relpipe/relpipe-in-ods.cpp: src/ODSCommand.h@e82aaf24b0fe (annotated)

0 71b902e1c5ee project skeleton (taken from relpipe-in-xmltable) František Kučera <franta-hg@frantovo.cz> parents: diff changeset	1	/**
71b902e1c5ee project skeleton (taken from relpipe-in-xmltable) František Kučera <franta-hg@frantovo.cz> parents: diff changeset	2	* Relational pipes
1 e82aaf24b0fe first version, prototype, works sometimes František Kučera <franta-hg@frantovo.cz> parents: 0 diff changeset	3	* Copyright © 2023 František Kučera (Frantovo.cz, GlobalCode.info)
0 71b902e1c5ee project skeleton (taken from relpipe-in-xmltable) František Kučera <franta-hg@frantovo.cz> parents: diff changeset	4	*
71b902e1c5ee project skeleton (taken from relpipe-in-xmltable) František Kučera <franta-hg@frantovo.cz> parents: diff changeset	5	* This program is free software: you can redistribute it and/or modify
71b902e1c5ee project skeleton (taken from relpipe-in-xmltable) František Kučera <franta-hg@frantovo.cz> parents: diff changeset	6	* it under the terms of the GNU General Public License as published by
71b902e1c5ee project skeleton (taken from relpipe-in-xmltable) František Kučera <franta-hg@frantovo.cz> parents: diff changeset	7	* the Free Software Foundation, version 3 of the License.
71b902e1c5ee project skeleton (taken from relpipe-in-xmltable) František Kučera <franta-hg@frantovo.cz> parents: diff changeset	8	*
71b902e1c5ee project skeleton (taken from relpipe-in-xmltable) František Kučera <franta-hg@frantovo.cz> parents: diff changeset	9	* This program is distributed in the hope that it will be useful,
71b902e1c5ee project skeleton (taken from relpipe-in-xmltable) František Kučera <franta-hg@frantovo.cz> parents: diff changeset	10	* but WITHOUT ANY WARRANTY; without even the implied warranty of
71b902e1c5ee project skeleton (taken from relpipe-in-xmltable) František Kučera <franta-hg@frantovo.cz> parents: diff changeset	11	* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
71b902e1c5ee project skeleton (taken from relpipe-in-xmltable) František Kučera <franta-hg@frantovo.cz> parents: diff changeset	12	* GNU General Public License for more details.
71b902e1c5ee project skeleton (taken from relpipe-in-xmltable) František Kučera <franta-hg@frantovo.cz> parents: diff changeset	13	*
71b902e1c5ee project skeleton (taken from relpipe-in-xmltable) František Kučera <franta-hg@frantovo.cz> parents: diff changeset	14	* You should have received a copy of the GNU General Public License
71b902e1c5ee project skeleton (taken from relpipe-in-xmltable) František Kučera <franta-hg@frantovo.cz> parents: diff changeset	15	* along with this program. If not, see <http://www.gnu.org/licenses/>.
71b902e1c5ee project skeleton (taken from relpipe-in-xmltable) František Kučera <franta-hg@frantovo.cz> parents: diff changeset	16	*/
71b902e1c5ee project skeleton (taken from relpipe-in-xmltable) František Kučera <franta-hg@frantovo.cz> parents: diff changeset	17	#pragma once
71b902e1c5ee project skeleton (taken from relpipe-in-xmltable) František Kučera <franta-hg@frantovo.cz> parents: diff changeset	18
71b902e1c5ee project skeleton (taken from relpipe-in-xmltable) František Kučera <franta-hg@frantovo.cz> parents: diff changeset	19	#include <cstdlib>
71b902e1c5ee project skeleton (taken from relpipe-in-xmltable) František Kučera <franta-hg@frantovo.cz> parents: diff changeset	20	#include <iostream>
71b902e1c5ee project skeleton (taken from relpipe-in-xmltable) František Kučera <franta-hg@frantovo.cz> parents: diff changeset	21	#include <string>
71b902e1c5ee project skeleton (taken from relpipe-in-xmltable) František Kučera <franta-hg@frantovo.cz> parents: diff changeset	22	#include <sstream>
71b902e1c5ee project skeleton (taken from relpipe-in-xmltable) František Kučera <franta-hg@frantovo.cz> parents: diff changeset	23	#include <vector>
71b902e1c5ee project skeleton (taken from relpipe-in-xmltable) František Kučera <franta-hg@frantovo.cz> parents: diff changeset	24	#include <algorithm>
