Add cache control headers for some static payloads (#4326)

packages/server/src/main/java/software/uncharted/terarium/hmiserver/configuration/Config.java

@@ -19,14 +19,23 @@
 @NoArgsConstructor
 public class Config {
 
-	/** The base url of the deployed application. Eg/ http://localhost:5173 or https://myapp.uncharted.software */
+	/**
+	 * The base url of the deployed application. Eg/ http://localhost:5173 or
+	 * https://myapp.uncharted.software
+	 */
 	String baseUrl;
 
 	/** A list of patterns for service requests that can be used via basic auth */
 	List<String> serviceRequestPatterns;
 
 	/**
-	 * A list of unauthenticated {@link org.springframework.util.AntPathMatcher} patterns for urls that should not be
+	 * Max age of the cache headers in seconds.
+	 */
+	Integer cacheHeadersMaxAge = 60 * 60 * 24;
+
+	/**
+	 * A list of unauthenticated {@link org.springframework.util.AntPathMatcher}
+	 * patterns for urls that should not be
 	 * authenticated via Spring Security
 	 */
 	List<String> unauthenticatedUrlPatterns;
@@ -40,13 +49,19 @@ public class Config {
 	/** Configuration values that are passed to the client */
 	ClientConfig clientConfig;
 
-	/** If queues should be declared durable. IF running Rabbit inside docker, this should be false */
+	/**
+	 * If queues should be declared durable. IF running Rabbit inside docker, this
+	 * should be false
+	 */
 	Boolean durableQueues = false;
 
 	/** The buffer size when uploading large files to the server */
 	Integer multipartFileBufferSize = 50 * 1024 * 1024;
 
-	/** The encryption key used to encrypt the download urls for retrieving files from the server */
+	/**
+	 * The encryption key used to encrypt the download urls for retrieving files
+	 * from the server
+	 */
 	String presignedUrlEncryptionKey;
 
 	/** The number of seconds that a presigned url signature is valid for */
@@ -74,15 +89,21 @@ public class Config {
 	@Accessors(chain = true)
 	public static class Caching {
 
-		/** If true, clear the cache on startup. Should be false in production environments */
+		/**
+		 * If true, clear the cache on startup. Should be false in production
+		 * environments
+		 */
 		Boolean clearOnStartup;
 	}
 
 	@Data
 	@Accessors(chain = true)
 	public static class Keycloak {
 
-		/** The url of the keycloak server. eg/ http://localhost:8081 or https://keycloak.uncharted.software */
+		/**
+		 * The url of the keycloak server. eg/ http://localhost:8081 or
+		 * https://keycloak.uncharted.software
+		 */
 		String url;
 		/** The realm name to use for authentication */
 		String realm;
@@ -108,19 +129,26 @@ public static class Keycloak {
 	@TSModel
 	public static class ClientConfig implements Serializable {
 
-		/** The base url of the deployed application. Mirror of {@link Config#baseUrl} */
+		/**
+		 * The base url of the deployed application. Mirror of {@link Config#baseUrl}
+		 */
 		String baseUrl;
 
 		/**
-		 * If true, we will log all client-side errors to the server. This is useful for debugging, but should be false
+		 * If true, we will log all client-side errors to the server. This is useful for
+		 * debugging, but should be false
 		 */
 		Boolean clientLogShippingEnabled;
 
-		/** The interval, in milliseconds, at which we will ship client-side logs to the server */
+		/**
+		 * The interval, in milliseconds, at which we will ship client-side logs to the
+		 * server
+		 */
 		Long clientLogShippingIntervalMillis;
 
 		/**
-		 * The interval, in milliseconds, at which we will send a heartbeat to connected clients for server-side-events
+		 * The interval, in milliseconds, at which we will send a heartbeat to connected
+		 * clients for server-side-events
 		 */
 		Long sseHeartbeatIntervalMillis;
 	}

packages/server/src/main/java/software/uncharted/terarium/hmiserver/controller/dataservice/ArtifactController.java

