Makes Measure Detection Great Again! Based on a clean 4-Tier architecture we contribute to a simplification of the process of "Measure Counting" (Vertaktung) done in modern music edition projects. AI powered, based on a modern technology stack and ready for you to use!
This application was generated using JHipster 7.4.0, you can find documentation and help at https://www.jhipster.tech/documentation-archive/v7.4.0.
-
Manage Projects and Editions with sheets of music
-
Automatic Measure Detection with the Measure Detector: https://github.com/OMR-Research/MeasureDetector
-
Provides a web-based front-end (AngularJS) where edition works can be managed.
-
Export into MEI / PDF with Annotations
-
User Management, Scheduling System, ...
This section guides you throw the installation of the application and its required dependencies.
-
doc/**: Contains all the required documentation generate with mkdocs which enables simple web rendering of the whole documentation contents
-
app/**: Contains the whole application
-
Clone the repository:
git clone https://github.com/lukas681/Measure-Detector-UI.git -
Make sure to have Docker installed. Use WSL 2 + Docker on WNX based systems. Then build the Measure Detector Docker container. Use the provided Makefile.
$ cd docker/measure-detector-docker $ make build-dockerOtherwise, just build the container manually:
$ cd measure-detector-docker # Build the image with the latest model $ docker build --no-cache --tag mdet:1 -t measure_detector . # Run in container (change port to `XXXX:8080` if needed): $ docker run --name mdet -d -p ${PORT}:8080 measure_detector -
(Optional) Test the REST interface:
curl --location --request POST 'localhost:8080/upload' --form 'image=@"path/to/example/image"'
You need to follow these steps, to set-up the application correctly:
- As the upload mechanism is not implemented optimally, yet, weshoudl increase the maximal receiving file length. (Affects the maximal size of a music document to be uploaded).
You can therefore free around 1/3 of your available (e.g. 5GB on a 16GB machine). To do so, open app/src/main/resources/config/application-xx.yml (dev, prod, xx) and modify the following line:
spring:
codec:
max-in-memory-size: 5GB
You can run the application inside a docker containre with the official jhipster docker images. therefore execute the following commands. Make sure to provide absolute paths as relative paths are not supported any more by docker.
cd measure-detector-ui/app
docker run --network="host" -v ${PWD}:/home/jhipster/app -p 8000:8000 --rm jhipster/jhipster .\gradlew bootRun --args='--spring.profiles.active=dev'
Voilâ, the application should already start ...
Note: --network="host" will embedd the host localhost inside the container, so that we can also reach the Measure Detector in the other container. Concerning security, this is not optimal and should be switched to a Virtual Network based approach (Docker Networking + Docker Compose)
in order to strat the application, call the gradle wrapper with the corresponding profile. It might be necessary, to also installed the openseadragon 2.4.8 type delarations manually to match those required by AnnotoriusJS.
cd app
./gradlew bootRun --args='--spring.profiles.active=dev'
# wait for finish
npm i --save-dev @types/[email protected]
npm start
--------------------------------------
Local: http://localhost:8080
External: http://192.168.2.148:8080
--------------------------------------
UI: http://localhost:3001
UI External: http://localhost:3001
--------------------------------------
Note: In most cases it is easier to use the development mode (uses a h2 file database). In case you want to deploy the application on a large scale level, consider using this prod profile. Currently, the productive mode is configured to use a local mysql instance. Observe the corresponding line in the application.yml:
liquibase:
contexts: prod
url: jdbc:mysql://localhost:3306/MeasureDetector?useUnicode=true&characterEncoding=utf8&useSSL=false&useLegacyDatetimeCode=false&serverTimezone=UTC&createDatabaseIfNotExist=true
In order to run against a local h2 database (Should also be sufficient as we do not store large data in the database so far...), just run in development mode.
You have access to the following interfaces acompanying the application (Standard Ports):
- Swagger: Login -> Administration/Api
- JobRunR (Tracks the background jobs) -> :8000
- Main Application -> 8080
- Measure Detector Listener: 8081
The following frameworks are used for building this application. In general, jhipster 7.4.0 was used to generate the backbone of the application
| Name | Version | Required |
|---|---|---|
| NodeJS | 14 | yes |
| Docker | yes | |
| Java | 14 | yes |
| TensorFlow | 1.13.1 | via gradle |
To build the final jar and optimize the MeasureDetector application for production, run:
./gradlew -Pprod clean bootJar
This will concatenate and minify the client CSS and JavaScript files. It will also modify index.html so it references these new files.
To ensure everything worked, run:
To launch your application's tests, run:
./gradlew test integrationTest jacocoTestReport
This function can be used to really debug all the necessary steps.
$ docker run ubuntu /bin/bash
apt update
apt install gradle
cd ~
git clone https://gitlab.lrz.de/ge82xib/measure-detector-ui
cd measure-detector-ui/app
chmod +x gradlew
./gradlew
Distributed under the MIT License. See LICENSE.md for more information.
Lukas Retschmeier - lukasretschmeier.de - [email protected]
Project Link: https://github.com/lukas681/Measure-Detector-UI
The Measure Detector DNN model was trained in this project https://github.com/OMR-Research/MeasureDetector/
The original repository was hosted at LRZ, but due to a better availability for the public moved to Github. The old repository can still be found here: