Accessing Reactive CRUD repository from MapStruct Mapper stuck in block() - java

I am somewhat new to Spring and have recently generated a JHipster monolith application with the WebFlux option. My current aim is to make it compatible with Firestore and implement some missing features like inserting document references. To do so, I am currently having the following structure:
A domain object class "Device" which holds a field String firmwareType;
A domain object class "FirmwareType"
A DTO object DeviceDTO which holds a field FirmwareType firmwareType;
Correspondingly, I also have the corresponding Repository (extending FirestoreReactiveRepository which extends ReactiveCrudRepository) and Controller implementations, which all work fine. To perform the conversion from a "full object" of FirmwareType in the DTO-object to a String firmwareTypeId; in the Device-object, I implemented a MapStruct Mapper:
#Mapper(unmappedTargetPolicy = org.mapstruct.ReportingPolicy.IGNORE, componentModel = "spring")
public abstract class DeviceMapper {
private final Logger logger = LoggerFactory.getLogger(DeviceMapper.class);
#Autowired
protected FirmwareTypeRepository fwTypeRepo;
public abstract Device dtoToDevice(DeviceDTO deviceDTO);
public abstract DeviceDTO deviceToDto(Device device);
public abstract List<DeviceDTO> devicesToDTOs(List<Device> devices);
public abstract List<Device> dtosToDevices(List<DeviceDTO> dtos);
public String map(FirmwareType value) {
if (value == null || value.getId() == null || value.getId().isEmpty()) {
return null;
}
return value.getId();
}
public FirmwareType map(String value) {
if (value == null || value.isEmpty()) {
return null;
}
return fwTypeRepo.findById(value).block(); // <<-- this gets stuck!
}
}
The FirmwareTypeRepository which is autowired as fwTypeRepo field:
#Repository
public interface FirmwareTypeRepository extends FirestoreReactiveRepository<FirmwareType> {
Mono<FirmwareType> findById(String id);
}
The corresponding map functions get called perfectly fine, but the fwTypeRepo.findById(..) call in the marked line (third-last line) seems to get stuck somewhere and never returns or throws an error. When the "fwTypeRepo" via its Controller-endpoint is called, it works without any issues.
I suppose it must be some kind of calling context issue or something? Is there another way to force a result by Mono synchronously than block?
Thanks for your help in advance, everyone!
Edit: At this point, I am sure it has something to do with Autowiring the Repository. It seems to not correctly do so / use the correct instance. While a customized Interface+Impl is called correctly, the underlying logic (from FirestoreReactive/ReactiveCrudRepository) doesn't seem to supply data correctly (also when #Autowire is used in other components!). I found some hints pointing at the package-structure but (i.e. Application class needs to be in a root package) but that isn't an issue.

Mapstruct is not reactive as i know so this approach won't work, you'll need mapstruct to return a mono that builds the object itself but that wouldn't make sense as it's a mapping lib which is only for doing blocking things.
Could try use 2 Mono/Mappers, 1 for each DB call and then just Mono.zip(dbCall1, dbCall2) and set the the mapped db call output into the other objects field.
var call1 = Mono.fromFuture(() -> db.loadObject1()).map(o -> mapper1.map(o));
var call2 = Mono.fromFuture(() -> db.loadObject2()).map(o -> mapper2.map(o));
Mono.zip(call1, call2)
.map(t -> {
var o1 = t.getT1();
var o2 = t.getT2();
o1.setField(o2);
});

