Modified backend and frontend for handling grouping of datasets #460

Hardiksh16 · 2024-11-12T11:03:27Z

No description provided.

MichaelRoeder

Writing "Ungrouped" or "Others" everywhere is not a good way to handle such a problem. Since the group is optional, the classes should offer a constructor without this attribute. The default value (e.g., "Others") should be provided in a static final class attribute.

There are also some other smaller changes needed. Please have a look at the single comments.

src/main/java/org/aksw/gerbil/annotator/AnnotatorConfigurationImpl.java

src/main/java/org/aksw/gerbil/dataset/InstanceListBasedDataset.java

src/main/java/org/aksw/gerbil/dataset/datahub/DatahubNIFConfig.java

src/main/java/org/aksw/gerbil/dataset/impl/nif/NIFFileDatasetConfig.java

src/main/properties/datasets.properties

src/test/java/org/aksw/gerbil/annotator/TestAnnotatorConfiguration.java

MichaelRoeder · 2024-11-13T09:03:13Z

src/test/java/org/aksw/gerbil/annotator/decorator/ErrorCountingAnnotatorDecoratorTest.java

@@ -76,7 +76,7 @@ public static class ErrorCausingAnnotatorConfig extends AbstractAdapterConfigura
        private int errorsPerHundred;

        public ErrorCausingAnnotatorConfig(int errorsPerHundred) {
-            super("Error causing topic system", false, ExperimentType.ERec);
+            super("Error causing topic system","Un Grouped", false, ExperimentType.ERec);


Please use the static attribute that you defined further above.

It doesn't look like that 😉

src/test/java/org/aksw/gerbil/annotator/decorator/ErrorCountingAnnotatorDecoratorTest.java

Hardiksh16 · 2024-11-13T10:14:36Z

I have read your comments on the code and would refactor them as per your comments @MichaelRoeder

MichaelRoeder

There is just a small change request to make things nicer 🙂

MichaelRoeder · 2024-11-19T10:58:41Z

src/main/java/org/aksw/gerbil/web/MainController.java

@@ -306,18 +302,20 @@ public ModelAndView experiment(@RequestParam(value = "id") String id, HttpServle
    }

    @RequestMapping("/datasets")
-    public @ResponseBody List<String> datasets(@RequestParam(value = "experimentType") String experimentType) {
+    public @ResponseBody Map<String, List<String>> datasets(@RequestParam(value = "experimentType") String experimentType) {


That is a solution that would work 👍

However, the current implementation handles something that the Spring framework can already do for you. I am afraid that I have described the goal in an ambiguous way, so let me try to explain what I would suggest.
The method should return Map<String, List<DatasetCofiguration>> and the Spring framework should apply it's "magic" to transform it into JSON 🪄

Of course, Spring cannot do magic things, so we still have to provide the serialization 😉
However, it is sufficient if we tell Spring that there is exactly one class that implements the serialization and Spring can simply make use of it whenever it is necessary. You can do that with the method that is shown under point 3.2 in https://www.baeldung.com/spring-boot-customize-jackson-objectmapper (if I am not mistaken, I didn't read the whole text 😅 so let me know if you encounter problems). This method should go into our RootConfig.

MichaelRoeder · 2024-11-19T10:59:11Z

src/main/java/org/aksw/gerbil/web/config/AdapterList.java

+            serializedConfigs.add(json);
+        }
+        return serializedConfigs;
+    }


This serialization is actually not necessary. See comment further above.

Please remove it, if it is not used anymore.

MichaelRoeder · 2024-11-19T11:00:53Z

src/test/java/org/aksw/gerbil/annotator/decorator/ErrorCountingAnnotatorDecoratorTest.java

@@ -76,7 +76,7 @@ public static class ErrorCausingAnnotatorConfig extends AbstractAdapterConfigura
        private int errorsPerHundred;

        public ErrorCausingAnnotatorConfig(int errorsPerHundred) {
-            super("Error causing topic system", false, ExperimentType.ERec);
+            super("Error causing topic system","Un Grouped", false, ExperimentType.ERec);


It doesn't look like that 😉

MichaelRoeder

The UI looks good but I think that we have a general misunderstanding how the workflow to achieve the targeted functionality should look like 😅
So let me bring up my points why I think the current solution is not exactly what it should be...

