add index search support for EML files in addition to DwC Archives; r…

…elated to #199

preston-cas/src/main/java/bio/guoda/preston/store/KeyTo1LevelZenodoBucket.java

@@ -10,7 +10,7 @@ public class KeyTo1LevelZenodoBucket implements KeyToPath {
 
     private final KeyToPath proxied;
 
-    private final AtomicReference<URI> lastZenodoBucket = new AtomicReference<URI>();
+    private final AtomicReference<URI> lastZenodoBucket = new AtomicReference<>();
 
     public KeyTo1LevelZenodoBucket(KeyToPath keyToPath) {
         this.proxied = keyToPath;

preston-serve/src/main/java/bio/guoda/preston/server/BadgeServlet.java

@@ -14,6 +14,7 @@
 import java.net.URISyntaxException;
 import java.nio.charset.StandardCharsets;
 import java.util.Map;
+import java.util.TreeMap;
 
 public class BadgeServlet extends RedirectingServlet {
 
@@ -29,21 +30,36 @@ protected void handleRequest(HttpServletResponse response,
                                  String sparqlEndpoint,
                                  String queryType,
                                  IRI requestedIdIRI,
-                                 int responseHttpStatus) throws IOException, ServletException {
+                                 int responseHttpStatus,
+                                 String resourceType) throws IOException, ServletException {
         try {
-            Map<String, String> provInfo = ProvUtil.findMostRecentContentId(
-                    requestedIdIRI,
-                    queryType,
-                    sparqlEndpoint);
+            Map<String, String> provInfo = ProvUtil
+                    .findMostRecentContentId(
+                            requestedIdIRI,
+                            queryType,
+                            sparqlEndpoint,
+                            resourceType
+                    );
+            final TreeMap<String, String> labelMap = new TreeMap<String, String>() {{
+                put(MimeTypes.MIME_TYPE_DWCA, "DwC Archive");
+                put(MimeTypes.MIME_TYPE_EML, "EML File");
+            }};
+            String typeLabel = labelMap.getOrDefault(resourceType, "DwC Archive");
             if (isOfKnownOrigin(provInfo)) {
                 URI uri = populateResponseHeader(response, getResolverEndpoint(), provInfo);
-                renderTemplate(response, uri, "origin-known.svg");
+                renderTemplate(
+                        response,
+                        uri,
+                        "origin-known.svg",
+                        typeLabel
+                );
                 log("found origin of [" + requestedIdIRI.getIRIString() + "]");
             } else {
                 renderTemplate(
                         response,
                         URI.create(getResolverEndpoint() + requestedIdIRI.getIRIString()),
-                        "origin-unknown.svg"
+                        "origin-unknown.svg",
+                        typeLabel
                 );
                 response.setStatus(HttpServletResponse.SC_NOT_FOUND);
                 log("did not find origin of [" + requestedIdIRI.getIRIString() + "]");
@@ -65,11 +81,13 @@ protected String getPrefix() {
     }
 
 
-    private void renderTemplate(HttpServletResponse response, URI uri, String templateResource) throws IOException {
+    private void renderTemplate(HttpServletResponse response, URI uri, String templateResource, String typeLabel) throws IOException {
         String badgeTemplate = IOUtils.toString(getClass().getResourceAsStream(templateResource), StandardCharsets.UTF_8);
-        String badge = StringUtils.replace(badgeTemplate, "{{REDIRECT_URL}}", uri.toString());
+        String badge = StringUtils
+                .replace(badgeTemplate, "{{REDIRECT_URL}}", uri.toString())
+                .replace("{{TYPE}}", typeLabel);
 
-        try(InputStream inputStream = IOUtils.toInputStream(badge, StandardCharsets.UTF_8)) {
+        try (InputStream inputStream = IOUtils.toInputStream(badge, StandardCharsets.UTF_8)) {
             IOUtils.copy(inputStream, response.getOutputStream());
             response.setHeader(HttpHeaders.CONTENT_TYPE, "image/svg+xml");
             response.setStatus(HttpServletResponse.SC_OK);

preston-serve/src/main/java/bio/guoda/preston/server/ProvUtil.java

@@ -31,17 +31,23 @@ public class ProvUtil {
     public static final String QUERY_TYPE_DOI = "doi";
     public static final List<String> QUERIES_SUPPORTED = Arrays.asList(QUERY_TYPE_DOI, QUERY_TYPE_UUID, QUERY_TYPE_URL, QUERY_TYPE_CONTENT_ID);
 