Related

Using Optionals correctly in service layer of spring boot application

I new to spring boot application development. I using service layer in my application, but came across the repository method that return Optional as shown below.
#Override
public Questionnaire getQuestionnaireById(Long questionnaireId) {
Questionnaire returnedQuestionnaire = null;
Optional<Questionnaire> questionnaireOptional = questionnaireRepository.findById(questionnaireId);
if(questionnaireOptional.isPresent()) {
returnedQuestionnaire = questionnaireOptional.get();
}
return returnedQuestionnaire;
}
My question is ,
whether I am using the Optional correctly here. And is it ok to check this optional (isPresent()) in the RestController and throughing exception is not present.Like below
public Optional<Questionnaire> getQuestionnaireById(Long questionnaireId) {
return questionnaireRepository.findById(questionnaireId);
}
I wouldn't go for either option tbh, especially not the first. You don't want to introduce null values inside your domain. Your domain should stay as simple as possible, readable and void of clutter like null checks.
You might want to read through the optional API for all your options, but personally I would go for something like this:
In repository:
public interface TodoBoardRepository {
Optional<Questionnaire> findByQuestionnaireId(String questionnaireId);
// ...
}
In service:
#Service
#RequiredArgsConstructor // Or generate constructor if you're not using Lombok
public class QuestionnaireService {
private final QuestionnaireRepository questionnaireRepository;
// ...
public Questionnaire getQuestionnaireById(Long questionnaireId) {
Questionnaire questionnaire = questionnaireRepository.findById(questionnaireId)
.orElseThrow(() -> new QuestionaireNotFoundException(questionnaireId));
// Do whatever you want to do with the Questionnaire...
return questionnaire;
}
}
I go with way 1 that you have mentioned. In case the object is not present, throw a validation exception or something. This approach also ensures that service layer is in charge of the logic and controller is just used your interacting with the outside world.

Access to protected field of parent object through child object via a query schema

I am using graphql-spqr So that I don't have to create schema.graphql files.
I have a base class in which most of my other classes inherit from. For example
#GraphQLInterface(name = "BaseResponse", implementationAutoDiscovery = true)
#ToString
#Data
public class BaseResponse {
#JsonInclude(Include.NON_NULL)
protected String responseCode;
#JsonInclude(Include.NON_NULL)
protected String responseDescription;
protected String hash;
}
Other classes inherit from the above base class as below
#ToString(callSuper = true)
#Data
#AllArgsConstructor
public class GetAllGroupResponse extends BaseResponse {
private List<Group> groups;
}
When I set the parent values of GetAllGroupResponse like below
getAllGroupResponse.setResponseCode(ResponseCodeEnum.SH_97.getRespCode());
getAllGroupResponse.setResponseDescription(ResponseCodeEnum.SH_97.getRespDescription());
I wish to retrieve the value of responseCode from my graphql query request
{
getAllPagable(pageNumber : 1, numberOfRecords : 3) { responseCode groups { groupName } }
}
but it throws an error below which tells me it can't see the responseCode variable because it is not a direct property of the class GetAllGroupResponse
Validation error of type FieldUndefined: Field 'responseCode' in type 'GetAllGroupResponse' is undefined # 'getAllPagable/responseCode'
PS. BaseResponse is from a different package/library project that was imported into the current project
UPDATE
Based on the GitHub issues raised and solutions provided. I created a bean as below
#Bean
public ExtensionProvider<GeneratorConfiguration, ResolverBuilder> resolverBuilderExtensionProvider() {
String[] packages = {"com.sheeft.microservices.library", "com.sheeft.mircoservice.groups"};
return (config, current) -> {
List<ResolverBuilder> resolverBuilders = new ArrayList<>();
//add a custom subtype of PublicResolverBuilder that only exposes a method if it's called "greeting"
resolverBuilders.add(new PublicResolverBuilder() {
#Override
public PublicResolverBuilder withBasePackages(String... basePackages) {
return super.withBasePackages(packages); //To change body of generated methods, choose Tools | Templates.
}
});
//add the default builder
resolverBuilders.add(new AnnotatedResolverBuilder().withBasePackages(packages));
return resolverBuilders;
};
}
and now I get an error which says the following
object type 'GetAllGroupResponse' does not implement interface 'BaseResponse' because field 'responseCode' is missing
So I decided to add the getResponseCode getter method manually and it runs successfully. When I the call the /graphql endpoint it throws an exception
Object required to be not null",
"trace": "graphql.AssertException: Object required to be not null
I think all you need is to set the base packages for your project. E.g.
generator.withBasePackages("your.root.package", "library.root.package")
Or, if you're using SPQR Spring Starter, add
graphql.spqr.base-packages=your.root.package,library.root.package
to your application.properties file, for the same effect.
This is needed when you have a hierarchy of classes across multiple packages (or even libraries, like in your case) to let SPQR know which packages should be exposed, because otherwise it might accidentally expose some framework code or even core Java things like getClass.

How to name a composite class like this?

