Enhance Multilingual Support with Chinese and German Label Recognition

label_studio_ml/examples/paddleocr/.dockerignore

@@ -0,0 +1,11 @@
+Dockerfile
+README.md
+*.pyc
+*.pyo
+*.pyd
+__pycache__
+.pytest_cache
+.idea
+docker-compose.yml
+data
+logs

label_studio_ml/examples/paddleocr/Dockerfile

@@ -0,0 +1,32 @@
+FROM python:3.10
+
+WORKDIR /tmp
+COPY requirements.txt .
+
+ENV PYTHONUNBUFFERED=True \
+    PORT=${PORT:-9090} \
+    PIP_CACHE_DIR=/.cache
+
+# Installing libGL
+RUN apt-get update && apt-get install -y \
+    libgl1-mesa-dev
+
+RUN apt-get install -y poppler-utils libpoppler-cpp-dev libglib2.0-0
+
+
+RUN pip install --upgrade pip \
+ && pip install -r requirements.txt
+
+
+#COPY uwsgi.ini /etc/uwsgi/
+COPY supervisord.conf /etc/supervisor/conf.d/
+
+WORKDIR /app
+
+COPY * /app/
+
+EXPOSE 9090
+
+CMD ["/usr/local/bin/supervisord", \
+     "-c", \
+     "/etc/supervisor/conf.d/supervisord.conf"]

