I have a Spring mvc #RestController class where the method parameter is annotated with #RequestBody. Something like this:
#RestController
#RequestMapping("/features")
public class FeatureController {
#PostMapping
public Feature createFeature(#RequestBody Feature feature) {
......
}
}
My Feature class has a private constructor built using a Builder pattern. So I created HttpMessageConverter called FeatureConverter and registered it properly using extendMessageConverters. The converter uses Jackson to parse the JSON and then uses Feature.Builder to create an instance of Feature.
My problem is that Spring registers an instance of MappingJackson2HttpMessageConverter before my custom FeatureConverter. As a result the parsing is attempted by MappingJackson2HttpMessageConverter even before it can reach FeatureConverter. Since the constructor is private so MappingJackson2HttpMessageConverter fails.
My question is how do I change the order so that FeatureConverter is asked before MappingJackson2HttpMessageConverter. Is there a proper way of doing it?
Spring MVC WebMvcConfigurerAdapter initialize HttpMessageConverters like that:
protected final List<HttpMessageConverter<?>> getMessageConverters() {
if (this.messageConverters == null) {
this.messageConverters = new ArrayList<HttpMessageConverter<?>>();
configureMessageConverters(this.messageConverters);
if (this.messageConverters.isEmpty()) {
addDefaultHttpMessageConverters(this.messageConverters);
}
extendMessageConverters(this.messageConverters);
}
return this.messageConverters;
}
See: Github - WebMvcConfigurationSupport
As you can see there is 3 method:
configureMessageConverters, empty by default it's where you are suppose to add converters
addDefaultHttpMessageConverters, where as the name said will create all default converters, the default MappingJackson2HttpMessageConverter is created there.
extendMessageConverters which as the documentation:
Override this method to extend or modify the list of converters after
it has been configured. This may be useful for example to allow
default converters to be registered and then insert a custom
converter through this method.
Is to use when you want to modify the list after having all the defaults converter added.
To answer your question, if you don't need defaults converters you just have to Override configureMessageConverters and add it:
#Override
public void configureMessageConverters(List<HttpMessageConverter<?>> converters) {
converters.add(new FeatureConverter());
}
If you need them you can either copy-paste the addDefaultHttpMessageConverters content and add it into configureMessageConverters and play with the order of initialisation there (ugly solution).
Otherwise you can use extendMessageConverters to add it after the default initialization and play with the order:
#Override
public void extendMessageConverters(List<HttpMessageConverter<?>> converters) {
converters.add(the_index_you_want, new FeatureConverter());
}
If I understand you right you want to use your Builder for deserialization of Feature object.
Maybe you can just use Jackson annotations (without FeatureConverter):
import com.fasterxml.jackson.databind.annotation.JsonDeserialize;
import com.fasterxml.jackson.databind.annotation.JsonPOJOBuilder;
#JsonDeserialize(builder = Feature.FeatureBuilder.class)
public class Feature {
private final String name;
public Feature(String name) {
this.name = name;
}
public String getName() {
return name;
}
/**
* Builder
*/
#JsonPOJOBuilder(withPrefix = "")
public static final class FeatureBuilder {
private String name = "";
private FeatureBuilder() {
}
public static FeatureBuilder create() {
return new FeatureBuilder();
}
public FeatureBuilder name(String name) {
this.name = name;
return this;
}
/**
* Build Feature object
*/
public Feature build() {
Feature feature = new Feature(this.name);
return feature;
}
}
}
If you use Lombok:
import com.fasterxml.jackson.databind.annotation.JsonDeserialize;
import com.fasterxml.jackson.databind.annotation.JsonPOJOBuilder;
import lombok.Builder;
import lombok.Getter;
#Getter
#Builder
#JsonDeserialize(builder = Feature.FeatureBuilder.class)
public class Feature {
private final String name;
/**
* Builder
*/
#JsonPOJOBuilder(withPrefix = "")
public static final class FeatureBuilder {
}
}
Related
I have a Spring Boot Controller -
#RestController
public class UserController {
#PostMapping
#ResponseStatus(CREATED)
public UserResponse register( #Valid #RequestBody UserRequest userRequest) {
//return ....