71b902e1c5ee project skeleton (taken from relpipe-in-xmltable) František Kučera <franta-hg@frantovo.cz> parents: diff changeset	25	#include <exception>
71b902e1c5ee project skeleton (taken from relpipe-in-xmltable) František Kučera <franta-hg@frantovo.cz> parents: diff changeset	26	#include <regex>
71b902e1c5ee project skeleton (taken from relpipe-in-xmltable) František Kučera <franta-hg@frantovo.cz> parents: diff changeset	27
71b902e1c5ee project skeleton (taken from relpipe-in-xmltable) František Kučera <franta-hg@frantovo.cz> parents: diff changeset	28	#include <libxml++-2.6/libxml++/libxml++.h>
71b902e1c5ee project skeleton (taken from relpipe-in-xmltable) František Kučera <franta-hg@frantovo.cz> parents: diff changeset	29
71b902e1c5ee project skeleton (taken from relpipe-in-xmltable) František Kučera <franta-hg@frantovo.cz> parents: diff changeset	30	#include <relpipe/writer/typedefs.h>
71b902e1c5ee project skeleton (taken from relpipe-in-xmltable) František Kučera <franta-hg@frantovo.cz> parents: diff changeset	31
71b902e1c5ee project skeleton (taken from relpipe-in-xmltable) František Kučera <franta-hg@frantovo.cz> parents: diff changeset	32	namespace relpipe {
71b902e1c5ee project skeleton (taken from relpipe-in-xmltable) František Kučera <franta-hg@frantovo.cz> parents: diff changeset	33	namespace in {
1 e82aaf24b0fe first version, prototype, works sometimes František Kučera <franta-hg@frantovo.cz> parents: 0 diff changeset	34	namespace ods {
0 71b902e1c5ee project skeleton (taken from relpipe-in-xmltable) František Kučera <franta-hg@frantovo.cz> parents: diff changeset	35
71b902e1c5ee project skeleton (taken from relpipe-in-xmltable) František Kučera <franta-hg@frantovo.cz> parents: diff changeset	36	using namespace relpipe::writer;
71b902e1c5ee project skeleton (taken from relpipe-in-xmltable) František Kučera <franta-hg@frantovo.cz> parents: diff changeset	37
1 e82aaf24b0fe first version, prototype, works sometimes František Kučera <franta-hg@frantovo.cz> parents: 0 diff changeset	38	class ODSCommand {
0 71b902e1c5ee project skeleton (taken from relpipe-in-xmltable) František Kučera <franta-hg@frantovo.cz> parents: diff changeset	39	private:
1 e82aaf24b0fe first version, prototype, works sometimes František Kučera <franta-hg@frantovo.cz> parents: 0 diff changeset	40	std::wstring_convert<codecvt_utf8<wchar_t>> convertor;
0 71b902e1c5ee project skeleton (taken from relpipe-in-xmltable) František Kučera <franta-hg@frantovo.cz> parents: diff changeset	41
1 e82aaf24b0fe first version, prototype, works sometimes František Kučera <franta-hg@frantovo.cz> parents: 0 diff changeset	42	xmlpp::Node::PrefixNsMap ns;
e82aaf24b0fe first version, prototype, works sometimes František Kučera <franta-hg@frantovo.cz> parents: 0 diff changeset	43	std::shared_ptr<RelationalWriter> writer;
e82aaf24b0fe first version, prototype, works sometimes František Kučera <franta-hg@frantovo.cz> parents: 0 diff changeset	44
e82aaf24b0fe first version, prototype, works sometimes František Kučera <franta-hg@frantovo.cz> parents: 0 diff changeset	45	string_t xpath(xmlpp::Node* node, std::string xpath) {
e82aaf24b0fe first version, prototype, works sometimes František Kučera <franta-hg@frantovo.cz> parents: 0 diff changeset	46	return convertor.from_bytes(node->eval_to_string(xpath, ns));
0 71b902e1c5ee project skeleton (taken from relpipe-in-xmltable) František Kučera <franta-hg@frantovo.cz> parents: diff changeset	47	}
71b902e1c5ee project skeleton (taken from relpipe-in-xmltable) František Kučera <franta-hg@frantovo.cz> parents: diff changeset	48
1 e82aaf24b0fe first version, prototype, works sometimes František Kučera <franta-hg@frantovo.cz> parents: 0 diff changeset	49	void processCell(xmlpp::Node* c, AttributeMetadata& am) {
e82aaf24b0fe first version, prototype, works sometimes František Kučera <franta-hg@frantovo.cz> parents: 0 diff changeset	50	string_t value = xpath(c, "@o:value");
