fix (core): Glob Resource Provider now supports e.g. **.java?charset=…

…UTF-8

java/dev/enola/ai/ollama/OllamaMain.java

@@ -20,6 +20,7 @@
 import dev.enola.common.function.MoreStreams;
 import dev.enola.common.io.resource.stream.GlobResourceProvider;
 import dev.enola.common.io.resource.stream.GlobResourceProviders;
+
 import io.github.amithkoujalgi.ollama4j.core.OllamaAPI;
 import io.github.amithkoujalgi.ollama4j.core.exceptions.OllamaBaseException;
 import io.github.amithkoujalgi.ollama4j.core.utils.OptionsBuilder;
@@ -64,7 +65,9 @@ public static void main(String[] args)
                                 modelName,
                                 "Here is a lot of source code: "
                                         + context
-                                        + "\nNow please answer the following question about this code, and avoid repeating yourself: "
+                                        + "\n"
+                                        + "Now please answer the following question about this"
+                                        + " code, and avoid repeating yourself: "
                                         + prompt,
                                 // "Do not provide any reasoning, just answer the question",
                                 options)

java/dev/enola/common/io/iri/URIs.java

@@ -17,21 +17,28 @@
  */
 package dev.enola.common.io.iri;
 
+import static java.util.Collections.emptyMap;
+
 import com.google.common.base.Splitter;
 import com.google.common.base.Strings;
 import com.google.common.net.MediaType;
 
 import java.net.URI;
+import java.net.URISyntaxException;
 import java.net.URLEncoder;
 import java.nio.charset.StandardCharsets;
+import java.nio.file.FileSystems;
 import java.nio.file.Path;
-import java.util.Collections;
 import java.util.HashMap;
 import java.util.Map;
 
 public final class URIs {
     // see also class dev.enola.common.io.iri.IRIs
 
+    // TODO Review if all this String instead of URI-based processing could be removed and replaced
+    // with first encoding invalid special characters in URIs; see the related TBD in
+    // FileGlobResourceProvider.
+
     // URI Query Parameter Names
     private static final String MEDIA_TYPE = "mediaType";
     private static final String CHARSET = "charset"; // as in MediaType#CHARSET_ATTRIBUTE
@@ -64,6 +71,20 @@ public static URI addQuery(URI uri, Map<String, String> parameters) {
         return uri;
     }
 
+    /**
+     * Returns an URI with everything except the query parameters of the uri (1st) parameter, but
+     * the query parameters of originalUriWithQuery (2nd) parameter - IFF the uri (1st) parameter
+     * has no query; otherwise just returns the uri (1st) parameter as-is.
+     *
+     * <p>See {@link URIsTest#testAddQueryGivenOriginalUriWithQuery()} for example.
+     */
+    public static URI addQuery(URI uri, URI originalUriWithQuery) {
+        if (Strings.isNullOrEmpty(originalUriWithQuery.getQuery())) return uri;
+        if (!Strings.isNullOrEmpty(uri.getQuery())) return uri;
+
+        return URI.create(uri.toString() + "?" + originalUriWithQuery.getQuery());
+    }
+
     public static String addQuery(String string, Map<String, String> parameters) {
         return addQuery(URI.create(string), parameters).toString();
     }
@@ -90,43 +111,117 @@ private static String encodeQueryParameterValue(String value) {
     }
 
     // package-local not public (for now)
-    static Map<String, String> getQueryMap(URI uri) {
-        if (uri == null) return Collections.emptyMap();
-        Map<String, String> map = new HashMap<>();
-        String query = uri.getQuery();
-        if (Strings.isNullOrEmpty(query)) {
-            var part = uri.getSchemeSpecificPart();
-            var qp = part.indexOf('?');
-            if (qp > -1) {
-                var fp = part.indexOf('#');
-                if (fp == -1) {
-                    query = part.substring(qp + 1);
-                } else {
-                    query = part.substring(qp + 1, fp);
-                }
+    static String getQueryString(String uri) {
+        var qp = uri.indexOf('?');
+        if (qp > -1) {
+            // Handle Glob URIs, like e.g. "file:/tmp//?.txt"
+            if (uri.indexOf('=', qp) == -1) return "";
+            var fp = uri.indexOf('#');
+            if (fp == -1) {
+                return uri.substring(qp + 1);
+            } else {
+                return uri.substring(qp + 1, fp);
             }
         }
+        return "";
+    }
+
+    public static Map<String, String> getQueryMap(String uri) {
+        return getQueryMapGivenQueryString(getQueryString(uri));
+    }
+
+    static Map<String, String> getQueryMapGivenQueryString(String query) {
         if (Strings.isNullOrEmpty(query)) {
-            return Collections.emptyMap();
+            return emptyMap();
         }
-        AMPERSAND_SPLITTER.split(query).forEach(queryParameter -> put(uri, queryParameter, map));
+        Map<String, String> map = new HashMap<>();
+        AMPERSAND_SPLITTER.split(query).forEach(queryParameter -> put(queryParameter, map));
         return map;
     }
 