label_studio_ml/examples/paddleocr/README.md

@@ -0,0 +1,137 @@
+## Interactive BBOX OCR using PaddleOCR
+Using an OCR engine for Interactive ML-Assisted Labelling, this functionality
+can speed up annotation for layout detection, classification and recognition
+models.
+
+PaddleOCR is used for OCR but minimal adaptation is needed to connect other OCR
+engines or models.
+
+Tested againt Label Studio 1.10.1, with basic support for both Label Studio
+Local File Storage and S3-compatible storage, with a example data storage with
+Minio.
+
+### Setup process
+0. Download and install Docker with Docker Compose. For MacOS and Windows users,
+   we suggest using Docker Desktop. You will also need to have git installed.
+
+1. Launch LabelStudio.
+
+   ```
+   docker run -it \
+      -p 8080:8080 \
+      -v `pwd`/mydata:/label-studio/data \
+      heartexlabs/label-studio:latest
+   ```
+
+   Optionally, you may enable local file serving in Label Studio
+
+   ```
+   docker run -it \
+      -p 8080:8080 \
+      -v `pwd`/mydata:/label-studio/data \
+      --env LABEL_STUDIO_LOCAL_FILES_SERVING_ENABLED=true \
+      --env LABEL_STUDIO_LOCAL_FILES_DOCUMENT_ROOT=/label-studio/data/images \
+      heartexlabs/label-studio:latest
+   ```
+   If you're using local file serving, be sure to get a copy of the API token from
+   Label Studio to connect the model.
+
+2. Create a new project for PaddleOCR OCR. In the project **Settings** set up the **Labeling Interface**.
+
+   Fill in the following template code. It's important to specify `smart="true"` in RectangleLabels.
+   ```
+   <View>    
+      <Image name="image" value="$ocr" zoom="true" zoomControl="false"
+            rotateControl="true" width="100%" height="100%"
+            maxHeight="auto" maxWidth="auto"/>
+
+      <RectangleLabels name="bbox" toName="image" strokeWidth="1" smart="true">
+         <Label value="Label1" background="green"/>
+         <Label value="Label2" background="blue"/>
+         <Label value="Label3" background="red"/>
+      </RectangleLabels>
+
+      <TextArea name="transcription" toName="image" 
+      editable="true" perRegion="true" required="false" 
+      maxSubmissions="1" rows="5" placeholder="Recognized Text" 
+      displayMode="region-list"/>
+   </View>
+   ```
+
+3. Download the Label Studio Machine Learning backend backend repository.
+   ```
+   git clone https://github.com/humansignal/label-studio-ml-backend
+   cd label-studio-ml-backend/label_studio_ml/examples/paddleocr
+   ```
+
+4. Configure the backend and the Minio server by editing the `example.env` file. If you opted to use Label Studio
+   Local File Storage, be sure to set the `LABEL_STUDIO_HOST` and `LABEL_STUDIO_ACCESS_TOKEN` variables. If you're
+   using the Minio storage example, set the `MINIO_ROOT_USER` AND `MINIO_ROOT_PASSWORD` variables, and make the 
+   `AWS_ACCESS_KEY_ID` AND `AWS_SECRET_ACCESS_KEY` variables equal to those values. You may optionally connect to your
+   own AWS cloud storage by setting those variables. Note that you may need to make additional software changes to the
+   `paddleocr_ch.py` file to match your particular infrastructure configuration.
+
+   ```
+   LABEL_STUDIO_HOST=http://host.docker.internal:8080
+   LABEL_STUDIO_ACCESS_TOKEN=<optional token for local file access>
+
+   AWS_ACCESS_KEY_ID=<set to MINIO_ROOT_USER for minio example>
+   AWS_SECRET_ACCESS_KEY=<set to MINIO_ROOT_PASSWORD for minio example>
+   AWS_ENDPOINT=http://host.docker.internal:9000
+
+   MINIO_ROOT_USER=<username>
+   MINIO_ROOT_PASSWORD=<password>
+   MINIO_API_CORS_ALLOW_ORIGIN=*
+   ```
+
+5. Start the PaddleOCR and minio servers.
+
+   ```
+   # build image
+   sudo docker build . -t paddleocr-backend:latest
+   # or
+   sudo docker compose build .
+
+   # only start paddleocr-backend
+   sudo docker compose up paddleocr-backend
+   # or and start with minio
+   docker compose up -d
+   # shutdown container
+   docker compose down 
+
+
+   ```
+
+6. Upload tasks.
+
+   If you're using the Label Studio Local File Storage option, upload images
+   directly to Label Studio using the Label Studio interface.
+
+   If you're using minio for task storage, log into the minio control panel at
+   `http://localhost:9001`. Create a new bucket, making a note of the name, and
+   upload your tasks to minio. Set the visibility of the tasks to be public.
+   Furtner configuration of your cloud storage is beyond the scope of this
+   tutorial, and you will want to configure your storage according to your
+   particular needs. 
+
+7. If using minio, In the project **Settings**, set up the **Cloud storage**.
+
+   Add your source S3 storage by connecting to the S3 Endpoint
+   `http://host.docker.internal:9000`, using the bucket name from the previous
+   step, and Access Key ID and Secret Access Key as configured in the previous
+   steps. For the minio example, uncheck **Use pre-signed URLS**. Check the
+   connection and save the storage.
+
+8. Open the **Machine Learning** settings and click **Add Model**.
+
+   Add the URL `http://host.docker.internal:9090` and save the model as an ML backend.
+
+9. To use this functionality, activate `Auto-Annotation` and use `Autotdetect` rectangle for drawing boxes
+
+Example below :
+
+![ls_demo_ocr](https://github.com/HumanSignal/label-studio-ml-backend/assets/1549611/fcc44c8b-12fd-495c-b0b4-9d5c0ceb2ed2)
+
+Reference links : 
+- https://labelstud.io/blog/Improve-OCR-quality-with-Tesseract-and-Label-Studio.html
+- https://labelstud.io/blog/release-130.html