}
}
Below is UserRequest.java
#Data
#NoArgsConstructor
#AllArgsConstructor
#Builder
public class UserRequest {
private String email;
//other property
}
I am sending below json in request body -
{
"email" : "TEST#Example.com",
//some other fields.
}
Sometime client send email in uppercase or in camel case so in userRquest I want to change value of email field to lowercase like test#example.com while de serializing to UserRequest Object.
Is there any easy way to do this. Can I introduce my own annotation like #ToLowerCase how I can create my own annotation and use that at field level in UserRequest.
There is no easy way just by introducing a new annotation #ToLowerCase,
because then you would also need to implement some annotation processor
for doing the real conversion work.
But you can achieve your goal in a slightly different way.
In your UserRequest class annotate the email property
with #JsonDeserialize and specify a converter there.
#JsonDeserialize(converter = ToLowerCaseConverter.class)
private String email;
You need to implement the converter class by yourself,
but it is easy by extending it from StdConverter.
public class ToLowerCaseConverter extends StdConverter<String, String> {
#Override
public String convert(String value) {
return value.toLowerCase();
}
}
Jackson will use the setter methods in your class.
Perform the conversion to lower case in the setter.
For example
public void setEmail(String newValue)
{
email = StringUtils.lowerCase(newValue);
}
StringUtils is an apache commons class.
You can make a general StringDeserializer and register it in ObjectMapper as shown below:-
StringDeserializer class
public final class StringDeserializer extends StdDeserializer<String> {
public StringDeserializer() {
super((Class<String>) null);
}
#Override
public String deserialize(JsonParser parser, DeserializationContext context) throws IOException {
JsonToken token = parser.getCurrentToken();
if (token == JsonToken.VALUE_STRING) {
String text = parser.getText();
return text == null ? null : text.toLowerCase().trim();
}
return null;
}
}
JacksonConfiguration class
#Configuration
public class JacksonConfiguration {
#Autowired
void mapper(ObjectMapper mapper) {
mapper.registerModule(initModule());
}
private Module initModule() {
SimpleModule module = new SimpleModule();
module.addDeserializer(String.class, new StringDeserializer());
return module;
}
}
The above code makes jackson deserialize all strings to lowercase and trimmed.
I went through the link: How to pass a Map<String, String> with application.properties and other related links multiple times, but still its not working.
I'm using Spring Boot and Spring REST example. Link Question: How to by default execute the latest version of endpoint in Spring Boot REST?.
I've created mapping something like this and simply read the mapping
get.customers={GET: '/app-data/customers', VERSION: 'v1'}
post.customers={POST: '/app-data/customers', VERSION: 'v1'}
get.customers.custId={GET: '/app-data/customers/{custId}', VERSION: 'v2'}
Code:
private String resolveLastVersion() {
// read from configuration or something
return "2";
}
Code:
#Component
#ConfigurationProperties
#PropertySource("classpath:restendpoint.properties")
public class PriorityProcessor {
private final Map<String, String> priorityMap = new HashMap<>();
public Map<String, String> getPriority() {
return priorityMap;
}
}
Code:
I suggest the following implementation:
#ConfigurationProperties(prefix="request")
public class ConfigurationProps {
private List<Mapping> mapping;
public List<Mapping> getMapping() {
return mapping;
}
public void setMapping(List<Mapping> mapping) {
this.mapping = mapping;
}
}
Class Mapping will denote the information about the single mapping:
public class Mapping {
private String method;
private String url;
private String version;
public Mapping(String method, String url, String version) {
this.method = method;
this.url = url;
this.version = version;
}
public Mapping() {
}
// getters setters here
}
On the Configuration or spring boot application class (the one with main method):
#EnableConfigurationProperties(ConfigurationProps.class)
In the properties file put:
request.mapping[0].method=get
request.mapping[0].url=/customers
request.mapping[0].version=1
request.mapping[1].method=post
request.mapping[1].url=/students
request.mapping[1].version=2
In Filter (I assume you followed my suggestion from the linked question):
#Component
#Order(1)
public class LatestVersionFilter implements Filter {
private List<Mapping> mappings;
public LatestVersionFilter(ConfigurationProps props) {
this.mappings = props.getMapping();
}
}
Feign default expander to convert param:
final class ToStringExpander implements Expander {
#Override
public String expand(Object value) {
return value.toString();
}
}
I want custom it to convert user to support GET param, like this
#FeignClient("xx")
interface UserService{
#RequestMapping(value="/users",method=GET)
public List<User> findBy(#ModelAttribute User user);
}
userService.findBy(user);
What can i do?