-    public static Map<String, String> findMostRecentContentId(IRI iri, String paramName, String sparqlEndpoint) throws IOException, URISyntaxException {
-        String response = findProvenance(iri, paramName, sparqlEndpoint);
+    public static Map<String, String> findMostRecentContentId(IRI iri, String paramName, String sparqlEndpoint, String resourceType) throws IOException, URISyntaxException {
+        String response = findProvenance(iri, paramName, sparqlEndpoint, resourceType);
         return extractProvenanceInfo(response);
     }
 
-    protected static String findProvenance(IRI iri, String paramName, String sparqlEndpoint) throws IOException, URISyntaxException {
-        InputStream resourceAsStream = RedirectingServlet.class.getResourceAsStream(paramName + ".rq");
+    protected static String findProvenance(IRI iri, String paramName, String sparqlEndpoint, String resourceType) throws IOException, URISyntaxException {
+        String queryTemplateName = paramName + ".rq";
+        InputStream resourceAsStream = RedirectingServlet.class.getResourceAsStream(queryTemplateName);
 
+        if (resourceAsStream == null) {
+            throw new IOException("failed to location query template [" + queryTemplateName + "]");
+        }
         String queryTemplate = IOUtils.toString(resourceAsStream, StandardCharsets.UTF_8);
 
-        String queryString = StringUtils.replace(queryTemplate, "?_" + paramName + "_iri", iri.toString());
+        String queryString = StringUtils
+                .replace(queryTemplate, "?_" + paramName + "_iri", iri.toString())
+                .replace("?_type", "\"" + resourceType + "\"");
 
         URI query = new URI("https", "example.org", "/query", "query=" + queryString, null);
 

preston-serve/src/main/java/bio/guoda/preston/server/RedirectingServlet.java

@@ -6,6 +6,8 @@
 import org.apache.commons.lang3.StringUtils;
 import org.apache.commons.rdf.api.IRI;
 import org.apache.http.HttpHeaders;
+import org.apache.http.NameValuePair;
+import org.apache.http.client.utils.URLEncodedUtils;
 
 import javax.servlet.ServletException;
 import javax.servlet.http.HttpServlet;
@@ -14,6 +16,7 @@
 import java.io.IOException;
 import java.net.URI;
 import java.net.URISyntaxException;
+import java.nio.charset.StandardCharsets;
 import java.util.ArrayList;
 import java.util.List;
 import java.util.Map;
@@ -27,6 +30,7 @@ public class RedirectingServlet extends HttpServlet {
     public static final String SEEN_AT = "seenAt";
     public static final String PROVENANCE_ID = "provenanceId";
     public static final String ARCHIVE_URL = "archiveUrl";
+    public static final String CONTENT_TYPE = "contentType";
     public static final String UUID = "uuid";
     public static final String DOI = "doi";
     public static final String CONTENT_ID = "contentId";
@@ -68,11 +72,21 @@ protected void doGet(
                     sparqlEndpoint,
                     queryType,
                     requestedIdIRI,
-                    redirectOnGetRequest(request));
+                    redirectOnGetRequest(request),
+                    getResourceType(request.getRequestURI()));
 
