src/HTTPHandler.h
author František Kučera <franta-hg@frantovo.cz>
Sun, 13 Mar 2022 01:17:59 +0100
branchv_0
changeset 6 59c9ca066322
parent 5 165f6162524d
child 7 0b0374746e48
permissions -rw-r--r--
write response headers as a relation
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
0
ce520a238309 new project relpipe-tr-http
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
     1
/**
ce520a238309 new project relpipe-tr-http
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
     2
 * Relational pipes
5
165f6162524d introduce HTTPClient wrapper around CURL
František Kučera <franta-hg@frantovo.cz>
parents: 4
diff changeset
     3
 * Copyright © 2022 František Kučera (Frantovo.cz, GlobalCode.info)
0
ce520a238309 new project relpipe-tr-http
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
     4
 *
ce520a238309 new project relpipe-tr-http
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
     5
 * This program is free software: you can redistribute it and/or modify
ce520a238309 new project relpipe-tr-http
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
     6
 * it under the terms of the GNU General Public License as published by
ce520a238309 new project relpipe-tr-http
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
     7
 * the Free Software Foundation, version 3 of the License.
ce520a238309 new project relpipe-tr-http
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
     8
 *
ce520a238309 new project relpipe-tr-http
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
     9
 * This program is distributed in the hope that it will be useful,
ce520a238309 new project relpipe-tr-http
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    10
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
ce520a238309 new project relpipe-tr-http
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    11
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
ce520a238309 new project relpipe-tr-http
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    12
 * GNU General Public License for more details.
ce520a238309 new project relpipe-tr-http
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    13
 *
ce520a238309 new project relpipe-tr-http
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    14
 * You should have received a copy of the GNU General Public License
ce520a238309 new project relpipe-tr-http
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    15
 * along with this program. If not, see <http://www.gnu.org/licenses/>.
ce520a238309 new project relpipe-tr-http
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    16
 */
ce520a238309 new project relpipe-tr-http
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    17
#pragma once
ce520a238309 new project relpipe-tr-http
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    18
ce520a238309 new project relpipe-tr-http
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    19
#include <memory>
ce520a238309 new project relpipe-tr-http
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    20
#include <string>
ce520a238309 new project relpipe-tr-http
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    21
#include <vector>
ce520a238309 new project relpipe-tr-http
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    22
#include <codecvt>
ce520a238309 new project relpipe-tr-http
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    23
#include <regex>
ce520a238309 new project relpipe-tr-http
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    24
#include <stdexcept>
ce520a238309 new project relpipe-tr-http
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    25
1
ddc12e789d00 link to the CURL library
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    26
#include <curl/curl.h>
ddc12e789d00 link to the CURL library
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    27
0
ce520a238309 new project relpipe-tr-http
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    28
#include <relpipe/common/type/typedefs.h>
ce520a238309 new project relpipe-tr-http
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    29
#include <relpipe/reader/TypeId.h>
ce520a238309 new project relpipe-tr-http
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    30
#include <relpipe/reader/handlers/RelationalReaderStringHandler.h>
ce520a238309 new project relpipe-tr-http
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    31
#include <relpipe/reader/handlers/AttributeMetadata.h>
ce520a238309 new project relpipe-tr-http
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    32
ce520a238309 new project relpipe-tr-http
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    33
#include <relpipe/writer/Factory.h>
1
ddc12e789d00 link to the CURL library
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    34
#include <relpipe/writer/TypeId.h>
0
ce520a238309 new project relpipe-tr-http
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    35
ce520a238309 new project relpipe-tr-http
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    36
#include <relpipe/cli/RelpipeCLIException.h>
ce520a238309 new project relpipe-tr-http
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    37
ce520a238309 new project relpipe-tr-http
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    38
#include "Configuration.h"
5
165f6162524d introduce HTTPClient wrapper around CURL
František Kučera <franta-hg@frantovo.cz>
parents: 4
diff changeset
    39
#include "HTTPClient.h"
0
ce520a238309 new project relpipe-tr-http
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    40
ce520a238309 new project relpipe-tr-http
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    41
namespace relpipe {
ce520a238309 new project relpipe-tr-http
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    42
namespace tr {
ce520a238309 new project relpipe-tr-http
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    43
namespace http {
ce520a238309 new project relpipe-tr-http
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    44
ce520a238309 new project relpipe-tr-http
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    45
class HTTPHandler : public relpipe::reader::handlers::RelationalReaderStringHandler {
ce520a238309 new project relpipe-tr-http
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    46
private:
ce520a238309 new project relpipe-tr-http
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    47
	std::wstring_convert<codecvt_utf8<wchar_t>> convertor; // TODO: support also other encodings.
