diff options
author | Jon Marius Venstad <jonmv@gmail.com> | 2022-04-07 15:31:44 +0200 |
---|---|---|
committer | Jon Marius Venstad <jonmv@gmail.com> | 2022-04-07 15:31:44 +0200 |
commit | c00d72bc4cdde1bde164e6174ab854f1ef4e7bfd (patch) | |
tree | 1cfe2296adc3ef7324b0c3238a5b7f0dc04ba7bb /vespajlib/src/main/java/ai | |
parent | f2850bb028ff11c91adb370f5bd03899ac5022ec (diff) |
Resolve conflicts
Diffstat (limited to 'vespajlib/src/main/java/ai')
-rw-r--r-- | vespajlib/src/main/java/ai/vespa/http/DomainName.java | 35 | ||||
-rw-r--r-- | vespajlib/src/main/java/ai/vespa/http/HttpURL.java | 458 | ||||
-rw-r--r-- | vespajlib/src/main/java/ai/vespa/http/package-info.java | 5 |
3 files changed, 498 insertions, 0 deletions
diff --git a/vespajlib/src/main/java/ai/vespa/http/DomainName.java b/vespajlib/src/main/java/ai/vespa/http/DomainName.java new file mode 100644 index 00000000000..f68ee8ffbcd --- /dev/null +++ b/vespajlib/src/main/java/ai/vespa/http/DomainName.java @@ -0,0 +1,35 @@ +// Copyright Yahoo. Licensed under the terms of the Apache 2.0 license. See LICENSE in the project root. +package ai.vespa.http; + +import ai.vespa.validation.PatternedStringWrapper; + +import java.util.regex.Pattern; + +import static ai.vespa.validation.Validation.requireLength; +import static ai.vespa.validation.Validation.requireMatch; + +/** + * A valid domain name, which can be used in a {@link java.net.URI}. + * + * @author jonmv + */ +public class DomainName extends PatternedStringWrapper<DomainName> { + + public static final Pattern labelPattern = Pattern.compile("([A-Za-z0-9]|[A-Za-z0-9][A-Za-z0-9-]{0,61}[A-Za-z0-9])"); + public static final Pattern domainNamePattern = Pattern.compile("(" + labelPattern + "\\.)*" + labelPattern); + + public static final DomainName localhost = DomainName.of("localhost"); + + private DomainName(String value) { + super(requireLength(value, "domain name length", 1, 255), domainNamePattern, "domain name"); + } + + public static DomainName of(String value) { + return new DomainName(value); + } + + public static String requireLabel(String label) { + return requireMatch(label, "domain name label", labelPattern); + } + +} diff --git a/vespajlib/src/main/java/ai/vespa/http/HttpURL.java b/vespajlib/src/main/java/ai/vespa/http/HttpURL.java new file mode 100644 index 00000000000..607d72f2550 --- /dev/null +++ b/vespajlib/src/main/java/ai/vespa/http/HttpURL.java @@ -0,0 +1,458 @@ +// Copyright Yahoo. Licensed under the terms of the Apache 2.0 license. See LICENSE in the project root. +package ai.vespa.http; + +import ai.vespa.validation.StringWrapper; + +import java.net.URI; +import java.net.URISyntaxException; +import java.util.ArrayList; +import java.util.Collections; +import java.util.LinkedHashMap; +import java.util.List; +import java.util.Map; +import java.util.Objects; +import java.util.OptionalInt; +import java.util.StringJoiner; +import java.util.function.Consumer; +import java.util.function.UnaryOperator; + +import static ai.vespa.validation.Validation.require; +import static ai.vespa.validation.Validation.requireInRange; +import static java.net.URLDecoder.decode; +import static java.net.URLEncoder.encode; +import static java.nio.charset.StandardCharsets.UTF_8; +import static java.util.Collections.unmodifiableMap; +import static java.util.Objects.requireNonNull; + +/** + * This is the best class for creating, manipulating and inspecting HTTP URLs, because: + * <ul> + * <li>It is more restrictive than {@link URI}, but with a richer construction API, reducing risk of blunder. + * <ul> + * <li>Scheme must be HTTP or HTTPS.</li> + * <li>Authority must be a {@link DomainName}, with an optional port.</li> + * <li>{@link Path} must be normalized at all times.</li> + * <li>Only {@link Query} is allowed, in addition to the above.</li> + * </ul> + * </li> + * <li> + * It contains all those helpful builder methods that {@link URI} has none of. + * <ul> + * <li>{@link Path} can be parsed, have segments or other paths appended, and cut.</li> + * <li>{@link Query} can be parsed, and keys and key-value pairs can be inserted or removed.</li> + * </ul> + * All these (except the parse methods) operate on <em>decoded</em> values. + * </li> + * <li>It makes it super-easy to use a {@link StringWrapper} for validation of path and query segments.