         }
     }
 
+    private String getResourceType(String requestURI) {
+        List<NameValuePair> parse = URLEncodedUtils.parse(requestURI, StandardCharsets.UTF_8);
+        return parse
+                .stream()
+                .filter(p -> StringUtils.equals(p.getName(), "type"))
+                .findFirst().map(NameValuePair::getValue)
+                .orElse(MimeTypes.MIME_TYPE_DWCA);
+    }
+
     protected String getPrefix() {
         return "/";
     }
@@ -92,12 +106,13 @@ protected void handleRequest(HttpServletResponse response,
                                  String sparqlEndpoint,
                                  String queryType,
                                  IRI requestedIdIRI,
-                                 int responseHttpStatus) throws IOException, ServletException {
+                                 int responseHttpStatus,
+                                 String resourceType) throws IOException, ServletException {
         try {
             Map<String, String> provInfo = ProvUtil.findMostRecentContentId(
                     requestedIdIRI,
                     queryType,
-                    sparqlEndpoint);
+                    sparqlEndpoint, resourceType);
             if (isOfKnownOrigin(provInfo)) {
                 populateResponseHeader(response, resolverEndpoint, provInfo);
                 response.setStatus(responseHttpStatus);
@@ -120,7 +135,7 @@ protected URI populateResponseHeader(HttpServletResponse response, String resolv
         String contentId = provInfo.get(CONTENT_ID);
         URI uri = getResolverURI(resolverEndpoint, contentId);
         response.setHeader(HttpHeaders.LOCATION, uri.toString());
-        response.setHeader(HttpHeaders.CONTENT_TYPE, MimeTypes.MIME_TYPE_DWCA);
+        response.setHeader(HttpHeaders.CONTENT_TYPE, provInfo.get(CONTENT_TYPE));
         response.setHeader(HttpHeaders.ETAG, contentId);
         response.setHeader(HttpHeaders.CONTENT_LOCATION, provInfo.get(ARCHIVE_URL));
         List<String> influencedBy = new ArrayList<>();

preston-serve/src/main/resources/bio/guoda/preston/server/contentId.rq

@@ -6,6 +6,7 @@ SELECT DISTINCT
   ?uuid
   ?doi
   ?contentId
+  ?contentType
   ?archiveUrl
   ?seenAt
   ?provenanceId
@@ -20,6 +21,7 @@ SELECT DISTINCT
                   ?activity
                 WHERE  {
                     BIND(?_contentId_iri as ?contentId)
+                    BIND(?_type as ?contentType)
                     BIND(<hash://sha256/5b7fa37bf8b64e7c935c4ff3389e36f8dd162f0705410dd719fd089e1ea253cd> as ?provenanceId)
                   {
                     SELECT ?uuid ?seenAt ?archiveUrl ?doi ?activity
@@ -37,7 +39,7 @@ SELECT DISTINCT
                 graph ?g1 {
                   ?uuid <http://www.w3.org/ns/prov#hadMember> ?archiveUrl .
                   ?uuid <http://www.w3.org/1999/02/22-rdf-syntax-ns#seeAlso> ?doi .
-                  ?archiveUrl <http://purl.org/dc/elements/1.1/format> "application/dwca" .
+                  ?archiveUrl <http://purl.org/dc/elements/1.1/format> ?_type .
                 }
             }
 

preston-serve/src/main/resources/bio/guoda/preston/server/doi.rq

@@ -6,20 +6,22 @@ SELECT DISTINCT
   ?doi
   ?uuid 
   ?contentId
-  ?archiveUrl 
+  ?contentType
+  ?archiveUrl
   ?seenAt
   ?provenanceId
   ?activity
 WHERE  {
   BIND(?_doi_iri as ?doi)
+  BIND(?_type as ?contentType)
   BIND(<hash://sha256/5b7fa37bf8b64e7c935c4ff3389e36f8dd162f0705410dd719fd089e1ea253cd> as ?provenanceId)
   {
     SELECT ?contentId ?seenAt ?archiveUrl ?uuid ?activity
     WHERE {
       graph ?g1 {
         ?uuid <http://www.w3.org/ns/prov#hadMember> ?archiveUrl .
         ?uuid <http://www.w3.org/1999/02/22-rdf-syntax-ns#seeAlso> ?_doi_iri .
-        ?archiveUrl <http://purl.org/dc/elements/1.1/format> "application/dwca" .
+        ?archiveUrl <http://purl.org/dc/elements/1.1/format> ?_type .
       }
       graph ?activity {
         ?activity <http://www.w3.org/ns/prov#used> ?archiveUrl .