ce520a238309 new project relpipe-tr-http
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    48
	shared_ptr<relpipe::writer::RelationalWriter> relationalWriter;
ce520a238309 new project relpipe-tr-http
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    49
	Configuration configuration;
ce520a238309 new project relpipe-tr-http
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    50
	std::vector<relpipe::reader::handlers::AttributeMetadata> currentReaderMetadata;
ce520a238309 new project relpipe-tr-http
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    51
	std::vector<relpipe::writer::AttributeMetadata> currentWriterMetadata;
6
59c9ca066322 write response headers as a relation
František Kučera <franta-hg@frantovo.cz>
parents: 5
diff changeset
    52
	std::vector<relpipe::common::type::StringX> responseHeaders;
0
ce520a238309 new project relpipe-tr-http
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    53
	size_t currentAttributeIndex = 0;
ce520a238309 new project relpipe-tr-http
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    54
	size_t currentRecordNumber = 1;
ce520a238309 new project relpipe-tr-http
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    55
6
59c9ca066322 write response headers as a relation
František Kučera <franta-hg@frantovo.cz>
parents: 5
diff changeset
    56
	void writeHeaders() {
59c9ca066322 write response headers as a relation
František Kučera <franta-hg@frantovo.cz>
parents: 5
diff changeset
    57
		if (responseHeaders.size()) {
59c9ca066322 write response headers as a relation
František Kučera <franta-hg@frantovo.cz>
parents: 5
diff changeset
    58
			relationalWriter->startRelation(L"response_header",{
59c9ca066322 write response headers as a relation
František Kučera <franta-hg@frantovo.cz>
parents: 5
diff changeset
    59
				// TODO: request ID instead of URL (or both)
59c9ca066322 write response headers as a relation
František Kučera <franta-hg@frantovo.cz>
parents: 5
diff changeset
    60
				{L"url", relpipe::writer::TypeId::STRING},
59c9ca066322 write response headers as a relation
František Kučera <franta-hg@frantovo.cz>
parents: 5
diff changeset
    61
				{L"name", relpipe::writer::TypeId::STRING},
59c9ca066322 write response headers as a relation
František Kučera <franta-hg@frantovo.cz>
parents: 5
diff changeset
    62
				{L"value", relpipe::writer::TypeId::STRING},
59c9ca066322 write response headers as a relation
František Kučera <franta-hg@frantovo.cz>
parents: 5
diff changeset
    63
			}, true);
59c9ca066322 write response headers as a relation
František Kučera <franta-hg@frantovo.cz>
parents: 5
diff changeset
    64
59c9ca066322 write response headers as a relation
František Kučera <franta-hg@frantovo.cz>
parents: 5
diff changeset
    65
			for (auto s : responseHeaders) relationalWriter->writeAttribute(s);
59c9ca066322 write response headers as a relation
František Kučera <franta-hg@frantovo.cz>
parents: 5
diff changeset
    66
59c9ca066322 write response headers as a relation
František Kučera <franta-hg@frantovo.cz>
parents: 5
diff changeset
    67
			responseHeaders.clear();
59c9ca066322 write response headers as a relation
František Kučera <franta-hg@frantovo.cz>
parents: 5
diff changeset
    68
		}
59c9ca066322 write response headers as a relation
František Kučera <franta-hg@frantovo.cz>
parents: 5
diff changeset
    69
	}
59c9ca066322 write response headers as a relation
František Kučera <franta-hg@frantovo.cz>
parents: 5
diff changeset
    70
0
ce520a238309 new project relpipe-tr-http
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    71
public:
ce520a238309 new project relpipe-tr-http
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    72
ce520a238309 new project relpipe-tr-http
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    73
	HTTPHandler(shared_ptr<relpipe::writer::RelationalWriter> relationalWriter, Configuration configuration) : relationalWriter(relationalWriter), configuration(configuration) {
ce520a238309 new project relpipe-tr-http
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    74
	}
ce520a238309 new project relpipe-tr-http
