1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
|
// Copyright 2017 Yahoo Holdings. Licensed under the terms of the Apache 2.0 license. See LICENSE in the project root.
package com.yahoo.vespa.http.client;
import com.yahoo.vespa.http.client.core.JsonReader;
import com.yahoo.vespa.http.client.core.XmlFeedReader;
import java.io.InputStream;
import java.util.concurrent.atomic.AtomicInteger;
/**
* API for feeding document operations (add, removes or updates) to one or many Vespa clusters.
* Use the factory to configure and set up an instance of this API.
* The feedclient does automatically error recovery and reconnects to hosts when
* connections die.
*
* A {@link FeedClientFactory} is provided to instantiate Sessions.
*
* @author dybis
* @see FeedClientFactory
* @see com.yahoo.text.Text#stripInvalidCharacters(String) to remove invalid characters from string fields before feeding
*/
public interface FeedClient extends AutoCloseable {
/**
* Streams a document to cluster(s). If the pipeline and buffers are full, this call will be blocking.
* Documents might time out before they are sent. Failed documents are not retried.
* Don't call stream() after close is called.
*
* @param documentId Document id of the document.
* @param documentData The document data as JSON or XML (as specified when using the factory to create the API)
*/
void stream(String documentId, CharSequence documentData);
/**
* Streams a document to cluster(s). If the pipeline and buffers are full, this call will be blocking.
* Documents might time out before they are sent. Failed documents are not retried.
* Don't call stream() after close is called.
*
* @param documentId Document id of the document.
* @param documentData The document data as JSON or XML (as specified when using the factory to create the API)
* @param context Any context, will be accessible in the result of the callback.
*/
void stream(String documentId, CharSequence documentData, Object context);
/**
* This callback is executed when new results are arriving or an error occur.
* Don't do any heavy lifting in this thread (no IO, disk, or heavy CPU usage).
* This call back will run in a different thread than your main program so use e.g.
* AtomicInteger for counters and follow general guides for thread-safe programming.
* There is an example implementation in class SimpleLoggerResultCallback.
*/
interface ResultCallback {
void onCompletion(String docId, Result documentResult);
/**
* Called with an exception whenever an endpoint specific error occurs during feeding.
* The error may or may not be transient - the operation will in both cases be retried until it's successful.
* This callback is intended for application level monitoring (logging, metrics, altering etc).
* Document specific errors will be reported back through {@link #onCompletion(String, Result)}.
*
* @see FeedEndpointException
* @param exception An exception specifying endpoint and cause. See {@link FeedEndpointException} for details.
*/
// TODO Vespa 7: Remove empty default implementation
default void onEndpointException(FeedEndpointException exception) {}
}
/**
* Waits for all results to arrive and closes the FeedClient. Don't call any other method after calling close().
* Does not throw any exceptions.
*/
@Override
void close();
/**
* Returns stats about the cluster.
*
* @return JSON string with information about cluster.
*/
String getStatsAsJson();
/**
* Utility function that takes an array of JSON documents and calls the FeedClient for each element.
*
* @param inputStream This can be a very large stream. The outer element is an array (of document operations).
* @param feedClient The feedClient that will receive the document operations.
* @param numSent increased per document sent to API (but no waiting for results).
*/
static void feedJson(InputStream inputStream, FeedClient feedClient, AtomicInteger numSent) {
JsonReader.read(inputStream, feedClient, numSent);
}
/**
* Utility function that takes an array of XML documents and calls the FeedClient for each element.
* The XML document has to be formatted with line space on each line (like "regular" XML, but stricter
* than the specifications of XML).
*
* @param inputStream This can be a very large stream.
* @param feedClient The feedClient that will receive the document operations.
* @param numSent increased per document sent to API (but no waiting for results).
*/
static void feedXml(InputStream inputStream, FeedClient feedClient, AtomicInteger numSent) {
try {
XmlFeedReader.read(inputStream, feedClient, numSent);
} catch (Exception e) {
throw new RuntimeException(e);
}
}
}
|