src/DeserializeHandler.h
author František Kučera <franta-hg@frantovo.cz>
Tue, 19 Apr 2022 23:17:15 +0200
branchv_0
changeset 1 d53041bb781b
parent 0 7e986fcf0d8f
child 2 d586db80b6cd
permissions -rw-r--r--
first version of deserializer
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
0
7e986fcf0d8f establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
     1
/**
7e986fcf0d8f establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
     2
 * Relational pipes
7e986fcf0d8f establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
     3
 * Copyright © 2022 František Kučera (Frantovo.cz, GlobalCode.info)
7e986fcf0d8f establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
     4
 *
7e986fcf0d8f establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
     5
 * This program is free software: you can redistribute it and/or modify
7e986fcf0d8f 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
7e986fcf0d8f establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
     7
 * the Free Software Foundation, version 3 of the License.
7e986fcf0d8f establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
     8
 *
7e986fcf0d8f establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
     9
 * This program is distributed in the hope that it will be useful,
7e986fcf0d8f establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    10
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
7e986fcf0d8f establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    11
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
7e986fcf0d8f establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    12
 * GNU General Public License for more details.
7e986fcf0d8f establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    13
 *
7e986fcf0d8f establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    14
 * You should have received a copy of the GNU General Public License
7e986fcf0d8f establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    15
 * along with this program. If not, see <http://www.gnu.org/licenses/>.
7e986fcf0d8f establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    16
 */
7e986fcf0d8f establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    17
#pragma once
7e986fcf0d8f establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    18
7e986fcf0d8f establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    19
#include <regex>
7e986fcf0d8f establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    20
#include <stdexcept>
7e986fcf0d8f establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    21
#include <sstream>
7e986fcf0d8f establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    22
#include <codecvt>
7e986fcf0d8f establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    23
#include <iomanip>
7e986fcf0d8f establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    24
7e986fcf0d8f establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    25
#include <relpipe/common/type/typedefs.h>
7e986fcf0d8f establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    26
#include <relpipe/reader/TypeId.h>
7e986fcf0d8f establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    27
#include <relpipe/reader/handlers/RelationalReaderStringHandler.h>
7e986fcf0d8f establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    28
#include <relpipe/reader/handlers/AttributeMetadata.h>
7e986fcf0d8f establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    29
7e986fcf0d8f establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    30
#include <relpipe/writer/Factory.h>
7e986fcf0d8f establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    31
7e986fcf0d8f establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    32
#include <relpipe/cli/RelpipeCLIException.h>
7e986fcf0d8f establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    33
7e986fcf0d8f establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    34
#include "Configuration.h"
1
d53041bb781b first version of deserializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    35
#include "UnionAllHandler.h"
0
7e986fcf0d8f establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    36
7e986fcf0d8f establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    37
namespace relpipe {
7e986fcf0d8f establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    38
namespace tr {
7e986fcf0d8f establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    39
namespace deserialize {
7e986fcf0d8f establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    40
7e986fcf0d8f establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    41
class DeserializeHandler : public relpipe::reader::handlers::RelationalReaderStringHandler {
7e986fcf0d8f establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    42
private:
7e986fcf0d8f establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    43
	Configuration configuration;
7e986fcf0d8f establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    44
	shared_ptr<relpipe::writer::RelationalWriter> writer;
1
d53041bb781b first version of deserializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    45
	std::wstring_convert<codecvt_utf8<wchar_t>> convertor;
d53041bb781b first version of deserializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    46
	UnionAllHandler unionAllHandler;
0
7e986fcf0d8f establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    47
1
d53041bb781b first version of deserializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    48
	class RelationContext {
d53041bb781b first version of deserializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    49
	public:
d53041bb781b first version of deserializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    50
		relpipe::common::type::StringX name;
d53041bb781b first version of deserializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    51
		std::vector<relpipe::reader::handlers::AttributeMetadata> readerMetadata;
d53041bb781b first version of deserializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    52
		std::vector<relpipe::writer::AttributeMetadata> writerMetadata;
d53041bb781b first version of deserializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    53
	} relationContext;
d53041bb781b first version of deserializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    54
d53041bb781b first version of deserializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    55
	class RecordContext {
d53041bb781b first version of deserializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    56
	public:
d53041bb781b first version of deserializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    57
		std::stringstream buffer;
d53041bb781b first version of deserializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    58
		size_t attributeIndex = 0;
d53041bb781b first version of deserializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    59
	} recordContext;
d53041bb781b first version of deserializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    60
d53041bb781b first version of deserializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    61
	char fromHex(wchar_t ch) {
d53041bb781b first version of deserializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    62
		if (L'0' <= ch && ch <= L'9') return ch - L'0';
d53041bb781b first version of deserializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    63
		else if (L'a' <= ch && ch <= L'f') return ch - L'a' + 10;
d53041bb781b first version of deserializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    64
		else throw relpipe::reader::RelpipeReaderException(L"Unable to decode hexadeximal string.");
d53041bb781b first version of deserializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    65
	}
d53041bb781b first version of deserializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    66
d53041bb781b first version of deserializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    67
	std::stringstream fromHex(const relpipe::common::type::StringX& hex) {
d53041bb781b first version of deserializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    68
		std::stringstream octets;
d53041bb781b first version of deserializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    69
d53041bb781b first version of deserializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    70
		char octet = 0;
d53041bb781b first version of deserializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    71
d53041bb781b first version of deserializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    72
		for (size_t i = 0, limit = hex.size(); i < limit; i++) {
d53041bb781b first version of deserializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    73
			if (i % 2 == 0) {
d53041bb781b first version of deserializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    74
				octet = fromHex(hex[i]) << 4;
d53041bb781b first version of deserializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    75
			} else {
d53041bb781b first version of deserializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    76
				octet += fromHex(hex[i]);
d53041bb781b first version of deserializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    77
				octets.put(octet);
d53041bb781b first version of deserializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    78
			}
d53041bb781b first version of deserializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    79
		}
d53041bb781b first version of deserializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    80
d53041bb781b first version of deserializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    81
		return octets;
d53041bb781b first version of deserializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    82
	}
d53041bb781b first version of deserializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    83
	
0
7e986fcf0d8f establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    84
public:
7e986fcf0d8f establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    85
1
d53041bb781b first version of deserializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    86
	DeserializeHandler(shared_ptr<relpipe::writer::RelationalWriter> writer, Configuration configuration) : writer(writer), configuration(configuration), unionAllHandler(writer, configuration) {
d53041bb781b first version of deserializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    87
		// TODO: configurable relation name?