</li> + * </ul> + * + * @author jonmv + */ +public class HttpURL { + + private final Scheme scheme; + private final DomainName domain; + private final int port; + private final Path path; + private final Query query; + + private HttpURL(Scheme scheme, DomainName domain, int port, Path path, Query query) { + this.scheme = requireNonNull(scheme); + this.domain = requireNonNull(domain); + this.port = requireInRange(port, "port number", -1, (1 << 16) - 1); + this.path = requireNonNull(path); + this.query = requireNonNull(query); + } + + public static HttpURL create(Scheme scheme, DomainName domain, int port, Path path, Query query) { + return new HttpURL(scheme, domain, port, path, query); + } + + public static HttpURL create(Scheme scheme, DomainName domain, int port, Path path) { + return create(scheme, domain, port, path, Query.empty()); + } + + public static HttpURL create(Scheme scheme, DomainName domain, int port) { + return create(scheme, domain, port, Path.empty(), Query.empty()); + } + + public static HttpURL create(Scheme scheme, DomainName domain) { + return create(scheme, domain, -1); + } + + public static HttpURL from(URI uri) { + return from(uri, HttpURL::requirePathSegment, HttpURL::requireNothing); + } + + public static HttpURL from(URI uri, Consumer<String> pathValidator, Consumer<String> queryValidator) { + if ( ! uri.normalize().equals(uri)) + throw new IllegalArgumentException("uri should be normalized, but got: " + uri); + + return create(Scheme.of(uri.getScheme()), + DomainName.of(requireNonNull(uri.getHost(), "URI must specify a host")), + uri.getPort(), + Path.parse(uri.getRawPath(), pathValidator), + Query.parse(uri.getRawQuery(), queryValidator)); + } + + public HttpURL withScheme(Scheme scheme) { + return create(scheme, domain, port, path, query); + } + + public HttpURL withDomain(DomainName domain) { + return create(scheme, domain, port, path, query); + } + + public HttpURL withPort(int port) { + return create(scheme, domain, port, path, query); + } + + public HttpURL withoutPort() { + return create(scheme, domain, -1, path, query); + } + + public HttpURL withPath(Path path) { + return create(scheme, domain, port, path, query); + } + + public HttpURL appendPath(Path path) { + return create(scheme, domain, port, this.path.append(path), query); + } + + public HttpURL withQuery(Query query) { + return create(scheme, domain, port, path, query); + } + + public HttpURL mergeQuery(Query query) { + return create(scheme, domain, port, path, this.query.merge(query)); + } + + public Scheme scheme() { + return scheme; + } + + public DomainName domain() { + return domain; + } + + public OptionalInt port() { + return port == -1 ? OptionalInt.empty() : OptionalInt.of(port); + } + + public Path path() { + return path; + } + + public Query query() { + return query; + } + + /** Returns an absolute, hierarchical URI representing this HTTP URL. */ + public URI asURI() { + try { + return new URI(scheme.name() + "://" + domain.value() + (port == -1 ? "" : ":" + port) + path.raw() + query.raw()); + } + catch (URISyntaxException e) { + throw new IllegalStateException("invalid URI, this should not happen", e); + } + } + + /** Require that the given string (possibly decoded multiple times) contains none of {@code '/', '?', '#'}, and isn't either of {@code "", ".", ".."}. */ + public static String requirePathSegment(String value) { + while ( ! value.equals(value = decode(value, UTF_8))); + require( ! value.contains("/"), value, "path segment decoded cannot contain '/'"); + require( ! value.contains("?"), value, "path segment decoded cannot contain '?'"); + require( ! value.contains("#"), value, "path segment decoded cannot contain '#'"); + return Path.requireNonNormalizable(value); + } + + private static void requireNothing(String value) { } + + public static class Path { + + private final List<String> segments; + private final boolean trailingSlash; + private final UnaryOperator<String> validator; + + private Path(List<String> segments, boolean trailingSlash, UnaryOperator<String> validator) { + this.segments = requireNonNull(segments); + this.trailingSlash = trailingSlash; + this.validator = requireNonNull(validator); + } + + /** Creates a new, empty path, with a trailing slash, using {@link HttpURL#requirePathSegment} for segment validation. */ + public static Path empty() { + return empty(HttpURL::requirePathSegment); + } + + /** Creates a new, empty path, with a trailing slash, using the indicated validator for segments. */ + public static Path empty(Consumer<String> validator) { + return new Path(List.of(), true, segmentValidator(validator)); + } + + /** Creates a new path with the given <em>decoded</em> segments. */ + public static Path from(List<String> segments) { + return from(segments, __ -> { }); + } + + /** Creates a new path with the given <em>decoded</em> segments, and the validator applied to each segment. */ + public static Path from(List<String> segments, Consumer<String> validator) { + return empty(validator).append(segments, true); + } + + /** Parses the given raw, normalized path string; this ignores whether the path is absolute or relative. */ + public static Path parse(String raw) { + return parse(raw, HttpURL::requirePathSegment); + } + + /** Parses the given raw, normalized path string; this ignores whether the path is absolute or relative.) */ + public static Path parse(String