ClickHouse/dbms/src/Storages/StorageCatBoostPool.h

86 lines
2.6 KiB
C++
Raw Normal View History

2017-10-30 14:38:14 +00:00
#pragma once
#include <Storages/IStorage.h>
#include <Core/Defines.h>
#include <ext/shared_ptr_helper.h>
namespace DB
{
class StorageCatBoostPool : private ext::shared_ptr_helper<StorageCatBoostPool>, public IStorage
{
public:
static StoragePtr create(const Context & context,
const String & column_description_file_name, const String & data_description_file_name);
2017-10-30 14:38:14 +00:00
std::string getName() const override { return "CatBoostPool"; }
std::string getTableName() const override { return table_name; }
const NamesAndTypesList & getColumnsListImpl() const override { return columns; }
BlockInputStreams read(const Names & column_names,
const SelectQueryInfo & query_info,
const Context & context,
QueryProcessingStage::Enum & processed_stage,
size_t max_block_size,
unsigned threads) override;
private:
String table_name;
NamesAndTypesList columns;
String column_description_file_name;
String data_description_file_name;
Block sample_block;
enum class DatasetColumnType
{
Target,
Num,
Categ,
Auxiliary,
DocId,
Weight,
Baseline
};
using ColumnTypesMap = std::map<std::string, DatasetColumnType>;
ColumnTypesMap getColumnTypesMap() const
{
return {
{"Target", DatasetColumnType::Target},
{"Num", DatasetColumnType::Num},
{"Categ", DatasetColumnType::Categ},
{"Auxiliary", DatasetColumnType::Auxiliary},
{"DocId", DatasetColumnType::DocId},
{"Weight", DatasetColumnType::Weight},
{"Baseline", DatasetColumnType::Baseline},
};
};
std::string getColumnTypesString(const ColumnTypesMap & columnTypesMap);
struct ColumnDescription
{
std::string column_name;
std::string alias;
2017-10-30 14:38:14 +00:00
DatasetColumnType column_type;
ColumnDescription() : column_type(DatasetColumnType::Num) {}
ColumnDescription(std::string column_name, std::string alias, DatasetColumnType column_type)
: column_name(std::move(column_name)), alias(std::move(alias)), column_type(column_type) {}
2017-10-30 14:38:14 +00:00
};
std::vector<ColumnDescription> columns_description;
void checkDatasetDescription();
void parseColumnDescription();
void createSampleBlockAndColumns();
protected:
StorageCatBoostPool(const Context & context, String column_description_file_name, String data_description_file_name);
2017-10-30 14:38:14 +00:00
};
}