/**

 * Relational pipes

 * Copyright © 2018 František Kučera (Frantovo.cz, GlobalCode.info)

 *

 * This program is free software: you can redistribute it and/or modify

 * it under the terms of the GNU General Public License as published by

 * the Free Software Foundation, version 3 of the License.

 *

 * This program is distributed in the hope that it will be useful,

 * but WITHOUT ANY WARRANTY; without even the implied warranty of

 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the

 * GNU General Public License for more details.

 *

 * You should have received a copy of the GNU General Public License

 * along with this program. If not, see <http://www.gnu.org/licenses/>.

 */

#include <cstdlib>

#include <vector>

#include <memory>

#include <locale>

#include <regex>

#include <algorithm>

#include <unistd.h>

#include <relpipe/writer/RelationalWriter.h>

#include <relpipe/writer/RelpipeWriterException.h>

#include <relpipe/writer/AttributeMetadata.h>

#include <relpipe/writer/Factory.h>

#include <relpipe/writer/TypeId.h>

#include <relpipe/cli/CLI.h>

#include "CSVCommand.h"

using namespace std;

using namespace relpipe::cli;

using namespace relpipe::writer;

namespace relpipe {

namespace in {

namespace csv {

bool CSVCommand::readValue(std::istream& input, std::stringstream& currentValue, bool& lastInRecord) {

	lastInRecord = false;

	char ch;

	input.get(ch);

	if (ch == '"') {

		while (input.get(ch)) {

			if (ch == '"') {

				input.get(ch);

				if (ch == '"') {

					currentValue << ch;

				} else {

					if (ch == '\r') input.get(ch);

					if (ch == '\n') lastInRecord = true;

					else if (ch != ',') throw RelpipeWriterException(L"Unexpected character (should be „\\n“ or „,“)");

					return true;

				}

			} else {

				currentValue << ch;

			}

		}

	} else if (ch == ',') {

		return true;

	} else if (ch == '\n') {

		lastInRecord = true;

		return true;

	} else if (ch == '\r') {

		input.get(ch);

		if (ch == '\n') {

			lastInRecord = true;

			return true;

		} else {

			throw RelpipeWriterException(L"Crazy carriage stuck during journey");

		}

	} else {

		for (currentValue << ch; input.get(ch);) {

			switch (ch) {

				case ',': return true;

				case '\r': break;

				case '\n':

					lastInRecord = true;

					return true;

				default: currentValue << ch;

			}

		}

	}

	return false;

}

/**

 * Data types might be encoded in the attribute names: name::type e.g. some_attribute::integer

 * 

 * TODO: share this code through relpipe-lib-infertypes (when available)

 */

void tryParseTypes(vector<AttributeMetadata>& metadata, std::shared_ptr<writer::RelationalWriter> writer, Configuration& configuration) {

	std::wregex pattern(L"(.*)::(.*)");

	std::wsmatch match;

	if (configuration.readTypes == Configuration::ReadTypes::AUTO || configuration.readTypes == Configuration::ReadTypes::TRUE) {

		bool hasTypes = true;

		std::vector<TypeId> types;

		std::vector<string_t> names;

		for (AttributeMetadata& am : metadata) {

			if (std::regex_match(am.attributeName, match, pattern)) {

				names.push_back(match[1]);

				if (configuration.readTypes == Configuration::ReadTypes::TRUE) {

					types.push_back(writer->toTypeId(match[2])); // must be valid type name otherwise exception is thrown

				} else {

					try {

						types.push_back(writer->toTypeId(match[2]));

					} catch (...) {

						hasTypes = false; // ignore exception and keep original names and default type (string)

					}

				}

			} else {

				hasTypes = false;

			}

		}

		if (hasTypes) {

			for (int i = 0, count = metadata.size(); i < count; i++) {

				metadata[i].attributeName = names[i];

				metadata[i].typeId = types[i];

			}

		} else if (configuration.readTypes == Configuration::ReadTypes::TRUE) {

			throw RelpipeWriterException(L"Types were expected in the CSV header, but not found.");

		}

	}

}

void CSVCommand::process(std::istream& input, std::shared_ptr<writer::RelationalWriter> writer, Configuration& configuration) {

	wstring_convert < codecvt_utf8<wchar_t>> convertor; // UTF-8 is required for CSV

	vector<AttributeMetadata> metadata;

	bool headerDone = false;

	bool lastInRecord = false;

	integer_t valueCount = 0;

	stringstream currentValue;

	while (readValue(input, currentValue, lastInRecord) && input.good()) {

		if (headerDone) {

			writer->writeAttribute(convertor.from_bytes(currentValue.str()));

			valueCount++;

		} else {

			AttributeMetadata am;

			am.attributeName = convertor.from_bytes(currentValue.str());

			am.typeId = TypeId::STRING;

			metadata.push_back(am);

			if (lastInRecord) {

				vector<string_t> firstLine;

				if (metadata.size() == configuration.attributes.size()) {

					for (int i = 0; i < metadata.size(); i++) {

						firstLine.push_back(metadata[i].attributeName);

						metadata[i].attributeName = configuration.attributes[i].name;

						metadata[i].typeId = configuration.attributes[i].type;

					}

				} else if (configuration.attributes.size() == 0) {

					// first line contains attribute names and maybe also types

					tryParseTypes(metadata, writer, configuration);

				} else {

					throw RelpipeWriterException(L"Declared attribute count (" + std::to_wstring(configuration.attributes.size()) + L") does not match with number of columns of the first line (" + std::to_wstring(metadata.size()) + L")");

				}

				headerDone = true;

				writer->startRelation(configuration.relation, metadata, true);

				if (firstLine.size()) {

					for (string_t value : firstLine) writer->writeAttribute(value);

				}

			}

		}

		currentValue.str("");

		currentValue.clear();

	}

	/**

	 * RFC 4180:

	 *  - Each line should contain the same number of fields throughout the file.

	 *  - The last field in the record must not be followed by a comma.

	 */

}

}

}

}