raw, Consumer<String> validator) { + Path base = new Path(List.of(), raw.endsWith("/"), segmentValidator(validator)); + if (raw.startsWith("/")) raw = raw.substring(1); + if (raw.isEmpty()) return base; + List<String> segments = new ArrayList<>(); + for (String segment : raw.split("/")) segments.add(decode(segment, UTF_8)); + if (segments.isEmpty()) requireNonNormalizable(""); // Raw path was only slashes. + return base.append(segments); + } + + private static UnaryOperator<String> segmentValidator(Consumer<String> validator) { + requireNonNull(validator, "segment validator cannot be null"); + return value -> { + requireNonNormalizable(value); + validator.accept(value); + return value; + }; + } + + private static String requireNonNormalizable(String segment) { + return require( ! (segment.isEmpty() || segment.equals(".") || segment.equals("..")), + segment, "path segments cannot be \"\", \".\", or \"..\""); + } + + /** Returns a copy of this where only the first segments are retained, and with a trailing slash. */ + public Path head(int count) { + return count == segments.size() ? this : new Path(segments.subList(0, count), true, validator); + } + + /** Returns a copy of this where only the last segments are retained. */ + public Path tail(int count) { + return count == segments.size() ? this : new Path(segments.subList(segments.size() - count, segments.size()), trailingSlash, validator); + } + + /** Returns a copy of this where the first segments are skipped. */ + public Path skip(int count) { + return count == 0 ? this : new Path(segments.subList(count, segments.size()), trailingSlash, validator); + } + + /** Returns a copy of this where the last segments are cut off, and with a trailing slash. */ + public Path cut(int count) { + return count == 0 ? this : new Path(segments.subList(0, segments.size() - count), true, validator); + } + + /** Returns a copy of this with the <em>decoded</em> segment appended at the end; it may not be either of {@code ""}, {@code "."} or {@code ".."}. */ + public Path append(String segment) { + return append(List.of(segment), trailingSlash); + } + + /** Returns a copy of this all segments of the other path appended, with a trailing slash as per the appendage. */ + public Path append(Path other) { + return append(other.segments, other.trailingSlash); + } + + /** Returns a copy of this all given segments appended, with a trailing slash as per this path. */ + public Path append(List<String> segments) { + return append(segments, trailingSlash); + } + + private Path append(List<String> segments, boolean trailingSlash) { + List<String> copy = new ArrayList<>(this.segments); + for (String segment : segments) copy.add(validator.apply(segment)); + return new Path(copy, trailingSlash, validator); + } + + /** Whether this path has a trailing slash. */ + public boolean hasTrailingSlash() { + return trailingSlash; + } + + /** Returns a copy of this which encodes a trailing slash. */ + public Path withTrailingSlash() { + return new Path(segments, true, validator); + } + + /** Returns a copy of this which does not encode a trailing slash. */ + public Path withoutTrailingSlash() { + return new Path(segments, false, validator); + } + + /** The <em>URL decoded</em> segments that make up this path; never {@code null}, {@code ""}, {@code "."} or {@code ".."}. */ + public List<String> segments() { + return Collections.unmodifiableList(segments); + } + + /** A raw path string which parses to this, by splitting on {@code "/"}, and then URL decoding. */ + private String raw() { + StringJoiner joiner = new StringJoiner("/", "/", trailingSlash ? "/" : "").setEmptyValue(trailingSlash ? "/" : ""); + for (String segment : segments) joiner.add(encode(segment, UTF_8)); + return joiner.toString(); + } + + /** Intentionally not usable for constructing new URIs. Use {@link HttpURL} for that instead. */ + @Override + public String toString() { + return "path '" + raw() + "'"; + } + + @Override + public boolean equals(Object o) { + if (this == o) return true; + if (o == null || getClass() != o.getClass()) return false; + Path path = (Path) o; + return trailingSlash == path.trailingSlash && segments.equals(path.segments); + } + + @Override + public int hashCode() { + return Objects.hash(segments, trailingSlash); + } + + } + + + public static class Query { + + private final Map<String, String> values; + private final UnaryOperator<String> validator; + + private Query(Map<String, String> values, UnaryOperator<String> validator) { + this.values = requireNonNull(values); + this.validator = requireNonNull(validator); + } + + /** Creates a new, empty query part. */ + public static Query empty() { + return empty(__ -> { }); + } + + /** Creates a new, empty query part, using the indicated string wrapper for keys and non-null values. */ + public static Query empty(Consumer<String> validator) { + return