Our team are using Spring Boot 2 with sql2o as db library. In the paste in our services, for trivial methods, we simply call the repository and returns the model. For example, if I have a Supplier table, I had in the service
#Override
public List<Supplier> findAll() {
return supplierRepository.findAll();
}
Now, since in our controllers we need 99% in the cases other objects correlated to the model, I would create a composite class that holds the model and other models. For example:
#Override
public List<UnknownName> findAll() {
List<Supplier> suppliers = supplierRepository.findAll();
List<UnknownName> res = new ArrayList<>();
UnknownName unknownName;
LegalOffice legalOffice;
if (suppliers != null) {
for (Supplier supplier in suppliers) {
unknownName = new UnknownName();
unknownName.setSupplier(supplier);
legalOffice = legalOfficeService.findByIdlegaloffice(supplier.getLegalofficeid);
unknownName.setLegalOffice(legalOffice);
res.add(unknownName);
}
}
return res;
}
What should the name of class UnknownName?
PS: I simplified the code for better redability, but I use a generic enrich() function that I call for all the find methods, so I don't have to dupe the code.
I would recommend SupplierDto or SupplierLegalOfficeDto. DTO stands for Data Transfer Objects and it's commonly used for enriched models (more here).
Also you shouldn't check suppliers for null as repository always returns a non-null list.
In the end, I adopted the suffix Aggregator, following the Domain-driven design wording.

Exposing hypermedia links on collection even it's empty using Spring Data Rest

First at all I read the previous question: Exposing link on collection entity in spring data REST
But the issue still persist without trick.
Indeed if I want to expose a link for a collections resources I'm using the following code:
#Component
public class FooProcessor implements ResourceProcessor<PagedResources<Resource<Foo>>> {
private final FooLinks fooLinks;
#Inject
public FooProcessor(FooLinks fooLinks) {
this.FooLinks = fooLinks;
}
#Override
public PagedResources<Resource<Foo>> process(PagedResources<Resource<Foo>> resource) {
resource.add(fooLinks.getMyCustomLink());
return resource;
}
}
That works correctly except when collection is empty...
The only way to works is to replace my following code by:
#Component
public class FooProcessor implements ResourceProcessor<PagedResources> {
private final FooLinks fooLinks;
#Inject
public FooProcessor(FooLinks fooLinks) {
this.FooLinks = fooLinks;
}
#Override
public PagedResources process(PagedResources resource) {
resource.add(fooLinks.getMyCustomLink());
return resource;
}
}
But by doing that the link will be exposed for all collections.
I can create condition for exposing only for what I want but I don't think is clean.
I think spring does some magic there trying to discover the type of the collection - on an empty collection you cannot tell which type it is of - so spring-data-rest cannot determine which ResourceProcessor to use.
I think I have seen in
org.springframework.data.rest.webmvc.ResourceProcessorHandlerMethodReturnValueHandler.ResourcesProcessorWrapper#isValueTypeMatch that they try to determine the type by looking at the first element in the collection and otherwise just stop processing:
if (content.isEmpty()) {
return false;
}
So I think you cannot solve this using spring-data-rest. For your controller you could fall back to writing a custom controller and use spring hateoas and implement your own ResourceAssemblerSupport to see the link also on empty collections.

