src/SerializeHandler.h
author František Kučera <franta-hg@frantovo.cz>
Sun, 17 Apr 2022 01:31:28 +0200
branchv_0
changeset 1 c0a1a008982c
parent 0 47ab821dd9ca
child 2 d2ba14aa4e20
permissions -rw-r--r--
first version of serializer
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
0
47ab821dd9ca establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
     1
/**
47ab821dd9ca establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
     2
 * Relational pipes
47ab821dd9ca establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
     3
 * Copyright © 2022 František Kučera (Frantovo.cz, GlobalCode.info)
47ab821dd9ca establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
     4
 *
47ab821dd9ca establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
     5
 * This program is free software: you can redistribute it and/or modify
47ab821dd9ca establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
     6
 * it under the terms of the GNU General Public License as published by
47ab821dd9ca establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
     7
 * the Free Software Foundation, version 3 of the License.
47ab821dd9ca establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
     8
 *
47ab821dd9ca establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
     9
 * This program is distributed in the hope that it will be useful,
47ab821dd9ca establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    10
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
47ab821dd9ca establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    11
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
47ab821dd9ca establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    12
 * GNU General Public License for more details.
47ab821dd9ca establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    13
 *
47ab821dd9ca establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    14
 * You should have received a copy of the GNU General Public License
47ab821dd9ca establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    15
 * along with this program. If not, see <http://www.gnu.org/licenses/>.
47ab821dd9ca establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    16
 */
47ab821dd9ca establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    17
#pragma once
47ab821dd9ca establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    18
47ab821dd9ca establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    19
#include <regex>
47ab821dd9ca establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    20
#include <stdexcept>
1
c0a1a008982c first version of serializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    21
#include <sstream>
c0a1a008982c first version of serializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    22
#include <codecvt>
c0a1a008982c first version of serializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    23
#include <iomanip>
0
47ab821dd9ca establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    24
47ab821dd9ca establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    25
#include <relpipe/common/type/typedefs.h>
47ab821dd9ca establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    26
#include <relpipe/reader/TypeId.h>
47ab821dd9ca establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    27
#include <relpipe/reader/handlers/RelationalReaderStringHandler.h>
47ab821dd9ca establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    28
#include <relpipe/reader/handlers/AttributeMetadata.h>
47ab821dd9ca establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    29
47ab821dd9ca establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    30
#include <relpipe/writer/Factory.h>
47ab821dd9ca establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    31
47ab821dd9ca establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    32
#include <relpipe/cli/RelpipeCLIException.h>
47ab821dd9ca establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    33
47ab821dd9ca establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    34
#include "Configuration.h"
47ab821dd9ca establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    35
47ab821dd9ca establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    36
namespace relpipe {
47ab821dd9ca establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    37
namespace tr {
47ab821dd9ca establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    38
namespace serialize {
47ab821dd9ca establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    39
47ab821dd9ca establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    40
class SerializeHandler : public relpipe::reader::handlers::RelationalReaderStringHandler {
47ab821dd9ca establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    41
private:
47ab821dd9ca establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    42
	Configuration configuration;
1
c0a1a008982c first version of serializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    43
	shared_ptr<relpipe::writer::RelationalWriter> writer;
c0a1a008982c first version of serializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    44
	std::wstring_convert<codecvt_utf8<wchar_t>> convertor;
c0a1a008982c first version of serializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    45
c0a1a008982c first version of serializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    46
	class RelationContext {
c0a1a008982c first version of serializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    47
	public:
c0a1a008982c first version of serializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    48
		relpipe::common::type::StringX name;
c0a1a008982c first version of serializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    49
		std::vector<relpipe::reader::handlers::AttributeMetadata> readerMetadata;
c0a1a008982c first version of serializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    50
		std::vector<relpipe::writer::AttributeMetadata> writerMetadata;
c0a1a008982c first version of serializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    51
	} relationContext;
c0a1a008982c first version of serializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    52
c0a1a008982c first version of serializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    53
	class RecordContext {
c0a1a008982c first version of serializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    54
	public:
c0a1a008982c first version of serializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    55
		std::stringstream buffer;
c0a1a008982c first version of serializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    56
		shared_ptr<relpipe::writer::RelationalWriter> writer;
c0a1a008982c first version of serializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    57
		size_t attributeIndex = 0;
c0a1a008982c first version of serializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    58
	} recordContext;
c0a1a008982c first version of serializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    59
c0a1a008982c first version of serializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    60
	relpipe::common::type::StringX toHex(const std::string& octets) {
c0a1a008982c first version of serializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    61
		std::stringstream hex;
c0a1a008982c first version of serializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    62
		hex << std::hex << std::setfill('0') << std::hex;
c0a1a008982c first version of serializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    63
		for (size_t i = 0, size = octets.size(); i < size; i++) hex << std::setw(2) << (0xff & octets[i]);
c0a1a008982c first version of serializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    64
		return convertor.from_bytes(hex.str());
c0a1a008982c first version of serializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    65
	}
c0a1a008982c first version of serializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    66
0
47ab821dd9ca establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    67
public:
47ab821dd9ca establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    68
1
c0a1a008982c first version of serializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    69
	SerializeHandler(shared_ptr<relpipe::writer::RelationalWriter> writer, Configuration configuration) : writer(writer), configuration(configuration) {
c0a1a008982c first version of serializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    70
		// TODO: configurable relation name
c0a1a008982c first version of serializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    71
		// TODO: configurable attribute name
c0a1a008982c first version of serializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    72
		// TODO: optional custom attributes with constant value or ordinal number
c0a1a008982c first version of serializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    73
		// TODO: optional serialization of only certain relations? and certain fields?