MainController: I do not get, why the controller returns a Map<String, ...>. I assume that is done to make it easier for the UI to show the groups, because the data already comes in the form that the UI would like to have it. However, this is actually not really the task of the controller. If the UI would like to have the data in a certain form, it is free to organize the datasets by groups by itself. It is not really the task of the server to do that especially if we keep in mind that other (future) clients of the API are not interested in the group-wise order of datasets.
MainController: the serialization is still explicitly done by ourselves while it should be done by the Spring framework. Yes, we can implement it by ourselves but what should we return in case of an error? At the moment, null is returned but that does not fit to the HTTP standard. So why shouldn't we let Spring do that for us, since it has already implemented all the error handling, etc...

I also added a comment that shows how I imagine it to look like.

MichaelRoeder · 2024-11-27T12:34:08Z

pom.xml

+            <groupId>org.springframework.boot</groupId>
+            <artifactId>spring-boot-starter</artifactId>
+            <version>2.7.15</version>
+        </dependency>


Can you please move this dependency to the other Spring dependencies? Just search for the  line to find them.

MichaelRoeder · 2024-11-27T12:44:26Z

src/main/java/org/aksw/gerbil/web/MainController.java

+            LOGGER.error("Error fetching datasets for ExperimentType: {}", experimentType, e);
+            return null;
+        }
+        return response;


This part should simply look like:

public @ResponseBody Map<String, List<DatasetConfiguration>> datasets(@RequestParam(value = "experimentType") String experimentType) { ExperimentType type = null; Map<String, List<DatasetConfiguration>> response = new TreeMap<>(); try { type = ExperimentType.valueOf(experimentType); } catch (IllegalArgumentException e) { LOGGER.warn("Got a request containing a wrong ExperimentType (\"{}\"). Ignoring it.", experimentType); return null; } return adapterManager.getDatasetDetailsForExperiment(type); }

Actually, the return null; is also not really nice... instead, we should return HTTP 400 with an error message.

MichaelRoeder · 2024-11-27T12:45:02Z

src/main/java/org/aksw/gerbil/web/config/AdapterList.java

+            serializedConfigs.add(json);
+        }
+        return serializedConfigs;
+    }


Please remove it, if it is not used anymore.

MichaelRoeder · 2024-11-27T12:47:30Z

src/test/java/org/aksw/gerbil/annotator/decorator/ErrorCountingAnnotatorDecoratorTest.java

 import org.junit.Assert;
 import org.junit.Test;

 import it.unipi.di.acube.batframework.utils.AnnotationException;

 public class ErrorCountingAnnotatorDecoratorTest {

+    private static String UNGROUPED = "Un Grouped";


Why a different value than DatasetsConfig.DEFAULT_DATASET_GROUP is needed?

Modified backend and frontend for handling grouping of datasets

d550337

Hardiksh16 requested a review from MichaelRoeder November 12, 2024 11:03

Hardiksh16 linked an issue Nov 12, 2024 that may be closed by this pull request

Organize datasets in groups #341

Open

Sorted the dataset in natural order

8a0fe16

MichaelRoeder requested changes Nov 13, 2024

View reviewed changes

Hardiksh16 added 2 commits November 13, 2024 20:19

Refactored Default_DataSet_Group

6130c24

updated unit test cases

0431bd9

Hardiksh16 marked this pull request as ready for review November 13, 2024 19:31

Hardiksh16 added 4 commits November 13, 2024 20:47

updated datasets.properties

6aac61e

formatted config.jsp

66014cf

updated AdapterConfiguration

4e15ef8

Added AdapterConfigSerializer

05c8ff9

MichaelRoeder requested changes Nov 19, 2024

View reviewed changes

Added Springboot Serialization, Refactored dataset API

acd333a

MichaelRoeder requested changes Nov 27, 2024

View reviewed changes

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Modified backend and frontend for handling grouping of datasets #460

Modified backend and frontend for handling grouping of datasets #460

Hardiksh16 commented Nov 12, 2024

MichaelRoeder left a comment

MichaelRoeder Nov 13, 2024

Hardiksh16 Nov 13, 2024

MichaelRoeder Nov 19, 2024

Hardiksh16 commented Nov 13, 2024

MichaelRoeder left a comment

MichaelRoeder Nov 19, 2024

MichaelRoeder Nov 19, 2024

MichaelRoeder Nov 27, 2024

MichaelRoeder Nov 19, 2024

MichaelRoeder left a comment

MichaelRoeder Nov 27, 2024

MichaelRoeder Nov 27, 2024

MichaelRoeder Nov 27, 2024

MichaelRoeder Nov 27, 2024

Modified backend and frontend for handling grouping of datasets #460

Are you sure you want to change the base?

Modified backend and frontend for handling grouping of datasets #460

Conversation

Hardiksh16 commented Nov 12, 2024

MichaelRoeder left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Hardiksh16 commented Nov 13, 2024

MichaelRoeder left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

MichaelRoeder left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment