src/GrepHandler.h
author František Kučera <franta-hg@frantovo.cz>
Tue, 11 May 2021 18:30:50 +0200
branchv_0
changeset 24 c69670b7b4ef
parent 18 5a3602fb1649
child 25 98be80d2e65b
permissions -rw-r--r--
new CLI interface: --relation, --attribute, --pattern
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
3
8731263d44f1 PassthroughHandler (string-based version)
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
     1
/**
8731263d44f1 PassthroughHandler (string-based version)
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
     2
 * Relational pipes
8731263d44f1 PassthroughHandler (string-based version)
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
     3
 * Copyright © 2018 František Kučera (Frantovo.cz, GlobalCode.info)
8731263d44f1 PassthroughHandler (string-based version)
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
     4
 *
8731263d44f1 PassthroughHandler (string-based version)
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
     5
 * This program is free software: you can redistribute it and/or modify
8731263d44f1 PassthroughHandler (string-based version)
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
     6
 * it under the terms of the GNU General Public License as published by
18
5a3602fb1649 fix license version: GNU GPLv3
František Kučera <franta-hg@frantovo.cz>
parents: 14
diff changeset
     7
 * the Free Software Foundation, version 3 of the License.
3
8731263d44f1 PassthroughHandler (string-based version)
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
     8
 *
8731263d44f1 PassthroughHandler (string-based version)
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
     9
 * This program is distributed in the hope that it will be useful,
8731263d44f1 PassthroughHandler (string-based version)
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    10
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
8731263d44f1 PassthroughHandler (string-based version)
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    11
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
8731263d44f1 PassthroughHandler (string-based version)
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    12
 * GNU General Public License for more details.
8731263d44f1 PassthroughHandler (string-based version)
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    13
 *
8731263d44f1 PassthroughHandler (string-based version)
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    14
 * You should have received a copy of the GNU General Public License
8731263d44f1 PassthroughHandler (string-based version)
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    15
 * along with this program. If not, see <http://www.gnu.org/licenses/>.
8731263d44f1 PassthroughHandler (string-based version)
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    16
 */
8731263d44f1 PassthroughHandler (string-based version)
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    17
#pragma once
8731263d44f1 PassthroughHandler (string-based version)
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    18
8731263d44f1 PassthroughHandler (string-based version)
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    19
#include <memory>
8731263d44f1 PassthroughHandler (string-based version)
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    20
#include <string>
8731263d44f1 PassthroughHandler (string-based version)
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    21
#include <vector>
8731263d44f1 PassthroughHandler (string-based version)
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    22
#include <iostream>
8731263d44f1 PassthroughHandler (string-based version)
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    23
#include <sstream>
8731263d44f1 PassthroughHandler (string-based version)
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    24
#include <locale>
8731263d44f1 PassthroughHandler (string-based version)
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    25
#include <codecvt>
8731263d44f1 PassthroughHandler (string-based version)
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    26
#include <regex>
8731263d44f1 PassthroughHandler (string-based version)
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    27
8731263d44f1 PassthroughHandler (string-based version)
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    28
#include <relpipe/reader/typedefs.h>
8731263d44f1 PassthroughHandler (string-based version)
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    29
#include <relpipe/reader/TypeId.h>
8731263d44f1 PassthroughHandler (string-based version)
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    30
#include <relpipe/reader/handlers/RelationalReaderStringHandler.h>
8731263d44f1 PassthroughHandler (string-based version)
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    31
#include <relpipe/reader/handlers/AttributeMetadata.h>
8731263d44f1 PassthroughHandler (string-based version)