First,you must write a expander like ToJsonExpander:
public class ToJsonExpander implements Param.Expander {
private static ObjectMapper objectMapper = new ObjectMapper();
public String expand(Object value) {
try {
return objectMapper.writeValueAsString(value);
} catch (JsonProcessingException e) {
throw new ExpanderException(e);
}
}
}
Second, write a AnnotatedParameterProcessor like JsonArgumentParameterProcessor to add expander for your processor.
public class JsonArgumentParameterProcessor implements AnnotatedParameterProcessor {
private static final Class<JsonArgument> ANNOTATION = JsonArgument.class;
public Class<? extends Annotation> getAnnotationType() {
return ANNOTATION;
}
public boolean processArgument(AnnotatedParameterContext context, Annotation annotation) {
MethodMetadata data = context.getMethodMetadata();
String name = ANNOTATION.cast(annotation).value();
String method = data.template().method();
Util.checkState(Util.emptyToNull(name) != null,
"JsonArgument.value() was empty on parameter %s", context.getParameterIndex());
context.setParameterName(name);
if (method != null && (HttpMethod.POST.matches(method) || HttpMethod.PUT.matches(method) || HttpMethod.DELETE.matches(method))) {
data.formParams().add(name);
} else {
`data.indexToExpanderClass().put(context.getParameterIndex(), ToJsonExpander.class);`
Collection<String> query = context.setTemplateParameter(name, data.template().queries().get(name));
data.template().query(name, query);
}
return true;
}
}
Third,add it to Feign configuration.
#Bean
public Contract feignContract(){
List<AnnotatedParameterProcessor> processors = new ArrayList<>();
processors.add(new JsonArgumentParameterProcessor());
processors.add(new PathVariableParameterProcessor());
processors.add(new RequestHeaderParameterProcessor());
processors.add(new RequestParamParameterProcessor());
return new SpringMvcContract(processors);
}
Now, you can use #JsonArgument to send model argument like:
public void saveV10(#JsonArgument("session") Session session);
I don't know what #ModelAttribute does but I was looking for a way to convert #RequestParam values so I did this:
import com.google.i18n.phonenumbers.PhoneNumberUtil;
import com.google.i18n.phonenumbers.Phonenumber;
import org.springframework.cloud.netflix.feign.FeignFormatterRegistrar;
import org.springframework.format.FormatterRegistry;
import org.springframework.stereotype.Component;
import static com.google.i18n.phonenumbers.PhoneNumberUtil.PhoneNumberFormat.E164;
#Component
public class PhoneNumberFeignFormatterRegistrar implements FeignFormatterRegistrar {
private final PhoneNumberUtil phoneNumberUtil;
public PhoneNumberFeignFormatterRegistrar(PhoneNumberUtil phoneNumberUtil) {
this.phoneNumberUtil = phoneNumberUtil;
}
#Override
public void registerFormatters(FormatterRegistry registry) {
registry.addConverter(Phonenumber.PhoneNumber.class, String.class, source -> phoneNumberUtil.format(source, E164));
}
}
Now stuff like the following works
import com.google.i18n.phonenumbers.Phonenumber;
import org.springframework.cloud.netflix.feign.FeignClient;
import org.springframework.hateoas.Resource;
import org.springframework.web.bind.annotation.RequestMapping;
import org.springframework.web.bind.annotation.RequestMethod;
import org.springframework.web.bind.annotation.RequestParam;
#FeignClient("data-service")
public interface DataClient {
#RequestMapping(method = RequestMethod.GET, value = "/phoneNumbers/search/findByPhoneNumber")
Resource<PhoneNumberRecord> getPhoneNumber(#RequestParam("phoneNumber") Phonenumber.PhoneNumber phoneNumber);
}
As the open feign issue and spring doc say:
The OpenFeign #QueryMap annotation provides support for POJOs to be used as GET parameter maps.