@@ -10,12 +10,14 @@
 import java.util.List;
 import java.util.Optional;
 import java.util.UUID;
+import java.util.concurrent.TimeUnit;
 import lombok.RequiredArgsConstructor;
 import lombok.extern.slf4j.Slf4j;
 import org.apache.http.HttpEntity;
 import org.apache.http.entity.ByteArrayEntity;
 import org.apache.http.entity.ContentType;
 import org.apache.http.entity.StringEntity;
+import org.springframework.http.CacheControl;
 import org.springframework.http.HttpStatus;
 import org.springframework.http.MediaType;
 import org.springframework.http.ResponseEntity;
@@ -32,6 +34,7 @@
 import org.springframework.web.bind.annotation.RestController;
 import org.springframework.web.multipart.MultipartFile;
 import org.springframework.web.server.ResponseStatusException;
+import software.uncharted.terarium.hmiserver.configuration.Config;
 import software.uncharted.terarium.hmiserver.models.dataservice.Artifact;
 import software.uncharted.terarium.hmiserver.models.dataservice.PresignedURL;
 import software.uncharted.terarium.hmiserver.models.dataservice.ResponseDeleted;
@@ -49,6 +52,8 @@
 @Transactional
 public class ArtifactController {
 
+	final Config config;
+
 	final ArtifactService artifactService;
 
 	final JsDelivrProxy gitHubProxy;
@@ -352,7 +357,15 @@ public ResponseEntity<byte[]> downloadFile(
 
 		try {
 			final Optional<byte[]> bytes = artifactService.fetchFileAsBytes(artifactId, filename);
-			return bytes.map(ResponseEntity::ok).orElseGet(() -> ResponseEntity.notFound().build());
+
+			if (bytes.isEmpty()) {
+				return ResponseEntity.notFound().build();
+			}
+			final CacheControl cacheControl = CacheControl.maxAge(
+				config.getCacheHeadersMaxAge(),
+				TimeUnit.SECONDS
+			).cachePublic();
+			return ResponseEntity.ok().cacheControl(cacheControl).body(bytes.get());
 		} catch (final Exception e) {
 			log.error("Unable to GET artifact data", e);
 			throw new ResponseStatusException(
@@ -390,7 +403,10 @@ public ResponseEntity<Integer> uploadFile(
 		return uploadArtifactHelper(artifactId, filename, fileEntity);
 	}
 
-	/** Downloads a file from GitHub given the path and owner name, then uploads it to the project. */
+	/**
+	 * Downloads a file from GitHub given the path and owner name, then uploads it
+	 * to the project.
+	 */
 	@PutMapping("/{id}/upload-artifact-from-github")
 	@Secured(Roles.USER)
 	@Operation(summary = "Uploads a file from GitHub to the artifact")
@@ -428,8 +444,8 @@ public ResponseEntity<Integer> uploadArtifactFromGithub(
 	/**
 	 * Uploads an artifact inside the entity to TDS via a presigned URL
 	 *
-	 * @param artifactId The ID of the artifact to upload to
-	 * @param fileName The name of the file to upload
+	 * @param artifactId         The ID of the artifact to upload to
+	 * @param fileName           The name of the file to upload
 	 * @param artifactHttpEntity The entity containing the artifact to upload
 	 * @return A response containing the status of the upload
 	 */

packages/server/src/main/java/software/uncharted/terarium/hmiserver/controller/dataservice/DatasetController.java

@@ -20,13 +20,15 @@
 import java.util.List;
 import java.util.Optional;
 import java.util.UUID;
+import java.util.concurrent.TimeUnit;
 import java.util.stream.Collectors;
 import lombok.RequiredArgsConstructor;
 import lombok.extern.slf4j.Slf4j;
 import org.apache.http.HttpEntity;
 import org.apache.http.entity.ByteArrayEntity;
 import org.apache.http.entity.ContentType;
 import org.apache.http.entity.StringEntity;
+import org.springframework.http.CacheControl;
 import org.springframework.http.HttpStatus;
 import org.springframework.http.MediaType;
 import org.springframework.http.ResponseEntity;
@@ -44,6 +46,7 @@
 import org.springframework.web.multipart.MultipartFile;
 import org.springframework.web.server.ResponseStatusException;
 import org.springframework.web.servlet.mvc.method.annotation.StreamingResponseBody;
+import software.uncharted.terarium.hmiserver.configuration.Config;
 import software.uncharted.terarium.hmiserver.models.dataservice.CsvAsset;
 import software.uncharted.terarium.hmiserver.models.dataservice.CsvColumnStats;
 import software.uncharted.terarium.hmiserver.models.dataservice.PresignedURL;
@@ -69,6 +72,8 @@ public class DatasetController {
 
 	private static final int DEFAULT_CSV_LIMIT = 100;
 