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    32
8731263d44f1 PassthroughHandler (string-based version)
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    33
#include <relpipe/writer/Factory.h>
8731263d44f1 PassthroughHandler (string-based version)
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    34
8
f66c759d1111 first working grep version
František Kučera <franta-hg@frantovo.cz>
parents: 7
diff changeset
    35
#include <relpipe/cli/RelpipeCLIException.h>
f66c759d1111 first working grep version
František Kučera <franta-hg@frantovo.cz>
parents: 7
diff changeset
    36
24
c69670b7b4ef new CLI interface: --relation, --attribute, --pattern
František Kučera <franta-hg@frantovo.cz>
parents: 18
diff changeset
    37
#include "Configuration.h"
c69670b7b4ef new CLI interface: --relation, --attribute, --pattern
František Kučera <franta-hg@frantovo.cz>
parents: 18
diff changeset
    38
3
8731263d44f1 PassthroughHandler (string-based version)
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    39
namespace relpipe {
8731263d44f1 PassthroughHandler (string-based version)
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    40
namespace tr {
7
aebaf590a838 use smart pointer
František Kučera <franta-hg@frantovo.cz>
parents: 3
diff changeset
    41
namespace grep {
3
8731263d44f1 PassthroughHandler (string-based version)
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    42
8
f66c759d1111 first working grep version
František Kučera <franta-hg@frantovo.cz>
parents: 7
diff changeset
    43
using namespace std;
3
8731263d44f1 PassthroughHandler (string-based version)
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    44
using namespace relpipe;
8731263d44f1 PassthroughHandler (string-based version)
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    45
using namespace relpipe::reader;
8731263d44f1 PassthroughHandler (string-based version)
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    46
using namespace relpipe::reader::handlers;
8731263d44f1 PassthroughHandler (string-based version)
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    47
14
68c084fb822f fix typo: Hadler → Handler
František Kučera <franta-hg@frantovo.cz>
parents: 8
diff changeset
    48
class GrepHandler : public RelationalReaderStringHandler {
3
8731263d44f1 PassthroughHandler (string-based version)
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    49
private:
8
f66c759d1111 first working grep version
František Kučera <franta-hg@frantovo.cz>
parents: 7
diff changeset
    50
	shared_ptr<writer::RelationalWriter> relationalWriter;
24
c69670b7b4ef new CLI interface: --relation, --attribute, --pattern
František Kučera <franta-hg@frantovo.cz>
parents: 18
diff changeset
    51
	Configuration configuration;
c69670b7b4ef new CLI interface: --relation, --attribute, --pattern
František Kučera <franta-hg@frantovo.cz>
parents: 18
diff changeset
    52
	RelationConfiguration* currentFilter = nullptr;
8
f66c759d1111 first working grep version
František Kučera <franta-hg@frantovo.cz>
parents: 7
diff changeset
    53
f66c759d1111 first working grep version
František Kučera <franta-hg@frantovo.cz>
parents: 7
diff changeset
    54
	vector<boolean_t> currentSearchableAttributes;
f66c759d1111 first working grep version
František Kučera <franta-hg@frantovo.cz>
parents: 7
diff changeset
    55
	vector<string_t> currentRecord;
f66c759d1111 first working grep version
František Kučera <franta-hg@frantovo.cz>
parents: 7
diff changeset
    56
	integer_t currentAttributeIndex = 0;
f66c759d1111 first working grep version
František Kučera <franta-hg@frantovo.cz>
parents: 7
diff changeset
    57
	boolean_t includeCurrentRecord = false;
f66c759d1111 first working grep version
František Kučera <franta-hg@frantovo.cz>
parents: 7
diff changeset
    58
3
8731263d44f1 PassthroughHandler (string-based version)
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    59
public:
8731263d44f1 PassthroughHandler (string-based version)
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    60
24
c69670b7b4ef new CLI interface: --relation, --attribute, --pattern
František Kučera <franta-hg@frantovo.cz>
parents: 18
diff changeset
    61
	GrepHandler(shared_ptr<writer::RelationalWriter> relationalWriter, Configuration configuration) : relationalWriter(relationalWriter), configuration(configuration) {
3
8731263d44f1 PassthroughHandler (string-based version)
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    62
	}
8731263d44f1 PassthroughHandler (string-based version)
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    63
8
f66c759d1111 first working grep version
František Kučera <franta-hg@frantovo.cz>
parents: 7
diff changeset
    64
	void startRelation(string_t name, vector<AttributeMetadata> attributes) override {
3
8731263d44f1 PassthroughHandler (string-based version)
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    65
		// TODO: move to a reusable method (or use same metadata on both reader and writer side?)
