metadata['LinearResponseLimit'] = float(metadata['LinearResponseLimit'])
+        if 'ImageWidth' in metadata:
+            metadata['ImageWidth'] = int(metadata['ImageWidth'])
+        if 'ImageHeight' in metadata:
+            metadata['ImageHeight'] = int(metadata['ImageHeight'])
 
-    return result
+        return metadata
+
+    finally:
+        os.remove(temp_input_filename)
 
 
 def convert_dng_to_jpg(dng_bytes: bytes) -> bytes:
@@ -162,27 +172,19 @@ async def create_jpg_preview(file: UploadFile = File(...)):
         raise HTTPException(status_code=400, detail=str(e))
 
 
-# Endpoint for metadata retrieval
 @router.post("/image/metadata")
 async def get_image_metadata(file: UploadFile = File(...)):
     try:
         contents = await file.read()
-        metadata = extract_metadata(contents)
+        metadata = get_metadata_with_exiftool(contents)
 
-        important_metadata = get_desired_metadata(metadata)
-
-        return important_metadata
-    except ValueError as e:
-        raise HTTPException(status_code=400, detail=str(e))
+        return metadata
+    except subprocess.CalledProcessError as e:
+        raise HTTPException(status_code=400, detail=f"Error processing image with exiftool: {str(e)}")
+    except json.JSONDecodeError as e:
+        raise HTTPException(status_code=500, detail=f"Error parsing exiftool output: {str(e)}")
     except Exception as e:
-        raise HTTPException(status_code=500, detail=str(e))
-
-
-# For debugging
-# docker cp $(docker ps --filter name=omi-postgres-odr-api -q):./app/cleaned_image_exiftool.dng ./cleaned_image_exiftool.dng
-# def save_image_locally(image_bytes: bytes, filename: str):
-#     with open(filename, 'wb') as f:
-#         f.write(image_bytes)
+        raise HTTPException(status_code=500, detail=f"Unexpected error: {str(e)}")
 
 
 def remove_metadata_with_exiftool(input_bytes):
@@ -252,9 +254,6 @@ async def clean_image_metadata(file: UploadFile = File(...)):
         contents = await file.read()
 
         cleaned_image_bytes = remove_metadata_with_exiftool(contents)
-        # For debugging
-        # save_image_locally(cleaned_image_bytes, 'cleaned_image_exiftool.dng')
-
         encoded_image = base64.b64encode(cleaned_image_bytes).decode('utf-8')
 
         return {

modules/odr_api/requirements.txt

@@ -1,6 +1,6 @@
 uvicorn==0.30.5
 fastapi==0.115.4
-pydantic[email]==2.8.2
+pydantic[email]==2.10.4
 sqlalchemy==2.0.31
 pytest==8.3.2
 starlette==0.40.0

modules/odr_caption/requirements.txt

@@ -4,7 +4,7 @@ qwen-vl-utils[decord]
 fastapi
 click
 openai
-pydantic
+pydantic==2.10.4
 pillow
 python-dotenv
 torch

modules/odr_core/requirements.txt

@@ -1,6 +1,6 @@
 alembic==1.13.2
 sqlalchemy==2.0.31
-pydantic==2.8.2
+pydantic==2.10.4
 pydantic-settings==2.4.0
 pytest==8.3.2
 loguru==0.7.2