oleksii-suprun / meter-tracker Goto Github PK

At the moment each build uploads cache to storage:

store build cache
0.00s
21.83schange detected (content changed, file is created, or file is deleted):
/home/travis/.gradle/caches/4.10/fileContent/fileContent.lock
/home/travis/.gradle/caches/4.10/fileHashes/fileHashes.bin
/home/travis/.gradle/caches/4.10/fileHashes/fileHashes.lock
/home/travis/.gradle/caches/4.10/javaCompile/javaCompile.lock
/home/travis/.gradle/caches/journal-1/file-access.bin
/home/travis/.gradle/caches/journal-1/journal-1.lock
changes detected, packing new archive
.
.
.
uploading archive

This should be fixed by removing all *.lock files, fileHashes.bin and file-access.bin before caching.

There is a need to implement dashboard page which will be used as a main screen of application. It should contain general overview about consumption and has a visualization that informations via charts.

There is a need to investigate integration with Google Photo API to use it as a source photo storage.
https://developers.google.com/picasa-web/docs/3.0/developers_guide_java#Photos

There is a need to introduce docker container for app

At the moment meter image uploading is very slow due to image hash calculation. There is a need to find a more efficient way to identify image uniqueness.

Currently method name is not obvious and it does not reflect method behaviour. It's better to name it like 'extractIndicationRegion', 'cropToIndication' or so.

Steps to reproduce:

1. Upload and recognise two images for same meter
2. Navigate to "Indications -> "
3. Click on "View image" button on indication's row

As a result 'Value' column in selected row contains value which is different from image.
Expected result: value and image must have identical values

Currently project is dependent on preinstalled opencv library. There is a need to reconfigure it to use existent opencv preset with included binaries (https://github.com/bytedeco/javacpp-presets/tree/master/opencv) as a dependency. This will help to avoid manual installation of opencv with all required bindings.

Currently there are a lot of places in code where nullable values are used. As a result there are a lot of null-checks to handle null values. The goal of this task is to perform deep code review and significantly decrease amount of using nullable variables and correspondent null-checks by replacing such variables with correspondent instances of java.util.Optional

There is a need to store consumption value for each indication. Consumption - is a difference between current indication value and previous one. It should be calculated indication value saving process.

1. Introduce new field 'consumption' for Indication entity
2. Extend indication saving logic with consumption calculation

Also, there is a need to consider situations where meter was replaced. Consumption in such situation could be less than zero. Such situations should be handled.

At the moment all resources for Swagger UI are placed and committed to GitHub. There is a need to investigate and find a way to avoid keeping them in repo.

The most likely solution would be downloading Swagger UI files during build time via Gradle as was done in following example using Maven: https://github.com/swagger-api/swagger-samples/blob/master/java/java-jaxrs-cxf/pom.xml

There is a need to introduce CI process using Travis CI.

There is a need to update README file with instructions related to builds for all available platforms. Such as arm, amd64, x84, etc.

There is a need to get rid of DbUnit dependency and reorganize tests to work without it. There are two potential resolutions for data-related tests:

1. Mock required data
2. Use Spring's @Sql annotation

Since embedded H2 does not support parallel connections, there is a need to reimplement DAO layer to use more advanced DB such as PostgreSQL.

The problem is that application returns indication region shaped correctly, but image inside it rotated to 90 degrees and stretched horizontally.

Most likely the root cause of the problem is in a bad corners ordering for perspective transformation.

Please replace github artefact with maven webjar dependency for Swagger UI
https://github.com/webjars/swagger-ui

There is a need to have an ability to log image transformation for debugging and troubleshooting.

Currently it's impossible to edit indication after it was saved. And this leads to a problem when indication was save with mistake. There is a need to implement edit feature for indications

At the moment, CI build take up to 17 minutes which is quite long period of time and it would be great to find approaches which could improve this value.

After quick look at the build log it was figured out that the most time consuming operation is building and installing OpenCV from sources. And speeding this up should be top priority for this task.

To achieve the task goal and speed-up CI build it's necessary to investigate and apply Travis CI features related to caching dependencies in build config.

At the moment filling data into test DB is implemented using a file with name 'data.sql'. It contains a set of insert operations which are executed during application context initialisation. Unfortunately, this approach is not convenient for development because test data are strongly rely on images which are processed via application and storing into DB. Updating 'data.sql' file or especially creating new one is not trivial and very time consuming operation.

To make test data creation process more simple and flexible, it's suggested to implement some mechanism which can pick up images from some folder and populate DB with corresponding data before test execution.

At the moment Google Drive directory id is getting during GDriveFileStorage object creation inside a constructor. As a result creation of GDriveFileStorage slows down Spring context start.

There is a need to implement lazy initialization for GDriveFileStorage

1. Cleanup CXF logging
2. Add liquibase loging to application logging