feat: support for XML datasets

.gitignore

@@ -15,3 +15,6 @@ buildNumber.properties
 
 # MacOS
 .DS_Store
+
+# Project
+tmp

README.md

@@ -18,9 +18,11 @@ The current implementation produces simple table output. More sophisticated exam
 ./mvnw spring-boot:run
 ```
 
-* [http://localhost:8080/](http://localhost:8080/)
-* Save `.csv` files in the [`datasets`](datasets) directory
-* Reload the main page to browse the available datasets
+* [http://localhost:8080/](http://localhost:8080/) to open the web interface
+* Save `.csv` and `.xml` files in the [`datasets`](datasets) directory  
+  Note that the application does not support duplicate base names. E.g. it cannot properly distinguish between
+  datasets `items.csv` and `items.xml`.
+* Reload the web page to refresh the datasets
 
 ### Spring Developer tools
 

app/pom.xml

@@ -35,6 +35,10 @@
             <groupId>org.springframework.boot</groupId>
             <artifactId>spring-boot-starter-test</artifactId>
         </dependency>
+        <dependency>
+            <groupId>net.sourceforge.htmlcleaner</groupId>
+            <artifactId>htmlcleaner</artifactId>
+        </dependency>
     </dependencies>
 
     <build>
@@ -48,7 +52,6 @@
                     <workingDirectory>${project.parent.basedir}</workingDirectory>
                     <profiles>
                         <profile>local</profile>
-                        <profile>test-set</profile>
                     </profiles>
                 </configuration>
                 <executions>

app/src/main/java/nl/runnable/dataset/Dataset.java

@@ -2,5 +2,20 @@
 
 import org.springframework.core.io.Resource;
 
-public record Dataset(String name, Resource resource) {
+public record Dataset(String name, Resource resource, Type type,
+                      boolean hasHtmlFormat, boolean hasPdfFormat, boolean hasCsvFormat) {
+
+    public enum Type {
+        CSV,
+        XML,
+    }
+
+    public static Dataset forCsv(String name, Resource resource) {
+        return new Dataset(name, resource, Type.CSV, true, true, true);
+    }
+
+    public static Dataset forXml(String name, Resource resource, boolean hasHtmlFormat, boolean hasPdfFormat) {
+        return new Dataset(name, resource, Type.XML, hasHtmlFormat, hasPdfFormat, false);
+    }
+
 }

app/src/main/java/nl/runnable/dataset/DatasetRepository.java

@@ -11,4 +11,7 @@ public interface DatasetRepository {
 
     Stream<Dataset> findAll();
 
+    boolean stylesheetExists(@NonNull String name);
+
+
 }

app/src/main/java/nl/runnable/dataset/DatasetSourceFactory.java

@@ -7,10 +7,10 @@
 
 public interface DatasetSourceFactory {
 
-    Source createIndexPageSource(Stream<Dataset> datasets);
+    Source createIndexPageSource(@NonNull Stream<Dataset> datasets);
 
     Source createDatasetPageSource(@NonNull Dataset dataset);
 
-    Source createDatasetSource(@NonNull Dataset dataset);
+    Source createDatasetXmlSource(@NonNull Dataset dataset);
 
 }

app/src/main/java/nl/runnable/dataset/XsltDatasetsProperties.java

@@ -1,7 +1,7 @@
 package nl.runnable.dataset;
 
 import lombok.Getter;
-import lombok.NonNull;
+import lombok.Setter;
 import org.springframework.boot.context.properties.ConfigurationProperties;
 import org.springframework.context.annotation.Configuration;
 
@@ -10,12 +10,7 @@
 public class XsltDatasetsProperties {
 
     @Getter
+    @Setter
     private String directory;
 
-    public void setDirectory(@NonNull String directory) {
-        while (directory.endsWith("/")) {
-            directory = directory.substring(0, directory.length() - 1);
-        }
-        this.directory = directory;
-    }
 }

app/src/main/java/nl/runnable/dataset/fop/FopConfig.java

@@ -20,7 +20,7 @@ public class FopConfig {
     private Resource fopConfig;
 
     @Bean
-    FopFactory fopFactory(SpringResourceResolver resourceResolver) {
+    FopFactory fopFactory(ResourceLoaderResourceResolver resourceResolver) {
         try {
             final var configurationBuilder = new DefaultConfigurationBuilder();
             final var configuration = configurationBuilder.build(fopConfig.getInputStream());

app/src/main/java/nl/runnable/dataset/fop/SpringResourceResolver.java → app/src/main/java/nl/runnable/dataset/fop/ResourceLoaderResourceResolver.java

@@ -13,7 +13,7 @@
 
 @Component
 @RequiredArgsConstructor
-class SpringResourceResolver implements ResourceResolver {
+class ResourceLoaderResourceResolver implements ResourceResolver {
 
     private final ResourceLoader resourceLoader;
 

app/src/main/java/nl/runnable/dataset/jdom/JdomDatasetSourceFactory.java

@@ -16,6 +16,7 @@
 import javax.xml.transform.TransformerException;
 import javax.xml.transform.TransformerFactory;
 import javax.xml.transform.sax.SAXSource;
+import javax.xml.transform.stream.StreamSource;
 import java.io.IOException;
 import java.io.InputStreamReader;
 import java.util.stream.Stream;