e82aaf24b0fe first version, prototype, works sometimes František Kučera <franta-hg@frantovo.cz> parents: 0 diff changeset	51	if (value.size()) writer->writeAttribute(value);
e82aaf24b0fe first version, prototype, works sometimes František Kučera <franta-hg@frantovo.cz> parents: 0 diff changeset	52	else writer->writeAttribute(xpath(c, "tx:p"));
e82aaf24b0fe first version, prototype, works sometimes František Kučera <franta-hg@frantovo.cz> parents: 0 diff changeset	53	// TODO: be aware of the current data types
0 71b902e1c5ee project skeleton (taken from relpipe-in-xmltable) František Kučera <franta-hg@frantovo.cz> parents: diff changeset	54	}
71b902e1c5ee project skeleton (taken from relpipe-in-xmltable) František Kučera <franta-hg@frantovo.cz> parents: diff changeset	55
1 e82aaf24b0fe first version, prototype, works sometimes František Kučera <franta-hg@frantovo.cz> parents: 0 diff changeset	56	void processRow(xmlpp::Node* r, std::vector<AttributeMetadata>& am) {
e82aaf24b0fe first version, prototype, works sometimes František Kučera <franta-hg@frantovo.cz> parents: 0 diff changeset	57	for (int i = 0; i < am.size(); i++) {
e82aaf24b0fe first version, prototype, works sometimes František Kučera <franta-hg@frantovo.cz> parents: 0 diff changeset	58	auto xpe = std::string("t:table-cell[")
e82aaf24b0fe first version, prototype, works sometimes František Kučera <franta-hg@frantovo.cz> parents: 0 diff changeset	59	+ std::to_string(i + 1)
e82aaf24b0fe first version, prototype, works sometimes František Kučera <franta-hg@frantovo.cz> parents: 0 diff changeset	60	+ "]";
e82aaf24b0fe first version, prototype, works sometimes František Kučera <franta-hg@frantovo.cz> parents: 0 diff changeset	61	auto cells = r->find(xpe, ns);
e82aaf24b0fe first version, prototype, works sometimes František Kučera <franta-hg@frantovo.cz> parents: 0 diff changeset	62	if (cells.size() == 1) {
e82aaf24b0fe first version, prototype, works sometimes František Kučera <franta-hg@frantovo.cz> parents: 0 diff changeset	63	processCell(cells[0], am[i]);
e82aaf24b0fe first version, prototype, works sometimes František Kučera <franta-hg@frantovo.cz> parents: 0 diff changeset	64	} else {
e82aaf24b0fe first version, prototype, works sometimes František Kučera <franta-hg@frantovo.cz> parents: 0 diff changeset	65	writer->writeAttribute(L"");
e82aaf24b0fe first version, prototype, works sometimes František Kučera <franta-hg@frantovo.cz> parents: 0 diff changeset	66	// TODO: support also other data types
e82aaf24b0fe first version, prototype, works sometimes František Kučera <franta-hg@frantovo.cz> parents: 0 diff changeset	67	}
e82aaf24b0fe first version, prototype, works sometimes František Kučera <franta-hg@frantovo.cz> parents: 0 diff changeset	68	}
e82aaf24b0fe first version, prototype, works sometimes František Kučera <franta-hg@frantovo.cz> parents: 0 diff changeset	69
e82aaf24b0fe first version, prototype, works sometimes František Kučera <franta-hg@frantovo.cz> parents: 0 diff changeset	70	// FIXME: support sparse data / missing values:
e82aaf24b0fe first version, prototype, works sometimes František Kučera <franta-hg@frantovo.cz> parents: 0 diff changeset	71	// <table:table-row table:number-rows-repeated="2">
e82aaf24b0fe first version, prototype, works sometimes František Kučera <franta-hg@frantovo.cz> parents: 0 diff changeset	72	// <table:table-cell table:number-columns-repeated="3"/>
e82aaf24b0fe first version, prototype, works sometimes František Kučera <franta-hg@frantovo.cz> parents: 0 diff changeset	73
0 71b902e1c5ee project skeleton (taken from relpipe-in-xmltable) František Kučera <franta-hg@frantovo.cz> parents: diff changeset	74	}
71b902e1c5ee project skeleton (taken from relpipe-in-xmltable) František Kučera <franta-hg@frantovo.cz> parents: diff changeset	75
1 e82aaf24b0fe first version, prototype, works sometimes František Kučera <franta-hg@frantovo.cz> parents: 0 diff changeset	76	void processTable(xmlpp::Node* t) {
e82aaf24b0fe first version, prototype, works sometimes František Kučera <franta-hg@frantovo.cz> parents: 0 diff changeset	77	auto relation = xpath(t, "@t:name");