8
f66c759d1111 first working grep version
František Kučera <franta-hg@frantovo.cz>
parents: 7
diff changeset
    66
		vector<writer::AttributeMetadata> writerMetadata;
3
8731263d44f1 PassthroughHandler (string-based version)
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    67
		for (AttributeMetadata readerMetadata : attributes) {
8731263d44f1 PassthroughHandler (string-based version)
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    68
			writerMetadata.push_back({readerMetadata.getAttributeName(), relationalWriter->toTypeId(readerMetadata.getTypeName())});
8731263d44f1 PassthroughHandler (string-based version)
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    69
		}
8
f66c759d1111 first working grep version
František Kučera <franta-hg@frantovo.cz>
parents: 7
diff changeset
    70
f66c759d1111 first working grep version
František Kučera <franta-hg@frantovo.cz>
parents: 7
diff changeset
    71
f66c759d1111 first working grep version
František Kučera <franta-hg@frantovo.cz>
parents: 7
diff changeset
    72
		currentRecord.resize(attributes.size());
f66c759d1111 first working grep version
František Kučera <franta-hg@frantovo.cz>
parents: 7
diff changeset
    73
		currentSearchableAttributes.resize(attributes.size(), false);
24
c69670b7b4ef new CLI interface: --relation, --attribute, --pattern
František Kučera <franta-hg@frantovo.cz>
parents: 18
diff changeset
    74
		currentFilter = nullptr;
c69670b7b4ef new CLI interface: --relation, --attribute, --pattern
František Kučera <franta-hg@frantovo.cz>
parents: 18
diff changeset
    75
		for (int i = 0; i < configuration.relationConfigurations.size(); i++) {
c69670b7b4ef new CLI interface: --relation, --attribute, --pattern
František Kučera <franta-hg@frantovo.cz>
parents: 18
diff changeset
    76
			if (regex_match(name, configuration.relationConfigurations[i].relationPattern)) {
c69670b7b4ef new CLI interface: --relation, --attribute, --pattern
František Kučera <franta-hg@frantovo.cz>
parents: 18
diff changeset
    77
				currentFilter = &configuration.relationConfigurations[i];
c69670b7b4ef new CLI interface: --relation, --attribute, --pattern
František Kučera <franta-hg@frantovo.cz>
parents: 18
diff changeset
    78
				break;
c69670b7b4ef new CLI interface: --relation, --attribute, --pattern
František Kučera <franta-hg@frantovo.cz>
parents: 18
diff changeset
    79
			}
c69670b7b4ef new CLI interface: --relation, --attribute, --pattern
František Kučera <franta-hg@frantovo.cz>
parents: 18
diff changeset
    80
		}
c69670b7b4ef new CLI interface: --relation, --attribute, --pattern
František Kučera <franta-hg@frantovo.cz>
parents: 18
diff changeset
    81
c69670b7b4ef new CLI interface: --relation, --attribute, --pattern
František Kučera <franta-hg@frantovo.cz>
parents: 18
diff changeset
    82
		if (currentFilter) {
8
f66c759d1111 first working grep version
František Kučera <franta-hg@frantovo.cz>
parents: 7
diff changeset
    83
			for (int i = 0; i < currentSearchableAttributes.size(); i++) {
24
c69670b7b4ef new CLI interface: --relation, --attribute, --pattern
František Kučera <franta-hg@frantovo.cz>
parents: 18
diff changeset
    84
				currentSearchableAttributes[i] = regex_match(attributes[i].getAttributeName(), currentFilter->attributePattern);
8
f66c759d1111 first working grep version
František Kučera <franta-hg@frantovo.cz>
parents: 7
diff changeset
    85
			}
f66c759d1111 first working grep version
František Kučera <franta-hg@frantovo.cz>
parents: 7
diff changeset
    86
		}
f66c759d1111 first working grep version
František Kučera <franta-hg@frantovo.cz>
parents: 7
diff changeset
    87
3
8731263d44f1 PassthroughHandler (string-based version)
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    88
		relationalWriter->startRelation(name, writerMetadata, true);
8731263d44f1 PassthroughHandler (string-based version)
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    89
	}
8731263d44f1 PassthroughHandler (string-based version)
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    90
8731263d44f1 PassthroughHandler (string-based version)
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