Spring Cloud OpenFeign provides an equivalent #SpringQueryMap annotation, which is used to annotate a POJO or Map parameter as a query parameter map since 2.1.0.
You can use it like this:
#GetMapping("user")
String getUser(#SpringQueryMap User user);
public class User {
private String name;
private int age;
...
}
I have an entity named EmployeeDepartment as below
#IdClass(EmployeeDepartmentPK.class) //EmployeeDepartmentPK is a serializeable object
#Entity
EmployeeDepartment{
#Id
private String employeeID;
#Id
private String departmentCode;
---- Getters, Setters and other props/columns
}
and I have a Spring Data Repository defined as as below
#RepositoryRestResource(....)
public interface IEmployeeDepartmentRepository extends PagingAndSortingRepository<EmployeeDepartment, EmployeeDepartmentPK> {
}
Further, I have a converter registered to convert from String to EmployeeDepartmentPK.
Now, for an entity, qualified by ID employeeID="abc123" and departmentCode="JBG", I expect the ID to use when SDR interface is called is abc123_JBG.
For example http://localhost/EmployeeDepartment/abc123_JBG should fetch me the result and indeed it does.
But, when I try to save an entity using PUT, the ID property available in BasicPersistentEntity class of Spring Data Commons is having a value of
abc123_JBG for departmentCode. This is wrong. I'm not sure if this is an expected behaviour.
Please help.
Thanks!
Currently Spring Data REST only supports compound keys that are represented as by a single field. That effectively means only #EmbeddedId is supported. I've filed DATAJPA-770 to fix that.
If you can switch to #EmbeddedId you still need to teach Spring Data REST the way you'd like to represent your complex identifier in the URI and how to transform the path segment back into an instance of your id type. To achieve that, implement a BackendIdConverter and register it as Spring bean.
#Component
class CustomBackendIdConverter implements BackendIdConverter {
#Override
public Serializable fromRequestId(String id, Class<?> entityType) {
// Make sure you validate the input
String[] parts = id.split("_");
return new YourEmbeddedIdType(parts[0], parts[1]);
}
#Override
public String toRequestId(Serializable source, Class<?> entityType) {
YourIdType id = (YourIdType) source;
return String.format("%s_%s", …);
}
#Override
public boolean supports(Class<?> type) {
return YourDomainType.class.equals(type);
}
}
If you can't use #EmbeddedId, you can still use #IdClass. For that, you need the BackendIdConverter as Oliver Gierke answered, but you also need to add a Lookup for your domain type:
#Configuration
public class IdClassAllowingConfig extends RepositoryRestConfigurerAdapter {
#Override
public void configureRepositoryRestConfiguration(RepositoryRestConfiguration config) {
config.withEntityLookup().forRepository(EmployeeDepartmentRepository.class, (EmployeeDepartment ed) -> {
EmployeeDepartmentPK pk = new EmployeeDepartmentPK();
pk.setDepartmentId(ed.getDepartmentId());
pk.setEmployeeId(ed.getEmployeeId());
return pk;
}, EmployeeDepartmentRepository::findOne);
}
}
Use #BasePathAwareController to customize Spring data rest controller.