e82aaf24b0fe first version, prototype, works sometimes František Kučera <franta-hg@frantovo.cz> parents: 0 diff changeset	78	std::vector<AttributeMetadata> metadata;
0 71b902e1c5ee project skeleton (taken from relpipe-in-xmltable) František Kučera <franta-hg@frantovo.cz> parents: diff changeset	79
1 e82aaf24b0fe first version, prototype, works sometimes František Kučera <franta-hg@frantovo.cz> parents: 0 diff changeset	80	for (xmlpp::Node* c : t->find("t:table-row[1]/t:table-cell", ns)) {
e82aaf24b0fe first version, prototype, works sometimes František Kučera <franta-hg@frantovo.cz> parents: 0 diff changeset	81	auto name = xpath(c, "tx:p");
e82aaf24b0fe first version, prototype, works sometimes František Kučera <franta-hg@frantovo.cz> parents: 0 diff changeset	82	if (name.size()) {
e82aaf24b0fe first version, prototype, works sometimes František Kučera <franta-hg@frantovo.cz> parents: 0 diff changeset	83	metadata.push_back({name, TypeId::STRING});
e82aaf24b0fe first version, prototype, works sometimes František Kučera <franta-hg@frantovo.cz> parents: 0 diff changeset	84	// TODO: detect and support other data types
e82aaf24b0fe first version, prototype, works sometimes František Kučera <franta-hg@frantovo.cz> parents: 0 diff changeset	85	}
0 71b902e1c5ee project skeleton (taken from relpipe-in-xmltable) František Kučera <franta-hg@frantovo.cz> parents: diff changeset	86	}
1 e82aaf24b0fe first version, prototype, works sometimes František Kučera <franta-hg@frantovo.cz> parents: 0 diff changeset	87
e82aaf24b0fe first version, prototype, works sometimes František Kučera <franta-hg@frantovo.cz> parents: 0 diff changeset	88	writer->startRelation(relation, metadata, true);
e82aaf24b0fe first version, prototype, works sometimes František Kučera <franta-hg@frantovo.cz> parents: 0 diff changeset	89
e82aaf24b0fe first version, prototype, works sometimes František Kučera <franta-hg@frantovo.cz> parents: 0 diff changeset	90	int i = 0;
e82aaf24b0fe first version, prototype, works sometimes František Kučera <franta-hg@frantovo.cz> parents: 0 diff changeset	91	for (xmlpp::Node* r : t->find("t:table-row", ns)) {
e82aaf24b0fe first version, prototype, works sometimes František Kučera <franta-hg@frantovo.cz> parents: 0 diff changeset	92	i++;
e82aaf24b0fe first version, prototype, works sometimes František Kučera <franta-hg@frantovo.cz> parents: 0 diff changeset	93	if (i == 1) continue; // skip header row
e82aaf24b0fe first version, prototype, works sometimes František Kučera <franta-hg@frantovo.cz> parents: 0 diff changeset	94	processRow(r, metadata);
e82aaf24b0fe first version, prototype, works sometimes František Kučera <franta-hg@frantovo.cz> parents: 0 diff changeset	95	}
0 71b902e1c5ee project skeleton (taken from relpipe-in-xmltable) František Kučera <franta-hg@frantovo.cz> parents: diff changeset	96	}
71b902e1c5ee project skeleton (taken from relpipe-in-xmltable) František Kučera <franta-hg@frantovo.cz> parents: diff changeset	97
71b902e1c5ee project skeleton (taken from relpipe-in-xmltable) František Kučera <franta-hg@frantovo.cz> parents: diff changeset	98	public:
71b902e1c5ee project skeleton (taken from relpipe-in-xmltable) František Kučera <franta-hg@frantovo.cz> parents: diff changeset	99
1 e82aaf24b0fe first version, prototype, works sometimes František Kučera <franta-hg@frantovo.cz> parents: 0 diff changeset	100	ODSCommand(std::shared_ptr<RelationalWriter> writer) : writer(writer) {
e82aaf24b0fe first version, prototype, works sometimes František Kučera <franta-hg@frantovo.cz> parents: 0 diff changeset	101	ns["o"] = "urn:oasis:names:tc:opendocument:xmlns:office:1.0";
e82aaf24b0fe first version, prototype, works sometimes František Kučera <franta-hg@frantovo.cz> parents: 0 diff changeset	102	ns["t"] = "urn:oasis:names:tc:opendocument:xmlns:table:1.0";
e82aaf24b0fe first version, prototype, works sometimes František Kučera <franta-hg@frantovo.cz> parents: 0 diff changeset	103	ns["tx"] = "urn:oasis:names:tc:opendocument:xmlns:text:1.0";