    91
	void attribute(const string_t& value) override {
24
c69670b7b4ef new CLI interface: --relation, --attribute, --pattern
František Kučera <franta-hg@frantovo.cz>
parents: 18
diff changeset
    92
		if (currentFilter) {
8
f66c759d1111 first working grep version
František Kučera <franta-hg@frantovo.cz>
parents: 7
diff changeset
    93
			currentRecord[currentAttributeIndex] = value;
f66c759d1111 first working grep version
František Kučera <franta-hg@frantovo.cz>
parents: 7
diff changeset
    94
f66c759d1111 first working grep version
František Kučera <franta-hg@frantovo.cz>
parents: 7
diff changeset
    95
			if (currentSearchableAttributes[currentAttributeIndex]) {
24
c69670b7b4ef new CLI interface: --relation, --attribute, --pattern
František Kučera <franta-hg@frantovo.cz>
parents: 18
diff changeset
    96
				includeCurrentRecord |= regex_search(value, currentFilter->valuePattern);
8
f66c759d1111 first working grep version
František Kučera <franta-hg@frantovo.cz>
parents: 7
diff changeset
    97
			}
f66c759d1111 first working grep version
František Kučera <franta-hg@frantovo.cz>
parents: 7
diff changeset
    98
f66c759d1111 first working grep version
František Kučera <franta-hg@frantovo.cz>
parents: 7
diff changeset
    99
			currentAttributeIndex++;
f66c759d1111 first working grep version
František Kučera <franta-hg@frantovo.cz>
parents: 7
diff changeset
   100
f66c759d1111 first working grep version
František Kučera <franta-hg@frantovo.cz>
parents: 7
diff changeset
   101
			if (currentAttributeIndex > 0 && currentAttributeIndex % currentSearchableAttributes.size() == 0) {
f66c759d1111 first working grep version
František Kučera <franta-hg@frantovo.cz>
parents: 7
diff changeset
   102
				if (includeCurrentRecord) for (string_t v : currentRecord) relationalWriter->writeAttribute(v);
f66c759d1111 first working grep version
František Kučera <franta-hg@frantovo.cz>
parents: 7
diff changeset
   103
				includeCurrentRecord = false;
f66c759d1111 first working grep version
František Kučera <franta-hg@frantovo.cz>
parents: 7
diff changeset
   104
			}
f66c759d1111 first working grep version
František Kučera <franta-hg@frantovo.cz>
parents: 7
diff changeset
   105
f66c759d1111 first working grep version
František Kučera <franta-hg@frantovo.cz>
parents: 7
diff changeset
   106
			currentAttributeIndex = currentAttributeIndex % currentSearchableAttributes.size();
f66c759d1111 first working grep version
František Kučera <franta-hg@frantovo.cz>
parents: 7
diff changeset
   107
		} else {
f66c759d1111 first working grep version
František Kučera <franta-hg@frantovo.cz>
parents: 7
diff changeset
   108
			relationalWriter->writeAttribute(value);
f66c759d1111 first working grep version
František Kučera <franta-hg@frantovo.cz>
parents: 7
diff changeset
   109
		}
3
8731263d44f1 PassthroughHandler (string-based version)
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
   110
	}
8731263d44f1 PassthroughHandler (string-based version)
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
   111
8731263d44f1 PassthroughHandler (string-based version)
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
   112
	void endOfPipe() {
8731263d44f1 PassthroughHandler (string-based version)
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
   113
8731263d44f1 PassthroughHandler (string-based version)
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
   114
	}
8731263d44f1 PassthroughHandler (string-based version)
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
   115
8731263d44f1 PassthroughHandler (string-based version)
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
   116
};
8731263d44f1 PassthroughHandler (string-based version)
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
   117
8731263d44f1 PassthroughHandler (string-based version)
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
   118
}
8731263d44f1 PassthroughHandler (string-based version)
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
   119
}
8731263d44f1 PassthroughHandler (string-based version)
František Kučera <franta-hg@frantovo.cz>
parents:
diff changeset
   120
}