new Query(Map.of(), entryValidator(validator)); + } + + /** Creates a new query part with the given <em>decoded</em> values. */ + public static Query from(Map<String, String> values) { + return from(values, __ -> { }); + } + + /** Creates a new query part with the given <em>decoded</em> values, and the validator applied to each pair. */ + public static Query from(Map<String, String> values, Consumer<String> validator) { + return empty(validator).merge(values); + } + + /** Parses the given raw query string. */ + public static Query parse(String raw) { + return parse(raw, __-> { }); + } + /** Parses the given raw query string, using the indicated string wrapper to hold keys and non-null values. */ + public static Query parse(String raw, Consumer<String> validator) { + if (raw == null) return empty(validator); + Map<String, String> values = new LinkedHashMap<>(); + for (String pair : raw.split("&")) { + int split = pair.indexOf("="); + String key, value; + if (split == -1) { key = pair; value = null; } + else { key = pair.substring(0, split); value = pair.substring(split + 1); } + values.put(decode(key, UTF_8), value == null ? null : decode(value, UTF_8)); + } + return empty(validator).merge(values); + } + + private static UnaryOperator<String> entryValidator(Consumer<String> validator) { + requireNonNull(validator); + return value -> { + validator.accept(value); + return value; + }; + } + + /** Returns a copy of this with the <em>decoded</em> non-null key pointing to the <em>decoded</em> non-null value. */ + public Query put(String key, String value) { + Map<String, String> copy = new LinkedHashMap<>(values); + copy.put(validator.apply(requireNonNull(key)), validator.apply(requireNonNull(value))); + return new Query(copy, validator); + } + + /** Returns a copy of this with the <em>decoded</em> non-null key pointing to "nothing". */ + public Query add(String key) { + Map<String, String> copy = new LinkedHashMap<>(values); + copy.put(validator.apply(requireNonNull(key)), null); + return new Query(copy, validator); + } + + /** Returns a copy of this without any key-value pair with the <em>decoded</em> key. */ + public Query remove(String key) { + Map<String, String> copy = new LinkedHashMap<>(values); + copy.remove(validator.apply(requireNonNull(key))); + return new Query(copy, validator); + } + + /** Returns a copy of this with all mappings from the other query added to this, possibly overwriting existing mappings. */ + public Query merge(Query other) { + return merge(other.values); + } + + /** Returns a copy of this with all given mappings added to this, possibly overwriting existing mappings. */ + public Query merge(Map<String, String> values) { + Map<String, String> copy = new LinkedHashMap<>(this.values); + values.forEach((key, value) -> copy.put(validator.apply(requireNonNull(key, "keys cannot be null")), + value == null ? null : validator.apply(value))); + return new Query(copy, validator); + } + + /** The <em>URL decoded</em> key-value pairs that make up this query; keys and values may be {@code ""}, and values are {@code null} when only key was specified. */ + public Map<String, String> entries() { + return unmodifiableMap(values); + } + + /** A raw query string, with {@code '?'} prepended, that parses to this, by splitting on {@code "&"}, then on {@code "="}, and then URL decoding; or the empty string if this is empty. */ + private String raw() { + StringJoiner joiner = new StringJoiner("&", "?", "").setEmptyValue(""); + values.forEach((key, value) -> joiner.add(encode(key, UTF_8) + + (value == null ? "" : "=" + encode(value, UTF_8)))); + return joiner.toString(); + } + + /** Intentionally not usable for constructing new URIs. Use {@link HttpURL} for that instead. */ + @Override + public String toString() { + return "query '" + raw() + "'"; + } + + @Override + public boolean equals(Object o) { + if (this == o) return true; + if (o == null || getClass() != o.getClass()) return false; + Query query = (Query) o; + return values.equals(query.values); + } + + @Override + public int hashCode() { + return Objects.hash(values); + } + + } + + + public enum Scheme { + http, + https; + public static Scheme of(String scheme) { + if (scheme.equalsIgnoreCase(http.name())) return http; + if (scheme.equalsIgnoreCase(https.name())) return https; + throw new IllegalArgumentException("scheme must be HTTP or HTTPS"); + } + } + +} diff --git a/vespajlib/src/main/java/ai/vespa/http/package-info.java b/vespajlib/src/main/java/ai/vespa/http/package-info.java new file mode 100644 index 00000000000..e5600c6f82d --- /dev/null +++ b/vespajlib/src/main/java/ai/vespa/http/package-info.java @@ -0,0 +1,5 @@ +// Copyright Yahoo. Licensed under the terms of the Apache 2.0 license. See LICENSE in the project root. +@ExportPackage +package ai.vespa.http; + +import com.yahoo.osgi.annotation.ExportPackage; |