Spring Partial Update Object Data Binding

We are trying to implement a special partial update function in Spring 3.2. We are using Spring for the backend and have a simple Javascript frontend. I've not been able to find a straight-forward solution to our requirements, which is The update() function should take in any number of field:values and update the persistence model accordingly.
We have in-line editing for all of our fields, so that when the user edits a field and confirms, an id and the modified field get passed to the controller as json. The controller should be able to take in any number of fields from the client (1 to n) and update only those fields.
e.g., when a user with id==1 edits his displayName, the data posted to the server looks like this:
{"id":"1", "displayName":"jim"}
Currently, we have an incomplete solution in the UserController as outlined below:
#RequestMapping(value = "/{id}", method = RequestMethod.POST )
public #ResponseBody ResponseEntity<User> update(#RequestBody User updateUser) {
dbUser = userRepository.findOne(updateUser.getId());
customObjectMerger(updateUser, dbUser);
userRepository.saveAndFlush(updateUuser);
...
}
The code here works, but has some issues: The #RequestBody creates a new updateUser, fills in the id and the displayName. CustomObjectMerger merges this updateUser with the corresponding dbUser from the database, updating the only fields included in updateUser.
The problem is that Spring populates some fields in updateUser with default values and other auto-generated field values, which, upon merging, overwrites valid data that we have in dbUser. Explicitly declaring that it should ignore these fields is not an option, as we want our update to be able to set these fields as well.
I am looking into some way to have Spring automatically merge ONLY the information explicitly sent into the update() function into the dbUser (without resetting default/auto field values). Is there any simple way to do this?
Update: I've already considered the following option which does almost what I'm asking for, but not quite. The problem is that it takes update data in as #RequestParam and (AFAIK) doesn't do JSON strings:
//load the existing user into the model for injecting into the update function
#ModelAttribute("user")
public User addUser(#RequestParam(required=false) Long id){
if (id != null) return userRepository.findOne(id);
return null;
}
....
//method declaration for using #MethodAttribute to pre-populate the template object
#RequestMapping(value = "/{id}", method = RequestMethod.POST )
public #ResponseBody ResponseEntity<User> update(#ModelAttribute("user") User updateUser){
....
}
I've considered re-writing my customObjectMerger() to work more appropriately with JSON, counting and having it take into consideration only the fields coming in from HttpServletRequest. but even having to use a customObjectMerger() in the first place feels hacky when spring provides almost exactly what I am looking, minus the lacking JSON functionality. If anyone knows of how to get Spring to do this, I'd greatly appreciate it!
I've just run into this same problem. My current solution looks like this. I haven't done much testing yet, but upon initial inspection it looks to be working fairly well.
#Autowired ObjectMapper objectMapper;
#Autowired UserRepository userRepository;
#RequestMapping(value = "/{id}", method = RequestMethod.POST )
public #ResponseBody ResponseEntity<User> update(#PathVariable Long id, HttpServletRequest request) throws IOException
{
User user = userRepository.findOne(id);
User updatedUser = objectMapper.readerForUpdating(user).readValue(request.getReader());
userRepository.saveAndFlush(updatedUser);
return new ResponseEntity<>(updatedUser, HttpStatus.ACCEPTED);
}
The ObjectMapper is a bean of type org.codehaus.jackson.map.ObjectMapper.
Hope this helps someone,
Edit:
Have run into issues with child objects. If a child object receives a property to partially update it will create a fresh object, update that property, and set it. This erases all the other properties on that object. I'll update if I come across a clean solution.
We are using #ModelAttribute to achive what you want to do.
Create a method annotated with#modelattribute which loads a user based on a pathvariable throguh a repository.
create a method #Requestmapping with a param #modelattribute
The point here is that the #modelattribute method is the initializer for the model. Then spring merges the request with this model since we declare it in the #requestmapping method.
This gives you partial update functionality.
Some , or even alot? ;) would argue that this is bad practice anyway since we use our DAOs directly in the controller and do not do this merge in a dedicated service layer. But currently we did not ran into issues because of this aproach.
I build an API that merge view objects with entities before call persiste or merge or update.
It's a first version but I think It's a start.
Just use the annotation UIAttribute in your POJO`S fields then use:
MergerProcessor.merge(pojoUi, pojoDb);
It works with native Attributes and Collection.
git: https://github.com/nfrpaiva/ui-merge
Following approach could be used.
For this scenario, PATCH method would be more appropriate since the entity will be partially updated.
In controller method, take the request body as string.
Convert that String to JSONObject. Then iterate over the keys and update matching variable with the incoming data.
import org.json.JSONObject;
#RequestMapping(value = "/{id}", method = RequestMethod.PATCH )
