[HLRC][ML] Add ML find file structure API (#35833)

Relates to #29827
elastic · Nov 23, 2018 · d0b5006 · d0b5006
1 parent c17fa7f
commit d0b5006
Show file tree

Hide file tree

Showing 14 changed files with 1,786 additions and 0 deletions.
diff --git a/client/rest-high-level/src/main/java/org/elasticsearch/client/MLRequestConverters.java b/client/rest-high-level/src/main/java/org/elasticsearch/client/MLRequestConverters.java
@@ -36,6 +36,7 @@
 import org.elasticsearch.client.ml.DeleteForecastRequest;
 import org.elasticsearch.client.ml.DeleteJobRequest;
 import org.elasticsearch.client.ml.DeleteModelSnapshotRequest;
+import org.elasticsearch.client.ml.FindFileStructureRequest;
 import org.elasticsearch.client.ml.FlushJobRequest;
 import org.elasticsearch.client.ml.ForecastJobRequest;
 import org.elasticsearch.client.ml.GetBucketsRequest;
@@ -70,6 +71,7 @@
 import org.elasticsearch.client.ml.job.util.PageParams;
 import org.elasticsearch.common.Strings;
 import org.elasticsearch.common.bytes.BytesReference;
+import org.elasticsearch.common.xcontent.XContentType;
 
 import java.io.IOException;
 