-    private static void put(URI uri, String queryParameter, Map<String, String> map) {
+    // package-local not public (for now)
+    static Map<String, String> getQueryMap(URI uri) {
+        if (uri == null) return emptyMap();
+
+        String query = uri.getQuery();
+        if (Strings.isNullOrEmpty(query)) {
+            var part = uri.getSchemeSpecificPart();
+            query = getQueryString(part);
+        }
+        return getQueryMapGivenQueryString(query);
+    }
+
+    private static void put(String queryParameter, Map<String, String> map) {
         var p = queryParameter.indexOf('=');
         if (p == -1) {
             map.put(queryParameter, null);
         } else {
             var key = queryParameter.substring(0, p);
             if (map.containsKey(key))
                 throw new IllegalArgumentException(
-                        uri.toString() + " ID URI ?query has duplicate key");
+                        "URI Query Parameter has duplicate key: " + queryParameter);
             var value = queryParameter.substring(p + 1);
             map.put(key.toLowerCase(), value);
         }
     }
 
+    /**
+     * Get a {@link Path} from an {@link URI}. This method is used internally by {@link
+     * dev.enola.common.io.resource.Resource} framework implementations, and typically shouldn't be
+     * called directly by users. Please see the {@link dev.enola.common.io.resource.FileResource}
+     * for more related background.
+     */
+    public static Path getFilePath(URI uri) {
+        // TODO Replace this with return Path.of(uri); but it needs more work...
+        // Both for relative file URIs and query parameters and ZIP files.
+        // Nota bene: https://stackoverflow.com/q/25032716/421602
+        // https://docs.oracle.com/javase/7/docs/technotes/guides/io/fsp/zipfilesystemprovider.html
+        // https://docs.oracle.com/en/java/javase/21/docs/api/jdk.zipfs/module-summary.html
+
+        var scheme = uri.getScheme();
+        var authority = uri.getAuthority();
+        var path = getPath(uri);
+        return getFilePath(scheme, authority, path);
+    }
+
+    public static Path getFilePath(String uri) {
+        var scheme = getScheme(uri);
+        var authority = ""; // TODO Implement getAuthority(String uri)
+        var path = getPath(uri);
+        return getFilePath(scheme, "", path);
+    }
+
+    private static Path getFilePath(String scheme, String authority, String path) {
+        // TODO Don't hard-code this to file: but use MoreFileSystems.URI_SCHEMAS, somehow...
+        if ("file".equals(scheme)) {
+            return FileSystems.getDefault().getPath(path);
+        } else
+            try {
+                // TODO Better null or "" for path?
+                URI fsURI = new URI(scheme, authority, "", null, null);
+                var fs = FileSystems.getFileSystem(fsURI);
+                return fs.getPath(path);
+            } catch (URISyntaxException e) {
+                // This is rather unexpected...
+                throw new IllegalStateException(
+                        "Failed to create FileSystem Authority URI: " + scheme + ":" + path, e);
+            }
+    }
+
+    static String getScheme(String iri) {
+        if (iri == null) return "";
+        var p = iri.indexOf(':');
+        if (p == -1) return "";
+        return iri.substring(0, p);
+    }
+
+    static String getSchemeSpecificPart(String iri) {
+        if (iri == null) return "";
+        var p = iri.indexOf(':');
+        if (p == -1) return "";
+        return iri.substring(p + 1);
+    }
+
     /**
      * Get the "path"-like component of any URI. Similar to {@link URI#getPath()}, but also works
      * e.g. for "non-standard" relative "file:hello.txt" URIs, and correctly chops off query
@@ -140,8 +235,16 @@ public static String getPath(URI uri) {
         return chopFragmentAndQuery(uri.getSchemeSpecificPart());
     }
 