public ResponseEntity<?> updateUserPartially(#RequestBody String rawJson, #PathVariable long id){
dbUser = userRepository.findOne(id);
JSONObject json = new JSONObject(rawJson);
Iterator<String> it = json.keySet().iterator();
while(it.hasNext()){
String key = it.next();
switch(key){
case "displayName":
dbUser.setDisplayName(json.get(key));
break;
case "....":
....
}
}
userRepository.save(dbUser);
...
}
Downside of this approach is, you have to manually validate the incoming values.
I've a customized and dirty solution employs java.lang.reflect package. My solution worked well for 3 years with no problem.
My method takes 2 arguments, objectFromRequest and objectFromDatabase both have the type Object.
The code simply does:
if(objectFromRequest.getMyValue() == null){
objectFromDatabase.setMyValue(objectFromDatabase.getMyValue); //change nothing
} else {
objectFromDatabase.setMyValue(objectFromRequest.getMyValue); //set the new value
}
A "null" value in a field from request means "don't change it!".
-1 value for a reference column which have name ending with "Id" means "Set it to null".
You can also add many custom modifications for your different scenarios.
public static void partialUpdateFields(Object objectFromRequest, Object objectFromDatabase) {
try {
Method[] methods = objectFromRequest.getClass().getDeclaredMethods();
for (Method method : methods) {
Object newValue = null;
Object oldValue = null;
Method setter = null;
Class valueClass = null;
String methodName = method.getName();
if (methodName.startsWith("get") || methodName.startsWith("is")) {
newValue = method.invoke(objectFromRequest, null);
oldValue = method.invoke(objectFromDatabase, null);
if (newValue != null) {
valueClass = newValue.getClass();
} else if (oldValue != null) {
valueClass = oldValue.getClass();
} else {
continue;
}
if (valueClass == Timestamp.class) {
valueClass = Date.class;
}
if (methodName.startsWith("get")) {
setter = objectFromRequest.getClass().getDeclaredMethod(methodName.replace("get", "set"),
valueClass);
} else {
setter = objectFromRequest.getClass().getDeclaredMethod(methodName.replace("is", "set"),
valueClass);
}
if (newValue == null) {
newValue = oldValue;
}
if (methodName.endsWith("Id")
&& (valueClass == Number.class || valueClass == Integer.class || valueClass == Long.class)
&& newValue.equals(-1)) {
setter.invoke(objectFromDatabase, new Object[] { null });
} else if (methodName.endsWith("Date") && valueClass == Date.class
&& ((Date) newValue).getTime() == 0l) {
setter.invoke(objectFromDatabase, new Object[] { null });
}
else {
setter.invoke(objectFromDatabase, newValue);
}
}
}
} catch (Exception e) {
e.printStackTrace();
}
}
In my DAO class, simcardToUpdate comes from http request:
simcardUpdated = (Simcard) session.get(Simcard.class, simcardToUpdate.getId());
MyUtil.partialUpdateFields(simcardToUpdate, simcardUpdated);
updatedEntities = Integer.parseInt(session.save(simcardUpdated).toString());
The main problem lies in your following code:
#RequestMapping(value = "/{id}", method = RequestMethod.POST )
public #ResponseBody ResponseEntity<User> update(#RequestBody User updateUser) {
dbUser = userRepository.findOne(updateUser.getId());
customObjectMerger(updateUser, dbUser);
userRepository.saveAndFlush(updateUuser);
...
}
In the above functions, you call some of your private functions & classes (userRepository, customObjectMerger, ...), but give no explanation how it works or how those functions look like. So I can only guess:
CustomObjectMerger merges this updateUser with the corresponding
dbUser from the database, updating the only fields included in
updateUser.
Here we don't know what happened in CustomObjectMerger (that's your function, and you don't show it). But from what you describe, I can make a guess: you copy all the properties from updateUser to your object at database. This is absolutely a wrong way, since when Spring map the object, it will fill all the data. And you only want to update some specific properties.
There are 2 options in your case:
1) Sending all the properties (including the unchanged properties) to the server. This may cost a little more bandwidth, but you still keep your way
2) You should set some special values as the default value for the User object (for example, id = -1, age = -1...). Then in customObjectMerger you just set the value that is not -1.
If you feel the 2 above solutions aren't satisfied, consider parsing the json request yourself, and don't bother with Spring object mapping mechanism. Sometimes it just confuse a lot.
Partial updates can be solved by using #SessionAttributes functionality, which are made to do what you did yourself with the customObjectMerger.
Look at my answer here, especially the edits, to get you started:
https://stackoverflow.com/a/14702971/272180
I've done this with a java Map and some reflection magic:
public static Entidade setFieldsByMap(Map<String, Object> dados, Entidade entidade) {
dados.entrySet().stream().
filter(e -> e.getValue() != null).
forEach(e -> {
try {
Method setter = entidade.getClass().
getMethod("set"+ Strings.capitalize(e.getKey()),
Class.forName(e.getValue().getClass().getTypeName()));
setter.invoke(entidade, e.getValue());
} catch (Exception ex) { // a lot of exceptions
throw new WebServiceRuntimeException("ws.reflection.error", ex);
}
});
return entidade;
}
And the entry point:
#Transactional
#PatchMapping("/{id}")
public ResponseEntity<EntityOutput> partialUpdate(#PathVariable String entity,
#PathVariable Long id, #RequestBody Map<String, Object> data) {
// ...
return new ResponseEntity<>(obj, HttpStatus.OK);
}

Categories