label_studio_ml/examples/paddleocr/_wsgi.py

@@ -0,0 +1,129 @@
+import json
+import os
+import argparse
+import logging
+import logging.config
+
+logging.config.dictConfig({
+  "version": 1,
+  "formatters": {
+    "standard": {
+      "format": "[%(asctime)s] [%(levelname)s] [%(name)s::%(funcName)s::%(lineno)d] %(message)s"
+    }
+  },
+  "handlers": {
+    "console": {
+      "class": "logging.StreamHandler",
+      "level": "DEBUG",
+      "stream": "ext://sys.stdout",
+      "formatter": "standard"
+    }
+  },
+  "root": {
+    "level": "ERROR",
+    "handlers": [
+      "console"
+    ],
+    "propagate": True
+  }
+})
+
+from label_studio_ml.api import init_app
+from paddleocr_ch import paddleocrLabeling
+
+
+_DEFAULT_CONFIG_PATH = os.path.join(os.path.dirname(__file__), 'config.json')
+
+
+def get_kwargs_from_config(config_path=_DEFAULT_CONFIG_PATH):
+    if not os.path.exists(config_path):
+        return dict()
+    with open(config_path) as f:
+        config = json.load(f)
+    assert isinstance(config, dict)
+    return config
+
+
+if __name__ == "__main__":
+    parser = argparse.ArgumentParser(description='Label studio')
+    parser.add_argument(
+        '-p', '--port', dest='port', type=int, default=9090,
+        help='Server port')
+    parser.add_argument(
+        '--host', dest='host', type=str, default='0.0.0.0',
+        help='Server host')
+    parser.add_argument(
+        '--kwargs', '--with', dest='kwargs', metavar='KEY=VAL', nargs='+', type=lambda kv: kv.split('='),
+        help='Additional LabelStudioMLBase model initialization kwargs')
+    parser.add_argument(
+        '-d', '--debug', dest='debug', action='store_true',
+        help='Switch debug mode')
+    parser.add_argument(
+        '--log-level', dest='log_level', choices=['DEBUG', 'INFO', 'WARNING', 'ERROR'], default=None,
+        help='Logging level')
+    parser.add_argument(
+        '--model-dir', dest='model_dir', default=os.path.dirname(__file__),
+        help='Directory where models are stored (relative to the project directory)')
+    parser.add_argument(
+        '--check', dest='check', action='store_true',
+        help='Validate model instance before launching server')
+
+    parser.add_argument('--basic-auth-user',
+                        default=os.environ.get('ML_SERVER_BASIC_AUTH_USER', None),
+                        help='Basic auth user')
+
+    parser.add_argument('--basic-auth-pass',
+                        default=os.environ.get('ML_SERVER_BASIC_AUTH_PASS', None),
+                        help='Basic auth pass')
+
+    args = parser.parse_args()
+
+    # setup logging level
+    if args.log_level:
+        logging.root.setLevel(args.log_level)
+
+    def isfloat(value):
+        try:
+            float(value)
+            return True
+        except ValueError:
+            return False
+
+    def parse_kwargs():
+        param = dict()
+        for k, v in args.kwargs:
+            if v.isdigit():
+                param[k] = int(v)
+            elif v == 'True' or v == 'true':
+                param[k] = True
+            elif v == 'False' or v == 'False':
+                param[k] = False
+            elif isfloat(v):
+                param[k] = float(v)
+            else:
+                param[k] = v
+        return param
+
+    kwargs = get_kwargs_from_config()
+
+    if args.kwargs:
+        kwargs.update(parse_kwargs())
+
+    if args.check:
+        print('Check "' + paddleocrLabeling.__name__ + '" instance creation..')
+        model = paddleocrLabeling(**kwargs)
+
+    app = init_app(
+        model_class=paddleocrLabeling,
+        basic_auth_user=args.basic_auth_user, 
+        basic_auth_pass=args.basic_auth_pass,
+        **kwargs
+    )
+
+    app.run(host=args.host, port=args.port, debug=args.debug)
+
+else:
+    # for uWSGI use
+    app = init_app(
+        model_class=paddleocrLabeling
+    )

label_studio_ml/examples/paddleocr/docker-compose.yml

@@ -0,0 +1,27 @@
+version: "3.8"
+
+services:
+
+  paddleocr-backend:
+    container_name: paddleocr-backend
+    build: .
+    env_file:
+      - example.env
+    ports:
+      - 9090:9090
+    volumes:
+      - "./data/server:/data"
+      - "./logs:/tmp"
+
+  minio:
+    container_name: minio
+    image: bitnami/minio:latest
+    env_file:
+      - example.env
+    ports:
+      - 9003:9000
+      - 9004:9001
+    volumes:
+      - ".data/storage:/data"
+    command: server /data --console-address ":9001"
+

label_studio_ml/examples/paddleocr/example.env

@@ -0,0 +1,11 @@
+LABEL_STUDIO_HOST=http://10.33.1.37:8089
+LABEL_STUDIO_ACCESS_TOKEN=9075ad7f2794f7d670a752dcf0eb1b6d177d08ff
+
+AWS_ACCESS_KEY_ID=admin
+AWS_SECRET_ACCESS_KEY=9075ad7f2
+AWS_ENDPOINT=http://10.33.1.37:9000
+
+MINIO_ROOT_USER=admin
+MINIO_ROOT_PASSWORD=9075ad7f2
+MINIO_API_CORS_ALLOW_ORIGIN=*
+MINIO_OPTS="--console-address :9001"