d53041bb781b first version of deserializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    88
		// TODO: configurable attribute name?
d53041bb781b first version of deserializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    89
		// TODO: optional custom attributes with constant value or ordinal number?
d53041bb781b first version of deserializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    90
		// TODO: optional deserialization of only certain relations? and certain fields?
d53041bb781b first version of deserializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    91
		// TODO: optional pass-through of certain relations?
0
7e986fcf0d8f establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    92
	}
7e986fcf0d8f establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    93
7e986fcf0d8f establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    94
	virtual ~DeserializeHandler() = default;
7e986fcf0d8f establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    95
7e986fcf0d8f establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    96
	void startRelation(relpipe::common::type::StringX name, std::vector<relpipe::reader::handlers::AttributeMetadata> attributes) override {
1
d53041bb781b first version of deserializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    97
		relationContext = RelationContext();
d53041bb781b first version of deserializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    98
d53041bb781b first version of deserializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
    99
		relationContext.name = name;
d53041bb781b first version of deserializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
   100
		relationContext.readerMetadata = attributes;
d53041bb781b first version of deserializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
   101
d53041bb781b first version of deserializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
   102
		for (relpipe::reader::handlers::AttributeMetadata readerMetadata : attributes) {
d53041bb781b first version of deserializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
   103
			relationContext.writerMetadata.push_back({readerMetadata.getAttributeName(), writer->toTypeId(readerMetadata.getTypeName())});
d53041bb781b first version of deserializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
   104
		}
0
7e986fcf0d8f establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
   105
	}
7e986fcf0d8f establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
   106
7e986fcf0d8f establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
   107
	void attribute(const relpipe::common::type::StringX& value) override {
1
d53041bb781b first version of deserializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
   108
		auto attributeName = relationContext.readerMetadata[recordContext.attributeIndex].getAttributeName();
d53041bb781b first version of deserializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
   109
d53041bb781b first version of deserializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
   110
		if (recordContext.attributeIndex == 0) recordContext = RecordContext();
d53041bb781b first version of deserializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
   111
d53041bb781b first version of deserializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
   112
		if (attributeName == L"data") recordContext.buffer = fromHex(value);
d53041bb781b first version of deserializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
   113
d53041bb781b first version of deserializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
   114
		recordContext.attributeIndex++;
d53041bb781b first version of deserializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
   115
d53041bb781b first version of deserializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
   116
		if (recordContext.attributeIndex % relationContext.readerMetadata.size() == 0) {
d53041bb781b first version of deserializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
   117
			std::stringstream octets = fromHex(value); // TODO: skip this step once we have octet-string data type + typed values instead of strings
d53041bb781b first version of deserializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
   118
			std::shared_ptr<relpipe::reader::RelationalReader> reader(relpipe::reader::Factory::create(octets));
d53041bb781b first version of deserializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
   119
			reader->addHandler(&unionAllHandler);
d53041bb781b first version of deserializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
   120
			reader->process();
d53041bb781b first version of deserializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
   121
			recordContext.attributeIndex = 0;
d53041bb781b first version of deserializer
František Kučera <franta-hg@frantovo.cz>
parents: 0
diff changeset
   122
		}
0
7e986fcf0d8f establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
   123
	}
7e986fcf0d8f establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
   124
7e986fcf0d8f establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
   125
	void endOfPipe() {
7e986fcf0d8f establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
   126
	}
7e986fcf0d8f establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
   127
7e986fcf0d8f establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
   128
};
7e986fcf0d8f establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
   129
7e986fcf0d8f establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
   130
}
7e986fcf0d8f establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
   131
}
7e986fcf0d8f establish project
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
   132
}