e82aaf24b0fe first version, prototype, works sometimes František Kučera <franta-hg@frantovo.cz> parents: 0 diff changeset	104	}
0 71b902e1c5ee project skeleton (taken from relpipe-in-xmltable) František Kučera <franta-hg@frantovo.cz> parents: diff changeset	105
1 e82aaf24b0fe first version, prototype, works sometimes František Kučera <franta-hg@frantovo.cz> parents: 0 diff changeset	106	void process(std::istream & input) {
0 71b902e1c5ee project skeleton (taken from relpipe-in-xmltable) František Kučera <franta-hg@frantovo.cz> parents: diff changeset	107	xmlpp::DomParser parser;
1 e82aaf24b0fe first version, prototype, works sometimes František Kučera <franta-hg@frantovo.cz> parents: 0 diff changeset	108	parser.parse_stream(input);
e82aaf24b0fe first version, prototype, works sometimes František Kučera <franta-hg@frantovo.cz> parents: 0 diff changeset	109
0 71b902e1c5ee project skeleton (taken from relpipe-in-xmltable) František Kučera <franta-hg@frantovo.cz> parents: diff changeset	110	xmlpp::Element* root = parser.get_document()->get_root_node();
1 e82aaf24b0fe first version, prototype, works sometimes František Kučera <franta-hg@frantovo.cz> parents: 0 diff changeset	111	auto spreadsheets = root->find("/o:document/o:body/o:spreadsheet", ns);
0 71b902e1c5ee project skeleton (taken from relpipe-in-xmltable) František Kučera <franta-hg@frantovo.cz> parents: diff changeset	112
1 e82aaf24b0fe first version, prototype, works sometimes František Kučera <franta-hg@frantovo.cz> parents: 0 diff changeset	113	if (spreadsheets.size() == 1) {
e82aaf24b0fe first version, prototype, works sometimes František Kučera <franta-hg@frantovo.cz> parents: 0 diff changeset	114	for (xmlpp::Node* table : spreadsheets[0]->find("t:table", ns)) {
e82aaf24b0fe first version, prototype, works sometimes František Kučera <franta-hg@frantovo.cz> parents: 0 diff changeset	115	processTable(table);
e82aaf24b0fe first version, prototype, works sometimes František Kučera <franta-hg@frantovo.cz> parents: 0 diff changeset	116	}
e82aaf24b0fe first version, prototype, works sometimes František Kučera <franta-hg@frantovo.cz> parents: 0 diff changeset	117	} else {
e82aaf24b0fe first version, prototype, works sometimes František Kučera <franta-hg@frantovo.cz> parents: 0 diff changeset	118	throw RelpipeWriterException(L"Invalid XML structure. "
e82aaf24b0fe first version, prototype, works sometimes František Kučera <franta-hg@frantovo.cz> parents: 0 diff changeset	119	"Expecting OpenDocument spreadsheet.");
0 71b902e1c5ee project skeleton (taken from relpipe-in-xmltable) František Kučera <franta-hg@frantovo.cz> parents: diff changeset	120	}
71b902e1c5ee project skeleton (taken from relpipe-in-xmltable) František Kučera <franta-hg@frantovo.cz> parents: diff changeset	121
1 e82aaf24b0fe first version, prototype, works sometimes František Kučera <franta-hg@frantovo.cz> parents: 0 diff changeset	122
0 71b902e1c5ee project skeleton (taken from relpipe-in-xmltable) František Kučera <franta-hg@frantovo.cz> parents: diff changeset	123	}
71b902e1c5ee project skeleton (taken from relpipe-in-xmltable) František Kučera <franta-hg@frantovo.cz> parents: diff changeset	124	};
71b902e1c5ee project skeleton (taken from relpipe-in-xmltable) František Kučera <franta-hg@frantovo.cz> parents: diff changeset	125
71b902e1c5ee project skeleton (taken from relpipe-in-xmltable) František Kučera <franta-hg@frantovo.cz> parents: diff changeset	126	}
71b902e1c5ee project skeleton (taken from relpipe-in-xmltable) František Kučera <franta-hg@frantovo.cz> parents: diff changeset	127	}
71b902e1c5ee project skeleton (taken from relpipe-in-xmltable) František Kučera <franta-hg@frantovo.cz> parents: diff changeset	128	}

author	František Kučera <franta-hg@frantovo.cz>
	Sun, 14 May 2023 02:13:46 +0200
branch	v_0
changeset 1	e82aaf24b0fe
parent 0	src/XMLTableCommand.h@71b902e1c5ee
child 2	d4e0472e8e5d
permissions	-rw-r--r--