Coverage Report

Created: 2025-07-27 00:06

/root/doris/be/src/http/http_client.h
Line
Count
Source (jump to first uncovered line)
1
// Licensed to the Apache Software Foundation (ASF) under one
2
// or more contributor license agreements.  See the NOTICE file
3
// distributed with this work for additional information
4
// regarding copyright ownership.  The ASF licenses this file
5
// to you under the Apache License, Version 2.0 (the
6
// "License"); you may not use this file except in compliance
7
// with the License.  You may obtain a copy of the License at
8
//
9
//   http://www.apache.org/licenses/LICENSE-2.0
10
//
11
// Unless required by applicable law or agreed to in writing,
12
// software distributed under the License is distributed on an
13
// "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
14
// KIND, either express or implied.  See the License for the
15
// specific language governing permissions and limitations
16
// under the License.
17
18
#pragma once
19
20
#include <curl/curl.h>
21
#include <curl/system.h>
22
#include <stdint.h>
23
24
#include <cstdio>
25
#include <functional>
26
#include <string>
27
#include <unordered_set>
28
29
#include "common/status.h"
30
#include "http/http_method.h"
31
32
namespace doris {
33
34
// Helper class to access HTTP resource
35
class HttpClient {
36
public:
37
    HttpClient();
38
    ~HttpClient();
39
40
    // you can call this function to execute HTTP request with retry,
41
    // if callback return OK, this function will end and return OK.
42
    // This function will return FAIL if three are more than retry_times
43
    // that callback return FAIL.
44
    static Status execute_with_retry(int retry_times, int sleep_time,
45
                                     const std::function<Status(HttpClient*)>& callback);
46
47
    // this function must call before other function,
48
    // you can call this multiple times to reuse this object
49
    Status init(const std::string& url, bool set_fail_on_error = true);
50
51
    void set_method(HttpMethod method);
52
53
8
    void set_basic_auth(const std::string& user, const std::string& passwd) {
54
8
        curl_easy_setopt(_curl, CURLOPT_HTTPAUTH, CURLAUTH_BASIC);
55
8
        curl_easy_setopt(_curl, CURLOPT_USERNAME, user.c_str());
56
8
        curl_easy_setopt(_curl, CURLOPT_PASSWORD, passwd.c_str());
57
8
    }
58
59
    // content_type such as "application/json"
60
0
    void set_content_type(const std::string content_type) {
61
0
        std::string scratch_str = "Content-Type: " + content_type;
62
0
        _header_list = curl_slist_append(_header_list, scratch_str.c_str());
63
0
        curl_easy_setopt(_curl, CURLOPT_HTTPHEADER, _header_list);
64
0
    }
65
66
11
    void set_payload(const std::string& post_body) {
67
11
        curl_easy_setopt(_curl, CURLOPT_POSTFIELDSIZE, (long)post_body.length());
68
11
        curl_easy_setopt(_curl, CURLOPT_COPYPOSTFIELDS, post_body.c_str());
69
11
    }
70
71
    // Currently, only fake SSL configurations are supported
72
0
    void use_untrusted_ssl() {
73
0
        curl_easy_setopt(_curl, CURLOPT_SSL_VERIFYPEER, 0L);
74
0
        curl_easy_setopt(_curl, CURLOPT_SSL_VERIFYHOST, 0L);
75
0
    }
76
77
    void set_speed_limit();
78
79
    // TODO(zc): support set header
80
    // void set_header(const std::string& key, const std::string& value) {
81
    // _cntl.http_request().SetHeader(key, value);
82
    // }
83
84
0
    std::string get_response_content_type() {
85
0
        char* ct = nullptr;
86
0
        auto code = curl_easy_getinfo(_curl, CURLINFO_CONTENT_TYPE, &ct);
87
0
        if (code == CURLE_OK && ct != nullptr) {
88
0
            return ct;
89
0
        }
90
0
        return std::string();
91
0
    }
92
93
    // Set the long gohead parameter to 1L to continue send authentication (user+password)
94
    // credentials when following locations, even when hostname changed.