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    75
ce520a238309 new project relpipe-tr-http
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    76
	virtual ~HTTPHandler() {
ce520a238309 new project relpipe-tr-http
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    77
	}
ce520a238309 new project relpipe-tr-http
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    78
ce520a238309 new project relpipe-tr-http
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    79
	void startRelation(relpipe::common::type::StringX name, std::vector<relpipe::reader::handlers::AttributeMetadata> attributes) override {
ce520a238309 new project relpipe-tr-http
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    80
6
59c9ca066322 write response headers as a relation
František Kučera <franta-hg@frantovo.cz>
parents: 5
diff changeset
    81
		writeHeaders(); // from previous relation
1
ddc12e789d00 link to the CURL library
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    82
3
9c710397ced6 return also the HTTP response code and URL
František Kučera <franta-hg@frantovo.cz>
parents: 2
diff changeset
    83
		relationalWriter->startRelation(name + L"_curl_info",{
9c710397ced6 return also the HTTP response code and URL
František Kučera <franta-hg@frantovo.cz>
parents: 2
diff changeset
    84
			{L"name", relpipe::writer::TypeId::STRING},
2
dccedac46e7e do HTTP request for each attribute value
František Kučera <franta-hg@frantovo.cz>
parents: 1
diff changeset
    85
			{L"value", relpipe::writer::TypeId::STRING}
1
ddc12e789d00 link to the CURL library
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    86
		}, true);
ddc12e789d00 link to the CURL library
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    87
3
9c710397ced6 return also the HTTP response code and URL
František Kučera <franta-hg@frantovo.cz>
parents: 2
diff changeset
    88
		relationalWriter->writeAttribute(L"curl version");
9c710397ced6 return also the HTTP response code and URL
František Kučera <franta-hg@frantovo.cz>
parents: 2
diff changeset
    89
		relationalWriter->writeAttribute(convertor.from_bytes(curl_version()));
9c710397ced6 return also the HTTP response code and URL
František Kučera <franta-hg@frantovo.cz>
parents: 2
diff changeset
    90
9c710397ced6 return also the HTTP response code and URL
František Kučera <franta-hg@frantovo.cz>
parents: 2
diff changeset
    91
		relationalWriter->startRelation(name + L"_curl_response",{
6
59c9ca066322 write response headers as a relation
František Kučera <franta-hg@frantovo.cz>
parents: 5
diff changeset
    92
			// TODO: request ID
59c9ca066322 write response headers as a relation
František Kučera <franta-hg@frantovo.cz>
parents: 5
diff changeset
    93
			// TODO: body in hexadecimal/binary format
3
9c710397ced6 return also the HTTP response code and URL
František Kučera <franta-hg@frantovo.cz>
parents: 2
diff changeset
    94
			{L"url", relpipe::writer::TypeId::STRING},
9c710397ced6 return also the HTTP response code and URL
František Kučera <franta-hg@frantovo.cz>
parents: 2
diff changeset
    95
			{L"body", relpipe::writer::TypeId::STRING},
5
165f6162524d introduce HTTPClient wrapper around CURL
František Kučera <franta-hg@frantovo.cz>
parents: 4
diff changeset
    96
			{L"response_code", relpipe::writer::TypeId::INTEGER},
3
9c710397ced6 return also the HTTP response code and URL
František Kučera <franta-hg@frantovo.cz>
parents: 2
diff changeset
    97
		}, true);
1
ddc12e789d00 link to the CURL library
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    98
0
ce520a238309 new project relpipe-tr-http
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    99
	}
ce520a238309 new project relpipe-tr-http
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
   100
ce520a238309 new project relpipe-tr-http
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
   101
	void attribute(const relpipe::common::type::StringX& value) override {
ce520a238309 new project relpipe-tr-http
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
   102
5
165f6162524d introduce HTTPClient wrapper around CURL
František Kučera <franta-hg@frantovo.cz>
parents: 4
diff changeset
   103