+    public static String getPath(String uri) {
+        return chopFragmentAndQuery(getSchemeSpecificPart(uri));
+    }
+
     private static String chopFragmentAndQuery(String ssp) {
         var chop = ssp.indexOf('?', 0);
+
+        // Handle Glob URIs, like e.g. "file:/tmp//?.txt"
+        if (ssp.indexOf('=', chop) == -1) chop = -1;
+
         if (chop == -1) chop = ssp.indexOf('#', 0);
         if (chop == -1) chop = ssp.length();
 

java/dev/enola/common/io/iri/URIsTest.java

@@ -22,6 +22,7 @@
 import com.google.common.collect.ImmutableMap;
 import com.google.common.io.Resources;
 import com.google.common.net.MediaType;
+import com.google.common.truth.Truth8;
 
 import dev.enola.common.io.iri.URIs.MediaTypeAndOrCharset;
 
@@ -31,12 +32,13 @@
 import java.io.File;
 import java.net.URI;
 import java.net.URISyntaxException;
+import java.nio.file.Path;
 
 public class URIsTest {
 
     @Test
     public void testGetQueryMap() throws URISyntaxException {
-        assertThat(URIs.getQueryMap(null)).isEmpty();
+        assertThat(URIs.getQueryMap((URI) null)).isEmpty();
         assertThat(URIs.getQueryMap(URI.create(""))).isEmpty();
         assertThat(URIs.getQueryMap(URI.create("http://www.vorburger.ch"))).isEmpty();
         assertThat(URIs.getQueryMap(URI.create("http://google.com?q=michi#fragment")))
@@ -49,6 +51,11 @@ public void testGetQueryMap() throws URISyntaxException {
                 .containsExactly("charset", "ASCII");
         assertThat(URIs.getQueryMap(URI.create("scheme:thing?ping=pong=pang#fragment")))
                 .containsExactly("ping", "pong=pang");
+
+        // Glob URIs
+        assertThat(URIs.getQueryMap("file:/tmp//?.txt")).isEmpty();
+        assertThat(URIs.getQueryMap("file:/tmp//?q=.")).containsExactly("q", ".");
+        // TODO assertThat(URIs.getQueryMap(URI.create("file:/tmp//?.txt"))).isEmpty();
     }
 
     @Test
@@ -97,6 +104,19 @@ public void testAddQuery() {
                 .isEqualTo("http://host/path?arg1=a&arg2=b&arg3=c");
     }
 
+    @Test
+    public void testAddQueryGivenOriginalUriWithQuery() {
+        var uri1 = URI.create("http://host/pathX");
+        var uri2 = URI.create("http://host/pathY?arg1=a");
+        var uri3 = URI.create("http://host/pathZ?arg2=b");
+        var uri4 = URI.create("http://host/pathZ");
+
+        assertThat(URIs.addQuery(uri1, uri2)).isEqualTo(URI.create("http://host/pathX?arg1=a"));
+        assertThat(URIs.addQuery(uri2, uri3)).isEqualTo(uri2);
+        assertThat(URIs.addQuery(uri1, uri4)).isEqualTo(uri1);
+        assertThat(URIs.addQuery(uri2, uri4)).isEqualTo(uri2);
+    }
+
     @Test
     public void testGetFilename() throws URISyntaxException {
         // Files
@@ -187,5 +207,37 @@ public void testGetPath() {
         assertThat(URIs.getPath(URI.create("whatever:/place/something.test")))
                 .isEqualTo("/place/something.test");
         assertThat(URIs.getPath(URI.create("whatever:something.test"))).isEqualTo("something.test");
+
+        // Glob URIs
+        assertThat(URIs.getPath("file:/tmp//?.txt")).isEqualTo("/tmp//?.txt");
+        assertThat(URIs.getPath("file:/tmp//{xy}?q=.")).isEqualTo("/tmp//{xy}");
+    }
+
+    @Test
+    public void testGetFilePath() {
+        // Truth8.assertThat(Path.of("/tmp//{xy}?.txt#yo"))
+        //        .isEqualTo(Path.of("/tmp/?.txt")); // TODO rm
+
+        Truth8.assertThat(URIs.getFilePath("file:/tmp/")).isEqualTo(Path.of("/tmp"));
+
+        // Glob URIs
+        Truth8.assertThat(URIs.getFilePath("file:/tmp//?.txt")).isEqualTo(Path.of("/tmp/?.txt"));
+        Truth8.assertThat(URIs.getFilePath("file:/tmp//{xy}?q=.")).isEqualTo(Path.of("/tmp/{xy}"));
+    }
+
+    @Test
+    public void testGetScheme() {
+        assertThat(URIs.getScheme("test:something")).isEqualTo("test");
+        assertThat(URIs.getScheme("rela/tive")).isEqualTo("");
+        assertThat(URIs.getScheme("/absolute/rela/tive")).isEqualTo("");
+        assertThat(URIs.getScheme(null)).isEqualTo("");
+    }
+
+    @Test
+    public void testGetSchemeSpecificPart() {
+        assertThat(URIs.getSchemeSpecificPart("test:something")).isEqualTo("something");
+        assertThat(URIs.getSchemeSpecificPart("rela/tive")).isEqualTo("");
+        assertThat(URIs.getSchemeSpecificPart("/absolute/rela/tive")).isEqualTo("");
+        assertThat(URIs.getSchemeSpecificPart(null)).isEqualTo("");
     }
 }

