C++ elasticlient

C++ elasticlient library is simple library for simplified work with Elasticsearch in C++. The library is based on C++ Requests: Curl for People.

Features

Elasticsearch client which work with unlimited nodes in one Elasticsearch cluster. If any node is dead it tries another one.
Elasticsearch client supports search, index, get, remove methods by default.
Posibility to perform not implemented method i.e multi GET or indices creation.
Support for Bulk API requests.
Support for Scroll API.

Dependencies

C++ Requests: Curl for People
JsonCpp
Google Test
Only for tests: C++ HTTP mock server library

Requirements

C++11 compatible compiler such as GCC (tested with version 4.9.2)
CMake (tested with version 3.5.2)

Building and testing on Unix like system

Step 1

Clone or download this repository to some directory and go to this directory.

git clone https://github.com/seznam/elasticlient
cd elasticlient

Step 2

Now if you want to use all above mentioned dependencies from system, you can skip this step.

git submodule update --init --recursive

Step 3

Build the library:

mkdir build
cd build
cmake ..
make
make test  # Optional, will run elasticlient tests

Following CMake configuration variables may be passed right before .. in cmake .. command.

-DUSE_ALL_SYSTEM_LIBS=YES - use all dependencies from system (default=NO)
-DUSE_SYSTEM_CPR=YES - use C++ Requests library from system (default=NO)
-DUSE_SYSTEM_JSONCPP=YES - use JsonCpp library from system (default=NO)
-DUSE_SYSTEM_GTEST=YES - use Google Test library from system (default=NO)
-DUSE_SYSTEM_HTTPMOCKSERVER=YES - use C++ HTTP mock server library from system (default=NO)
-DBUILD_ELASTICLIENT_TESTS=YES - build elasticlient library tests (default=YES)
-DBUILD_ELASTICLIENT_EXAMPLE=YES - build elasticlient library example hello-world program (default=YES)

How to use

Basic Hello world example

If you build the library with -DBUILD_ELASTICLIENT_EXAMPLE=YES you can find compiled binary (hello-world) of this example in build/bin directory.

This example will at first index the document into elasticsearch cluster. After that it will retrieve the same document and on the end it will remove the document.

#include <string>
#include <vector>
#include <iostream>
#include <cpr/response.h>
#include <elasticlient/client.h>


int main() {
    // Prepare Client for nodes of one Elasticsearch cluster
    elasticlient::Client client({"http://elastic1.host:9200/"}); // last / is mandatory

    std::string document {"{\"message\": \"Hello world!\"}"};

    // Index the document, index "testindex" must be created before
    cpr::Response indexResponse = client.index("testindex", "docType", "docId", document);
    // 200
    std::cout << indexResponse.status_code << std::endl;
    // application/json; charset=UTF-8
    std::cout << indexResponse.header["content-type"] << std::endl;
    // Elasticsearch response (JSON text string)
    std::cout << indexResponse.text << std::endl;

    // Retrieve the document
    cpr::Response retrievedDocument = client.get("testindex", "docType", "docId");
    // 200
    std::cout << retrievedDocument.status_code << std::endl;
    // application/json; charset=UTF-8
    std::cout << retrievedDocument.header["content-type"] << std::endl;
    // Elasticsearch response (JSON text string) where key "_source" contain:
    // {"message": "Hello world!"}
    std::cout << retrievedDocument.text << std::endl;

    // Remove the document
    cpr::Response removedDocument = client.remove("testindex", "docType", "docId");
    // 200
    std::cout << removedDocument.status_code << std::endl;
    // application/json; charset=UTF-8
    std::cout << removedDocument.header["content-type"] << std::endl;
    // Elasticsearch response (JSON text string)
    std::cout << removedDocument.text << std::endl;

    return 0;
}

Usage of Bulk indexer

#include <memory>
#include <string>
#include <vector>
#include <iostream>
#include <elasticlient/client.h>
#include <elasticlient/bulk.h>


int main() {
    // Prepare Client for nodes of one Elasticsearch cluster
    std::shared_ptr<elasticlient::Client> client = std::make_shared<elasticlient::Client>(
        std::vector<std::string>({"http://elastic1.host:9200/"}));  // last / is mandatory
    elasticlient::Bulk bulkIndexer(client);

    elasticlient::SameIndexBulkData bulk("testindex", 100);
    bulk.indexDocument("docType", "docId0", "{\"data\": \"data0\"}");
    bulk.indexDocument("docType", "docId1", "{\"data\": \"data1\"}");
    bulk.indexDocument("docType", "docId2", "{\"data\": \"data2\"}");
    // another unlimited amount of indexDocument() calls...

    size_t errors = bulkIndexer.perform(bulk);
    std::cout << "When indexing " << bulk.size() << " documents, "
              << errors << " errors occured" << std::endl;
    bulk.clear();
    return 0;
}

Usage of Scroll API

#include <memory>
#include <string>
#include <vector>
#include <iostream>
#include <json/json.h>
#include <elasticlient/client.h>
#include <elasticlient/scroll.h>


int main() {
    // Prepare Client for nodes of one Elasticsearch cluster
    std::shared_ptr<elasticlient::Client> client = std::make_shared<elasticlient::Client>(
        std::vector<std::string>({"http://elastic1.host:9200/"}));  // last / is mandatory
    elasticlient::Scroll scrollInstance(client);

    std::string searchQuery{"{\"sort\": [\"_doc\"],"
                            " \"query\": {"
                            "   \"prefix\": {"
                            "     \"data\": \"data\"}}}"};

    // Scroll all documents of type: docType from testindex which corresponding searchQuery
    scrollInstance.init("testindex", "docType", searchQuery);

    Json::StyledWriter writer;

    Json::Value res;
    bool isSuccessful = true;
    // Will scroll for all suitable documents
    while ((isSuccessful = scrollInstance.next(res))) {
        if (res["hits"].empty()) {
            // last scroll, no more results
            break;
        }
        std::cout << writer.write(res["hits"]) << std::endl;
    }

    scrollInstance.clear();
}

How to use logging feature (debugging)

#include <iostream>
#include <elasticlient/client.h>
#include <elasticlient/logging.h>


/// Very simple log callback (only print message to stdout)
void logCallback(elasticlient::LogLevel logLevel, const std::string &msg) {
    if (logLevel != elasticlient::LogLevel::DEBUG) {
        std::cout << "LOG " << (unsigned) logLevel << ": " << msg << std::endl;
    }
}


int main() {
    // Set logging function for elasticlient library
    elasticlient::setLogFunction(logCallback);

    // Elasticlient will now print all debug messages on stdout.
    // It is necessary to set logging function for elasticlient for each thread where you want
    // use this logging feature!
}

License

Elasticlient is licensed under the MIT License (MIT). Only the cmake/Modules/FindJsonCpp.cmake is originally licensed under Boost Software License, for more information see header of the file.

linecode / elasticlient Goto Github PK

elasticlient's Introduction