c0a1a008982c first version of serializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    74
		// TODO: optional pass-through of certain relations?
c0a1a008982c first version of serializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    75
		// TODO: multiple modes? one output record per a) one input record, b) several input records, c) one input relation, … ?
c0a1a008982c first version of serializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    76
c0a1a008982c first version of serializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    77
		writer->startRelation(L"message",{
c0a1a008982c first version of serializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    78
			{L"data", relpipe::writer::TypeId::STRING} // TODO: octet-string (when supported) instead of HEX
c0a1a008982c first version of serializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    79
		}, true);
0
47ab821dd9ca establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    80
	}
47ab821dd9ca establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    81
47ab821dd9ca establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    82
	virtual ~SerializeHandler() = default;
47ab821dd9ca establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    83
47ab821dd9ca establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    84
	void startRelation(relpipe::common::type::StringX name, std::vector<relpipe::reader::handlers::AttributeMetadata> attributes) override {
1
c0a1a008982c first version of serializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    85
		relationContext = RelationContext();
c0a1a008982c first version of serializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    86
c0a1a008982c first version of serializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    87
		relationContext.name = name;
c0a1a008982c first version of serializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    88
		relationContext.readerMetadata = attributes;
c0a1a008982c first version of serializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    89
c0a1a008982c first version of serializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    90
		recordContext = RecordContext();
c0a1a008982c first version of serializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    91
		recordContext.writer.reset(relpipe::writer::Factory::create(recordContext.buffer));
c0a1a008982c first version of serializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    92
c0a1a008982c first version of serializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    93
		for (relpipe::reader::handlers::AttributeMetadata readerMetadata : attributes) {
c0a1a008982c first version of serializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    94
			relationContext.writerMetadata.push_back({readerMetadata.getAttributeName(), writer->toTypeId(readerMetadata.getTypeName())});
c0a1a008982c first version of serializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    95
		}
0
47ab821dd9ca establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    96
	}
47ab821dd9ca establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    97
47ab821dd9ca establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    98
	void attribute(const relpipe::common::type::StringX& value) override {
1
c0a1a008982c first version of serializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    99
		if (recordContext.attributeIndex == 0) {
c0a1a008982c first version of serializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
   100
			recordContext.writer->startRelation(relationContext.name, relationContext.writerMetadata, true);
c0a1a008982c first version of serializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
   101
		}
0
47ab821dd9ca establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
   102
1
c0a1a008982c first version of serializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
   103
		recordContext.writer->writeAttribute(value); // TODO: typed values instead of strings
c0a1a008982c first version of serializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
   104
		recordContext.attributeIndex++;
c0a1a008982c first version of serializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
   105
c0a1a008982c first version of serializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
   106
		if (recordContext.attributeIndex % relationContext.readerMetadata.size() == 0) {
c0a1a008982c first version of serializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
   107
			writer->writeAttribute(toHex(recordContext.buffer.str()));
c0a1a008982c first version of serializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
   108
			recordContext = RecordContext();
c0a1a008982c first version of serializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
   109
			recordContext.writer.reset(relpipe::writer::Factory::create(recordContext.buffer));
c0a1a008982c first version of serializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
   110
		}
0
47ab821dd9ca establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
   111
	}
47ab821dd9ca establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
   112
47ab821dd9ca establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
   113
	void endOfPipe() {
1
c0a1a008982c first version of serializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
   114
		// TODO: check number of attribute values
0
47ab821dd9ca establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
   115
	}
47ab821dd9ca establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
   116
47ab821dd9ca establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
   117
};
47ab821dd9ca establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
   118
47ab821dd9ca establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
   119
}
47ab821dd9ca establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
   120
}
47ab821dd9ca establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
   121
}