#BasePathAwareController
public class CustInfoCustAcctController {
#Autowired
CustInfoCustAcctRepository cicaRepo;
#RequestMapping(value = "/custInfoCustAccts/{id}", method = RequestMethod.GET)
public #ResponseBody custInfoCustAccts getOne(#PathVariable("id") String id) {
String[] parts = id.split("_");
CustInfoCustAcctKey key = new CustInfoCustAcctKey(parts[0],parts[1]);
return cicaRepo.getOne(key);
}
}
It's work fine for me with sample uri /api/custInfoCustAccts/89232_70
A more generic approach would be following -
package com.pratham.persistence.config;
import com.fasterxml.jackson.core.JsonProcessingException;
import com.fasterxml.jackson.databind.ObjectMapper;
import com.sun.istack.NotNull;
import lombok.RequiredArgsConstructor;
import org.springframework.data.rest.webmvc.spi.BackendIdConverter;
import org.springframework.lang.NonNull;
import org.springframework.stereotype.Component;
import javax.persistence.EmbeddedId;
import java.io.Serializable;
import java.lang.reflect.Field;
import java.util.Arrays;
import java.util.Base64;
import java.util.Optional;
import static java.nio.charset.StandardCharsets.UTF_8;
/**
* Customization of how composite ids are exposed in URIs.
* The implementation will convert the Ids marked with {#link EmbeddedId} to base64 encoded json
* in order to expose them properly within URI.
*
* #author im-pratham
*/
#Component
#RequiredArgsConstructor
public class EmbeddedBackendIdConverter implements BackendIdConverter {
private final ObjectMapper objectMapper;
#Override
public Serializable fromRequestId(String id, Class<?> entityType) {
return getFieldWithEmbeddedAnnotation(entityType)
.map(Field::getType)
.map(ret -> {
try {
String decodedId = new String(Base64.getUrlDecoder().decode(id));
return (Serializable) objectMapper.readValue(decodedId, (Class) ret);
} catch (JsonProcessingException ignored) {
return null;
}
})
.orElse(id);
}
#Override
public String toRequestId(Serializable id, Class<?> entityType) {
try {
String json = objectMapper.writeValueAsString(id);
return Base64.getUrlEncoder().encodeToString(json.getBytes(UTF_8));
} catch (JsonProcessingException ignored) {
return id.toString();
}
}
#Override
public boolean supports(#NonNull Class<?> entity) {
return isEmbeddedIdAnnotationPresent(entity);
}
private boolean isEmbeddedIdAnnotationPresent(Class<?> entity) {
return getFieldWithEmbeddedAnnotation(entity)
.isPresent();
}
#NotNull
private static Optional<Field> getFieldWithEmbeddedAnnotation(Class<?> entity) {
return Arrays.stream(entity.getDeclaredFields())
.filter(method -> method.isAnnotationPresent(EmbeddedId.class))
.findFirst();
}
}
Is it possible to use Spring's #Value annotation to read and write property values of a custom class type?
For example:
#Component
#PropertySource("classpath:/data.properties")
public class CustomerService {
#Value("${data.isWaiting:#{false}}")
private Boolean isWaiting;
// is this possible for a custom class like Customer???
// Something behind the scenes that converts Custom object to/from property file's string value via an ObjectFactory or something like that?
#Value("${data.customer:#{null}}")
private Customer customer;
...
}
EDITED SOLUTION
Here is how I did it using Spring 4.x APIs...
Created new PropertyEditorSupport class for Customer class:
public class CustomerPropertiesEditor extends PropertyEditorSupport {
// simple mapping class to convert Customer to String and vice-versa.
private CustomerMap map;
#Override
public String getAsText()
{
Customer customer = (Customer) this.getValue();
return map.transform(customer);
}
#Override
public void setAsText(String text) throws IllegalArgumentException
{
Customer customer = map.transform(text);
super.setValue(customer);
}
}
Then in application's ApplicationConfig class:
#Bean
public CustomEditorConfigurer customEditorConfigurer() {
Map<Class<?>, Class<? extends PropertyEditor>> customEditors =
new HashMap<Class<?>, Class<? extends PropertyEditor>>(1);
customEditors.put(Customer.class, CustomerPropertiesEditor.class);
CustomEditorConfigurer configurer = new CustomEditorConfigurer();
configurer.setCustomEditors(customEditors);
return configurer;
}
Cheers,
PM
You have to create a class extending PropertyEditorSupport.