+	final Config config;
+
 	final DatasetService datasetService;
 	final ClimateDataProxy climateDataProxy;
 
@@ -364,7 +369,11 @@ public ResponseEntity<CsvAsset> getCsv(
 			csv.size()
 		);
 
-		return ResponseEntity.ok(csvAsset);
+		final CacheControl cacheControl = CacheControl.maxAge(
+			config.getCacheHeadersMaxAge(),
+			TimeUnit.SECONDS
+		).cachePublic();
+		return ResponseEntity.ok().cacheControl(cacheControl).body(csvAsset);
 	}
 
 	@GetMapping("/{id}/download-file")
@@ -475,7 +484,10 @@ public ResponseEntity<PresignedURL> getDownloadURL(
 		}
 	}
 
-	/** Downloads a CSV file from github given the path and owner name, then uploads it to the dataset. */
+	/**
+	 * Downloads a CSV file from github given the path and owner name, then uploads
+	 * it to the dataset.
+	 */
 	@PutMapping("/{id}/upload-csv-from-github")
 	@Secured(Roles.USER)
 	@Operation(summary = "Uploads a CSV file from github to a dataset")
@@ -522,10 +534,11 @@ public ResponseEntity<ResponseStatus> uploadCsvFromGithub(
 	}
 
 	/**
-	 * Uploads a CSV file to the dataset. This will grab a presigned URL from TDS then push the file to S3.
+	 * Uploads a CSV file to the dataset. This will grab a presigned URL from TDS
+	 * then push the file to S3.
 	 *
 	 * @param datasetId ID of the dataset to upload t
-	 * @param filename CSV file to upload
+	 * @param filename  CSV file to upload
 	 * @return Response
 	 */
 	@PutMapping(value = "/{id}/upload-csv", consumes = MediaType.MULTIPART_FORM_DATA_VALUE)
@@ -675,16 +688,19 @@ public ResponseEntity<PresignedURL> getUploadURL(
 	}
 
 	/**
-	 * Uploads a CSV file to the dataset. This will grab a presigned URL from TDS then push the file to S3 via a
+	 * Uploads a CSV file to the dataset. This will grab a presigned URL from TDS
+	 * then push the file to S3 via a
 	 * presigned URL and update the dataset with the headers.
 	 *
-	 * <p>If the headers fail to update there will be a print to the log, however, the response will just be the status
+	 * <p>
+	 * If the headers fail to update there will be a print to the log, however, the
+	 * response will just be the status
 	 * of the original csv upload.
 	 *
 	 * @param datasetId ID of the dataset to upload to
-	 * @param filename CSV file to upload
+	 * @param filename  CSV file to upload
 	 * @param csvEntity CSV file as an HttpEntity
-	 * @param headers headers of the CSV file
+	 * @param headers   headers of the CSV file
 	 * @return Response from the upload
 	 */
 	private ResponseEntity<ResponseStatus> uploadCSVAndUpdateColumns(

packages/server/src/main/java/software/uncharted/terarium/hmiserver/controller/dataservice/DocumentController.java