		std::shared_ptr<HTTPClient> http(HTTPClient::open());
2
dccedac46e7e do HTTP request for each attribute value
František Kučera <franta-hg@frantovo.cz>
parents: 1
diff changeset
   104
5
165f6162524d introduce HTTPClient wrapper around CURL
František Kučera <franta-hg@frantovo.cz>
parents: 4
diff changeset
   105
		HTTPClient::Request request;
165f6162524d introduce HTTPClient wrapper around CURL
František Kučera <franta-hg@frantovo.cz>
parents: 4
diff changeset
   106
		request.method = HTTPClient::Method::GET;
165f6162524d introduce HTTPClient wrapper around CURL
František Kučera <franta-hg@frantovo.cz>
parents: 4
diff changeset
   107
		request.url = convertor.to_bytes(value);
4
602462d04c57 print somehow HTTP headers on STDERR
František Kučera <franta-hg@frantovo.cz>
parents: 3
diff changeset
   108
5
165f6162524d introduce HTTPClient wrapper around CURL
František Kučera <franta-hg@frantovo.cz>
parents: 4
diff changeset
   109
		HTTPClient::Response response = http->exchange(request);
165f6162524d introduce HTTPClient wrapper around CURL
František Kučera <franta-hg@frantovo.cz>
parents: 4
diff changeset
   110
		relpipe::common::type::Integer responseCode = response.responseCode;
4
602462d04c57 print somehow HTTP headers on STDERR
František Kučera <franta-hg@frantovo.cz>
parents: 3
diff changeset
   111
5
165f6162524d introduce HTTPClient wrapper around CURL
František Kučera <franta-hg@frantovo.cz>
parents: 4
diff changeset
   112
		relationalWriter->writeAttribute(value);
165f6162524d introduce HTTPClient wrapper around CURL
František Kučera <franta-hg@frantovo.cz>
parents: 4
diff changeset
   113
		relationalWriter->writeAttribute(convertor.from_bytes(response.body));
3
9c710397ced6 return also the HTTP response code and URL
František Kučera <franta-hg@frantovo.cz>
parents: 2
diff changeset
   114
		relationalWriter->writeAttribute(&responseCode, typeid (responseCode));
6
59c9ca066322 write response headers as a relation
František Kučera <franta-hg@frantovo.cz>
parents: 5
diff changeset
   115
59c9ca066322 write response headers as a relation
František Kučera <franta-hg@frantovo.cz>
parents: 5
diff changeset
   116
		for (size_t i = 0; i < response.headers.size(); i += 2) {
59c9ca066322 write response headers as a relation
František Kučera <franta-hg@frantovo.cz>
parents: 5
diff changeset
   117
			responseHeaders.push_back(convertor.from_bytes(request.url));
59c9ca066322 write response headers as a relation
František Kučera <franta-hg@frantovo.cz>
parents: 5
diff changeset
   118
			responseHeaders.push_back(convertor.from_bytes(response.headers[i]));
59c9ca066322 write response headers as a relation
František Kučera <franta-hg@frantovo.cz>
parents: 5
diff changeset
   119
			responseHeaders.push_back(convertor.from_bytes(response.headers[i + 1]));
59c9ca066322 write response headers as a relation
František Kučera <franta-hg@frantovo.cz>
parents: 5
diff changeset
   120
		}
0
ce520a238309 new project relpipe-tr-http
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
   121
	}
ce520a238309 new project relpipe-tr-http
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
   122
ce520a238309 new project relpipe-tr-http
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
   123
	void endOfPipe() {
6
59c9ca066322 write response headers as a relation
František Kučera <franta-hg@frantovo.cz>
parents: 5
diff changeset
   124
		writeHeaders(); // from last relation
0
ce520a238309 new project relpipe-tr-http
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
   125
	}
ce520a238309 new project relpipe-tr-http
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
   126
ce520a238309 new project relpipe-tr-http
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
   127
};
ce520a238309 new project relpipe-tr-http
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
   128
ce520a238309 new project relpipe-tr-http
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
   129
}
ce520a238309 new project relpipe-tr-http
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
   130
}
ce520a238309 new project relpipe-tr-http
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
   131
}