public class CustomerEditor extends PropertyEditorSupport {
#Override
public void setAsText(String text) {
Customer c = new Customer();
// Parse text and set customer fields...
setValue(c);
}
}
It's possible but reading Spring documentation. You could see this example:
Example usage
#Configuration
#PropertySource("classpath:/com/myco/app.properties")
public class AppConfig {
#Autowired
Environment env;
#Bean
public TestBean testBean() {
TestBean testBean = new TestBean();
testBean.setName(env.getProperty("testbean.name"));
return testBean;
}
}
See details here
Spring can read properties and load them directly into a class.
Moreover, you can add #ConfigurationProperties(prefix = "data") on top of the class, instead of wiring each nested property one by one, by making the code cleaner.
Given all that, here is the final example with explanations:
// File: CustomerConfig.java
#Configuration
// Set property source file path (optional)
#PropertySource("classpath:/data.properties")
// Put prefix = "data" here so that Spring read properties under "data.*"
#ConfigurationProperties(prefix = "data")
public class CustomerConfig {
// Note: Property name here is the same as in the file (data.customer)
// Spring will automatically read and put "data.customer.*" properties into this object
private Customer customer;
// Other configs can be added here too... without wiring one-by-one
public setCustomer(Customer customer){
this.customer = customer;
}
public getCustomer(){
return this.customer;
}
}
That's it, now you have "data.customer.*" properties, loaded and accessible via CustomerConfig.getCustomer().
To integrate it into your service (based on your example code):
// File: CustomerService.java
#Component
#PropertySource("classpath:/data.properties")
public class CustomerService {
#Value("${data.isWaiting:#{false}}")
private Boolean isWaiting;
#Autowired // Inject configs, either with #Autowired or using constructor injection
private CustomerConfig customerConfig;
public void myMethod() {
// Now its available for use
System.out.println(customerConfig.getCustomer().toString());
}
}
This way no "magical hack" is required to read configs into a class.
Take a look at the #ConfigurationProperties documentation/examples, and this post for more useful info.
Note: I'd suggest against using PropertyEditorSupport, since
a) it was built for different purpose, may change in future by breaking the code
b) it requires manual "handling" code inside => possible bugs
Instead, use what was built right for that purpose (Spring already has it), in order to both make the code easier to understand, and to gain possible inner improvements/optimizations which might be done in the future (or present).
Further improvements: Your CustomerService seems to be cluttered with configs (#PropertyService) too. I'd suggest reading those properties via another class too (similarly) then wiring that class here, instead of doing all in the CustomerService.
If you want to use it with lists, there is a workaround using array instead.
Define your property as Customer[] instead of List then:
in ApplicationConfig class:
#Bean
public CustomEditorConfigurer customEditorConfigurer() {
Map<Class<?>, Class<? extends PropertyEditor>> customEditors =
new HashMap<Class<?>, Class<? extends PropertyEditor>>(1);
customEditors.put(Customer.class, CustomerPropertiesEditor.class);
customEditors.put(Customer[].class, CustomerPropertiesEditor.class);
CustomEditorConfigurer configurer = new CustomEditorConfigurer();
configurer.setCustomEditors(customEditors);
return configurer;
}
In CustomerEditor:
public class CustomerEditor extends PropertyEditorSupport {
public static final String DEFAULT_SEPARATOR = ",";
#Override
public void setAsText(String text) {
String[] array = StringUtils.delimitedListToStringArray(text, this.separator);
if (this.emptyArrayAsNull && array.length == 0) {
super.setValue((Object) null);
} else {
if (this.trimValues) {
array = StringUtils.trimArrayElements(array);
}
// Convert String[] to Customer[]
super.setValue(...);
}
}
}
If you want to use an existing converter/constructor, you can just call it within your expression.
For example:
#Value("#{T(org.test.CutomerMap).transform('${serialized.customer}')}")
private Customer customer;