@@ -33,6 +34,10 @@ public Source createIndexPageSource(Stream<Dataset> datasets) {
         datasets.map(dataset -> {
                     final var element = new Element("dataset");
                     element.setAttribute("name", dataset.name());
+                    element.setAttribute("type", dataset.type().name());
+                    element.setAttribute("hasHtmlFormat", Boolean.toString(dataset.hasHtmlFormat()));
+                    element.setAttribute("hasPdfFormat", Boolean.toString(dataset.hasPdfFormat()));
+                    element.setAttribute("hasCsvFormat", Boolean.toString(dataset.hasCsvFormat()));
                     return element;
                 })
                 .forEach(datasetsElement::addContent);
@@ -47,7 +52,7 @@ public Source createIndexPageSource(Stream<Dataset> datasets) {
     @Override
     public Source createDatasetPageSource(@NonNull Dataset dataset) {
         try {
-            final var source = createDatasetSource(dataset);
+            final var source = createDatasetXmlSource(dataset);
             final var transformer = TransformerFactory.newInstance().newTransformer();
             final var result = new JDOMResult();
             transformer.transform(source, result);
@@ -64,12 +69,17 @@ public Source createDatasetPageSource(@NonNull Dataset dataset) {
     }
 
     @Override
-    public Source createDatasetSource(Dataset dataset) {
+    public Source createDatasetXmlSource(Dataset dataset) {
         try {
-            final InputSource inputSource = new InputSource(new InputStreamReader(dataset.resource().getInputStream()));
-            final var saxSource = new SAXSource(inputSource);
-            saxSource.setXMLReader(new CsvXmlReader());
-            return saxSource;
+            return switch (dataset.type()) {
+                case CSV -> {
+                    final InputSource inputSource = new InputSource(new InputStreamReader(dataset.resource().getInputStream()));
+                    final var saxSource = new SAXSource(inputSource);
+                    saxSource.setXMLReader(new CsvXmlReader());
+                    yield saxSource;
+                }
+                case XML -> new StreamSource(dataset.resource().getInputStream());
+            };
         } catch (IOException e) {
             throw new RuntimeException(e);
         }

app/src/main/java/nl/runnable/dataset/spring/ResourceLoaderDatasetRepository.java

@@ -15,41 +15,65 @@
 import java.util.stream.Stream;
 
 @Component
-@RequiredArgsConstructor
 @Slf4j
+@RequiredArgsConstructor
 class ResourceLoaderDatasetRepository implements DatasetRepository {
 
     private final ResourcePatternResolver resourcePatternResolver;
 
-    @Value("${xslt-datasets.directory}")
     private String directory;
 
+    @Value("${xslt-datasets.directory}")
+    public void setDirectory(String directory) {
+        while (directory.endsWith("/")) {
+            directory = directory.substring(0, directory.length() - 1);
+        }
+        this.directory = directory;
+    }
+
     @Override
     public Optional<Dataset> find(String name) {
-        final var filename = "%s/%s.csv".formatted(directory, name);
-        final var resource = resourcePatternResolver.getResource(filename);
-        if (resource.exists()) {
-            return Optional.of(new Dataset(resolveName(resource), resource));
-        } else {
-            return Optional.empty();
+        final var csv = resourcePatternResolver.getResource("%s/%s.csv".formatted(directory, name));
+        if (csv.exists()) {
+            return Optional.of(Dataset.forCsv(resolveName(csv), csv));
         }
+
+        final var xml = resourcePatternResolver.getResource("%s/%s.xml".formatted(directory, name));
+        if (xml.exists()) {
+            final var hasHtmlFormat = stylesheetExists("%s.html".formatted(name));
+            final var hasPdfFormat = stylesheetExists("%s.fo".formatted(name));
+            return Optional.of(Dataset.forXml(name, xml, hasHtmlFormat, hasPdfFormat));
+        }
+
+        return Optional.empty();
     }
 
     @Override
     public Stream<Dataset> findAll() {
         try {
-            final var pattern = "%s/*.csv".formatted(directory);
-            return Stream.of(resourcePatternResolver.getResources(pattern))
-                    .map(resource -> {
-                        String name = resolveName(resource);
-                        return new Dataset(name, resource);
+            final var csvs = Stream.of(resourcePatternResolver.getResources("%s/*.csv".formatted(directory)))
+                    .map(csv -> Dataset.forCsv(resolveName(csv), csv));
+            final var xmls = Stream.of(resourcePatternResolver.getResources("%s/*.xml".formatted(directory)))
+                    .map(xml -> {
+                        final var name = resolveName(xml);
+                        final var hasHtmlFormat = stylesheetExists("%s.html".formatted(name));