java/dev/enola/common/io/resource/FileResource.java

@@ -27,7 +27,6 @@
 
 import java.io.IOException;
 import java.net.URI;
-import java.net.URISyntaxException;
 import java.nio.file.*;
 import java.time.Instant;
 import java.util.Arrays;
@@ -67,43 +66,20 @@ public Resource getResource(URI uri) {
     private final Path path;
     private final OpenOption[] openOptions;
 
-    private static Path pathFromURI(URI uri) {
-        // TODO Replace this with return Path.of(uri); but it needs more work...
-        // Both for relative file URIs and query parameters and ZIP files.
-        // Nota bene: https://stackoverflow.com/q/25032716/421602
-        // https://docs.oracle.com/javase/7/docs/technotes/guides/io/fsp/zipfilesystemprovider.html
-        // https://docs.oracle.com/en/java/javase/21/docs/api/jdk.zipfs/module-summary.html
-
-        var path = URIs.getPath(uri);
-        var scheme = uri.getScheme();
-        if ("file".equals(scheme)) {
-            return FileSystems.getDefault().getPath(path);
-        } else
-            try {
-                URI fsURI = new URI(scheme, uri.getAuthority(), "", null, null);
-                var fs = FileSystems.getFileSystem(fsURI);
-                return fs.getPath(path);
-            } catch (URISyntaxException e) {
-                // This is rather unexpected...
-                throw new IllegalStateException(
-                        "Failed to create FileSystem Authority URI: " + uri.toString(), e);
-            }
-    }
-
     public FileResource(URI uri, MediaType mediaType, OpenOption... openOptions) {
         super(uri, mediaType);
-        this.path = pathFromURI(uri);
+        this.path = URIs.getFilePath(uri);
         this.openOptions = safe(openOptions);
     }
 
     public FileResource(URI uri, OpenOption... openOptions) {
-        super(uri, MoreFiles.asByteSource(pathFromURI(uri), openOptions));
-        this.path = pathFromURI(uri);
+        super(uri, MoreFiles.asByteSource(URIs.getFilePath(uri), openOptions));
+        this.path = URIs.getFilePath(uri);
         this.openOptions = safe(openOptions);
     }
 
     private static OpenOption[] safe(OpenOption[] openOptions) {
-        if (openOptions.length == 0) return EMPTY_OPTIONS;
+        if (openOptions.length == 0) return EMPTY_OPTIONS; // skipcq: JAVA-S1049
         else return Arrays.copyOf(openOptions, openOptions.length);
     }
 

java/dev/enola/common/io/resource/ResourceProvider.java

@@ -48,10 +48,18 @@ public interface ResourceProvider extends ProviderFromIRI<Resource> {
         return getResource(uri);
     }
 
+    default @Nullable ReadableResource getReadableResource(String iri) {
+        return get(iri);
+    }
+
     default @Nullable WritableResource getWritableResource(URI uri) {
         return getResource(uri);
     }
 
+    default @Nullable WritableResource getWritableResource(String iri) {
+        return get(iri);
+    }
+
     // -------------------------------------------
 
     default @Nullable ReadableResource getReadableResource(URL url) {

java/dev/enola/common/io/resource/stream/FileGlobPathWalker.java

@@ -42,7 +42,6 @@ static Stream<Path> walk(Path globPath) throws IOException {
             if (starPos > 0) basePath = Path.of(globString.substring(0, starPos - 1));
             else basePath = new File(".").toPath();
 
-            // TODO Use URI instead of Path, and FileResource's#pathFromURI() !!!
             // Inspired by File.newDirectoryStream(), but matching full path, not just getFileName()
             var fs = basePath.getFileSystem();
             var matcher = fs.getPathMatcher("glob:" + globString);