@@ -13,6 +13,7 @@
 import java.util.List;
 import java.util.Optional;
 import java.util.UUID;
+import java.util.concurrent.TimeUnit;
 import lombok.RequiredArgsConstructor;
 import lombok.extern.slf4j.Slf4j;
 import org.apache.commons.io.IOUtils;
@@ -21,6 +22,7 @@
 import org.apache.http.entity.ContentType;
 import org.apache.http.entity.StringEntity;
 import org.springframework.beans.factory.annotation.Value;
+import org.springframework.http.CacheControl;
 import org.springframework.http.HttpStatus;
 import org.springframework.http.MediaType;
 import org.springframework.http.ResponseEntity;
@@ -37,13 +39,12 @@
 import org.springframework.web.bind.annotation.RestController;
 import org.springframework.web.multipart.MultipartFile;
 import org.springframework.web.server.ResponseStatusException;
+import software.uncharted.terarium.hmiserver.configuration.Config;
 import software.uncharted.terarium.hmiserver.controller.services.DownloadService;
 import software.uncharted.terarium.hmiserver.models.dataservice.PresignedURL;
 import software.uncharted.terarium.hmiserver.models.dataservice.ResponseDeleted;
 import software.uncharted.terarium.hmiserver.models.dataservice.ResponseStatus;
 import software.uncharted.terarium.hmiserver.models.dataservice.document.DocumentAsset;
-import software.uncharted.terarium.hmiserver.models.dataservice.document.DocumentExtraction;
-import software.uncharted.terarium.hmiserver.models.dataservice.document.ExtractionAssetType;
 import software.uncharted.terarium.hmiserver.proxies.jsdelivr.JsDelivrProxy;
 import software.uncharted.terarium.hmiserver.proxies.skema.SkemaRustProxy;
 import software.uncharted.terarium.hmiserver.proxies.skema.SkemaUnifiedProxy;
@@ -64,6 +65,7 @@
 @Transactional
 public class DocumentController {
 
+	final Config config;
 	final ReBACService reBACService;
 	final CurrentUserService currentUserService;
 	final Messages messages;
@@ -380,7 +382,7 @@ public ResponseEntity<ResponseDeleted> deleteDocument(
 	 * Uploads an artifact inside the entity to TDS via a presigned URL
 	 *
 	 * @param documentId The ID of the document to upload to
-	 * @param fileName The name of the file to upload
+	 * @param fileName   The name of the file to upload
 	 * @param fileEntity The entity containing the file to upload
 	 * @return A response containing the status of the upload
 	 */
@@ -453,7 +455,10 @@ public ResponseEntity<Void> uploadDocument(
 		}
 	}
 
-	/** Downloads a file from GitHub given the path and owner name, then uploads it to the project. */
+	/**
+	 * Downloads a file from GitHub given the path and owner name, then uploads it
+	 * to the project.
+	 */
 	@PutMapping("/{documentId}/upload-document-from-github")
 	@Secured(Roles.USER)
 	@Operation(summary = "Uploads a document from github")
@@ -516,7 +521,14 @@ public ResponseEntity<byte[]> downloadDocument(
 	) {
 		try {
 			final Optional<byte[]> bytes = documentAssetService.fetchFileAsBytes(id, filename);
-			return bytes.map(ResponseEntity::ok).orElse(null);
+			if (bytes.isEmpty()) {
+				return ResponseEntity.notFound().build();
+			}
+			final CacheControl cacheControl = CacheControl.maxAge(
+				config.getCacheHeadersMaxAge(),
+				TimeUnit.SECONDS
+			).cachePublic();
+			return ResponseEntity.ok().cacheControl(cacheControl).body(bytes.get());
 		} catch (final Exception e) {
 			final String error = "Unable to download document";
 			log.error(error, e);
@@ -564,7 +576,7 @@ public ResponseEntity<String> getDocumentFileAsText(
 	 * Post Images to Equations Unified service to get an AMR
 	 *
 	 * @param documentId document id
-	 * @param filename filename of the image
+	 * @param filename   filename of the image
 	 * @return LaTeX representation of the equation
 	 */
 	@GetMapping("/{id}/image-to-equation")
@@ -617,11 +629,12 @@ public ResponseEntity<String> postImageToEquation(
 	}
 
 	/**
-	 * Uploads a PDF file to a document asset and then fires and forgets the extraction
+	 * Uploads a PDF file to a document asset and then fires and forgets the
+	 * extraction
 	 *
-	 * @param doi DOI of the document
+	 * @param doi      DOI of the document
 	 * @param filename filename of the PDF
-	 * @param docId document id
+	 * @param docId    document id
 	 * @return extraction job id
 	 */
 	private void uploadPDFFileToDocumentThenExtract(