95
0
    void set_unrestricted_auth(int gohead) {
96
0
        curl_easy_setopt(_curl, CURLOPT_UNRESTRICTED_AUTH, gohead);
97
0
    }
98
99
6
    void set_timeout_ms(int64_t timeout_ms) {
100
6
        curl_easy_setopt(_curl, CURLOPT_TIMEOUT_MS, timeout_ms);
101
6
    }
102
103
    // used to get content length
104
    // return -1 as error
105
3
    Status get_content_length(uint64_t* length) const {
106
3
        curl_off_t cl;
107
3
        auto code = curl_easy_getinfo(_curl, CURLINFO_CONTENT_LENGTH_DOWNLOAD_T, &cl);
108
3
        if (!code) {
109
3
            if (cl < 0) {
110
0
                return Status::InternalError(
111
0
                        fmt::format("failed to get content length, it should be a positive value, "
112
0
                                    "actual is : {}",
113
0
                                    cl));
114
0
            }
115
3
            *length = (uint64_t)cl;
116
3
            return Status::OK();
117
3
        }
118
0
        return Status::InternalError("failed to get content length. err code: {}", code);
119
3
    }
120
121
    // Get the value of the header CONTENT-MD5. The output is empty if no such header exists.
122
    Status get_content_md5(std::string* md5) const;
123
124
3
    long get_http_status() const {
125
3
        long code;
126
3
        curl_easy_getinfo(_curl, CURLINFO_RESPONSE_CODE, &code);
127
3
        return code;
128
3
    }
129
130
    // execute a head method
131
0
    Status head() {
132
0
        set_method(HEAD);
133
0
        return execute();
134
0
    }
135
136
    // helper function to download a file, you can call this function to download
137
    // a file to local_path
138
    Status download(const std::string& local_path);
139
    Status download_multi_files(const std::string& local_dir,
140
                                const std::unordered_set<std::string>& expected_files);
141
142
    Status execute_post_request(const std::string& payload, std::string* response);
143
144
    Status execute_delete_request(const std::string& payload, std::string* response);
145
146
    // execute a simple method, and its response is saved in response argument
147
    Status execute(std::string* response);
148
149
    // execute remote call action
150
    Status execute(const std::function<bool(const void* data, size_t length)>& callback = {});
151
152
    // execute remote call action with retry, like execute_with_retry but keep the http client instance
153
    Status execute(int retry_times, int sleep_time,
154
                   const std::function<Status(HttpClient*)>& callback);
155
156
    size_t on_response_data(const void* data, size_t length);
157
158
    // The file name of the variant column with the inverted index contains %
159
    // such as: 020000000000003f624c4c322c568271060f9b5b274a4a95_0_10133@properties%2Emessage.idx
160
    //  {rowset_id}_{seg_num}_{index_id}_{variant_column_name}{%2E}{extracted_column_name}.idx
161
    // We need to handle %, otherwise it will cause an HTTP 404 error.
162
    // Because the percent ("%") character serves as the indicator for percent-encoded octets,
163
    // it must be percent-encoded as "%25" for that octet to be used as data within a URI.
164
    // https://datatracker.ietf.org/doc/html/rfc3986
165
    Status _escape_url(const std::string& url, std::string* escaped_url);
166
167
private:
168
    const char* _to_errmsg(CURLcode code) const;
169
    const char* _get_url() const;
170
171
private:
172
    CURL* _curl = nullptr;
173
    using HttpCallback = std::function<bool(const void* data, size_t length)>;
174
    const HttpCallback* _callback = nullptr;
175
    char _error_buf[CURL_ERROR_SIZE];
176
    curl_slist* _header_list = nullptr;
177
    HttpMethod _method = GET;
178
};
179
180
} // namespace doris