+                        final var hasPdfFormat = stylesheetExists("%s.fo".formatted(name));
+                        return Dataset.forXml(name, xml, hasHtmlFormat, hasPdfFormat);
                     });
+            return Stream.concat(csvs, xmls).sorted((a, b) -> a.name().compareToIgnoreCase(b.name()));
         } catch (IOException e) {
             log.warn("Error resolving resources: {}", e.getMessage());
             return Stream.empty();
         }
     }
 
+    @Override
+    public boolean stylesheetExists(@NonNull String name) {
+        final var xslt = resourcePatternResolver.getResource("%s/%s.xslt".formatted(directory, name));
+        return xslt.exists();
+    }
+
+
     @NonNull
     private static String resolveName(@NonNull Resource resource) {
         var name = resource.getFilename();

app/src/main/java/nl/runnable/dataset/web/DatasetController.java

@@ -1,62 +1,73 @@
 package nl.runnable.dataset.web;
 
 import lombok.RequiredArgsConstructor;
+import nl.runnable.dataset.Dataset;
 import nl.runnable.dataset.DatasetRepository;
 import nl.runnable.dataset.DatasetSourceFactory;
 import org.springframework.stereotype.Controller;
 import org.springframework.web.bind.annotation.*;
 import org.springframework.web.servlet.ModelAndView;
 
+import javax.xml.transform.Source;
+import java.util.Map;
+import java.util.NoSuchElementException;
+
+import static nl.runnable.dataset.web.DatasetXsltViewResolver.datasetViewName;
+
 @Controller
 @RequestMapping("/datasets")
 @RequiredArgsConstructor
 public class DatasetController {
 
-    private final DatasetRepository repository;
+    private final DatasetRepository datasetRepository;
 
     private final DatasetSourceFactory datasetSourceFactory;
 
     @GetMapping
     public ModelAndView index() {
         final var modelAndView = new ModelAndView("dataset/index.html");
-        final var datasets = repository.findAll();
+        final var datasets = datasetRepository.findAll();
         modelAndView.addObject(datasetSourceFactory.createIndexPageSource(datasets));
         return modelAndView;
     }
 
     @GetMapping("/{name}.html")
     public ModelAndView datasetHtml(@PathVariable(value = "name") String name) {
-        final var dataset = repository.find(name).orElseThrow();
-        final var source = datasetSourceFactory.createDatasetPageSource(dataset);
-
-        final var modelAndView = new ModelAndView("dataset/table.html");
-        modelAndView.addObject(source);
-        return modelAndView;
+        final var dataset = datasetRepository.find(name).orElseThrow();
+
+        var viewName = "%s.html".formatted(name);
+        final Source source;
+        if (datasetRepository.stylesheetExists(viewName)) {
+            viewName = datasetViewName(viewName);
+            source = datasetSourceFactory.createDatasetXmlSource(dataset);
+        } else {
+            viewName = "dataset/table.html";
+            source = datasetSourceFactory.createDatasetPageSource(dataset);
+        }
+
+        return new ModelAndView(viewName, Map.of("source", source));
     }
 
     @GetMapping("/{name}.pdf")
     public ModelAndView datasetPdf(@PathVariable(value = "name") String name) {
-        final var dataset = repository.find(name).orElseThrow();
-        final var source = datasetSourceFactory.createDatasetSource(dataset);
-
-        final var modelAndView = new ModelAndView("dataset/table.fo");
-        modelAndView.addObject(source);
-        return modelAndView;
+        final var dataset = datasetRepository.find(name).orElseThrow();
+        final var source = datasetSourceFactory.createDatasetXmlSource(dataset);
+        return new ModelAndView("dataset/table.fo", Map.of("source", source));
     }
 
     @GetMapping("/{name}.xml")
     public ModelAndView datasetXml(@PathVariable(value = "name") String name) {
-        final var dataset = repository.find(name).orElseThrow();
-        final var source = datasetSourceFactory.createDatasetSource(dataset);
-
-        final var modelAndView = new ModelAndView("dataset/identity.xml");
-        modelAndView.addObject(source);
-        return modelAndView;
+        final var dataset = datasetRepository.find(name).orElseThrow();
+        final var source = datasetSourceFactory.createDatasetXmlSource(dataset);
+        return new ModelAndView("dataset/identity.xml", Map.of("source", source));
     }
 
     @GetMapping("/{name}.csv")
     public ModelAndView datasetCsv(@PathVariable(value = "name") String name) {
-        final var dataset = repository.find(name).orElseThrow();
+        final var dataset = datasetRepository.find(name).orElseThrow();
+        if (dataset.type() != Dataset.Type.CSV) {
+            throw new NoSuchElementException();
+        }
 
         final var modelAndView = new ModelAndView();
         modelAndView.setView(new ResourceView(dataset.resource(), "text/csv"));