@@ -648,4 +650,65 @@ static Request deleteFilter(DeleteFilterRequest deleteFilterRequest) {
         Request request = new Request(HttpDelete.METHOD_NAME, endpoint);
         return request;
     }
+
+    static Request findFileStructure(FindFileStructureRequest findFileStructureRequest) {
+        String endpoint = new EndpointBuilder()
+            .addPathPartAsIs("_xpack")
+            .addPathPartAsIs("ml")
+            .addPathPartAsIs("find_file_structure")
+            .build();
+        Request request = new Request(HttpPost.METHOD_NAME, endpoint);
+
+        RequestConverters.Params params = new RequestConverters.Params(request);
+        if (findFileStructureRequest.getLinesToSample() != null) {
+            params.putParam(FindFileStructureRequest.LINES_TO_SAMPLE.getPreferredName(),
+                findFileStructureRequest.getLinesToSample().toString());
+        }
+        if (findFileStructureRequest.getTimeout() != null) {
+            params.putParam(FindFileStructureRequest.TIMEOUT.getPreferredName(), findFileStructureRequest.getTimeout().toString());
+        }
+        if (findFileStructureRequest.getCharset() != null) {
+            params.putParam(FindFileStructureRequest.CHARSET.getPreferredName(), findFileStructureRequest.getCharset());
+        }
+        if (findFileStructureRequest.getFormat() != null) {
+            params.putParam(FindFileStructureRequest.FORMAT.getPreferredName(), findFileStructureRequest.getFormat().toString());
+        }
+        if (findFileStructureRequest.getColumnNames() != null) {
+            params.putParam(FindFileStructureRequest.COLUMN_NAMES.getPreferredName(),
+                Strings.collectionToCommaDelimitedString(findFileStructureRequest.getColumnNames()));
+        }
+        if (findFileStructureRequest.getHasHeaderRow() != null) {
+            params.putParam(FindFileStructureRequest.HAS_HEADER_ROW.getPreferredName(),
+                findFileStructureRequest.getHasHeaderRow().toString());
+        }
+        if (findFileStructureRequest.getDelimiter() != null) {
+            params.putParam(FindFileStructureRequest.DELIMITER.getPreferredName(),
+                findFileStructureRequest.getDelimiter().toString());
+        }
+        if (findFileStructureRequest.getQuote() != null) {
+            params.putParam(FindFileStructureRequest.QUOTE.getPreferredName(), findFileStructureRequest.getQuote().toString());
+        }
+        if (findFileStructureRequest.getShouldTrimFields() != null) {
+            params.putParam(FindFileStructureRequest.SHOULD_TRIM_FIELDS.getPreferredName(),
+                findFileStructureRequest.getShouldTrimFields().toString());
+        }
+        if (findFileStructureRequest.getGrokPattern() != null) {
+            params.putParam(FindFileStructureRequest.GROK_PATTERN.getPreferredName(), findFileStructureRequest.getGrokPattern());
+        }
+        if (findFileStructureRequest.getTimestampFormat() != null) {
+            params.putParam(FindFileStructureRequest.TIMESTAMP_FORMAT.getPreferredName(), findFileStructureRequest.getTimestampFormat());
+        }
+        if (findFileStructureRequest.getTimestampField() != null) {
+            params.putParam(FindFileStructureRequest.TIMESTAMP_FIELD.getPreferredName(), findFileStructureRequest.getTimestampField());
+        }
+        if (findFileStructureRequest.getExplain() != null) {
+            params.putParam(FindFileStructureRequest.EXPLAIN.getPreferredName(), findFileStructureRequest.getExplain().toString());
+        }
+
+        BytesReference sample = findFileStructureRequest.getSample();
+        BytesRef source = sample.toBytesRef();
+        HttpEntity byteEntity = new ByteArrayEntity(source.bytes, source.offset, source.length, createContentType(XContentType.JSON));
+        request.setEntity(byteEntity);
+        return request;
+    }
 }
diff --git a/client/rest-high-level/src/main/java/org/elasticsearch/client/MachineLearningClient.java b/client/rest-high-level/src/main/java/org/elasticsearch/client/MachineLearningClient.java
@@ -31,6 +31,8 @@
 import org.elasticsearch.client.ml.DeleteJobRequest;
 import org.elasticsearch.client.ml.DeleteJobResponse;
 import org.elasticsearch.client.ml.DeleteModelSnapshotRequest;
+import org.elasticsearch.client.ml.FindFileStructureRequest;
+import org.elasticsearch.client.ml.FindFileStructureResponse;
 import org.elasticsearch.client.ml.FlushJobRequest;
 import org.elasticsearch.client.ml.FlushJobResponse;
 import org.elasticsearch.client.ml.ForecastJobRequest;
@@ -1711,4 +1713,45 @@ public void deleteFilterAsync(DeleteFilterRequest request, RequestOptions option
             listener,
             Collections.emptySet());
     }
+
+    /**
+     * Finds the structure of a file
+     * <p>
+     * For additional info
+     * see <a href="http://www.elastic.co/guide/en/elasticsearch/reference/current/ml-find-file-structure.html">
+     *     ML Find File Structure documentation</a>
+     *
+     * @param request The find file structure request
+     * @param options  Additional request options (e.g. headers), use {@link RequestOptions#DEFAULT} if nothing needs to be customized
+     * @return the response containing details of the file structure
+     * @throws IOException when there is a serialization issue sending the request or receiving the response
+     */
+    public FindFileStructureResponse findFileStructure(FindFileStructureRequest request, RequestOptions options) throws IOException {
+        return restHighLevelClient.performRequestAndParseEntity(request,
+            MLRequestConverters::findFileStructure,
+            options,
+            FindFileStructureResponse::fromXContent,
+            Collections.emptySet());
+    }
+
+    /**
+     * Finds the structure of a file asynchronously and notifies the listener on completion
+     * <p>
+     * For additional info
+     * see <a href="http://www.elastic.co/guide/en/elasticsearch/reference/current/ml-find-file-structure.html">
+     *         ML Find File Structure documentation</a>
+     *
+     * @param request The find file structure request
+     * @param options  Additional request options (e.g. headers), use {@link RequestOptions#DEFAULT} if nothing needs to be customized
+     * @param listener Listener to be notified upon request completion
+     */
+    public void findFileStructureAsync(FindFileStructureRequest request, RequestOptions options,
+                                       ActionListener<FindFileStructureResponse> listener) {
+        restHighLevelClient.performRequestAsyncAndParseEntity(request,
+            MLRequestConverters::findFileStructure,
+            options,
+            FindFileStructureResponse::fromXContent,
+            listener,
+            Collections.emptySet());
+    }
 }