Conversion program search system and conversion program search method

A conversion program search method for, according to a name of conversion-source data, a name of post-conversion data, and a conversion example including the conversion-source data and the post-conversion data, identifying a conversion path which is information indicating a relationship between the conversion-source data, the post-conversion data, input data of one or more conversion programs, and output data outputted by the one or more conversion programs, the conversion program search method comprising:

• generating a first signature with respect to each of the conversion-source data and the post-conversion data, by using first-signature generation unit, the first signature being information indicating, in a binary manner, whether one piece of data meets each of a predetermined plurality of description formats;
• generating a second signature based on a difference between the first signature of the conversion-source data and the first signature of the post-conversion data generated by the first-signature generation unit, by using second-signature generation unit, the second signature being metadata about the conversion program;
• storing the second signature in program information storage unit, by using the second-signature generation unit, the program information storage unit previously storing names of input data and output data of the conversion program and execution definition information for executing the conversion program;
• when receiving a name of conversion-source data, a name of post-conversion data, and the conversion example including conversion-source data and post-conversion data, selecting a conversion program candidate to be used in a data conversion process, based on a similarity between the received conversion-source data name and a name of input data of the conversion program and a similarity between the received post-conversion data name and a name of output data of the conversion program, by using program candidate selection unit;
• with respect to combinations of the candidates selected by the program candidate selection unit, associating a name of conversion-source data with the name of input data of the conversion program and associating the name of output data of the conversion program with a name of a post-conversion data, by using conversion path candidate generation unit and, when a combination of the selected candidates contains a plurality of conversion programs, generating a conversion path candidate by associating the name of output data of the conversion program with a name of input data of another candidate conversion program, by using the conversion path candidate generation unit;
• calculating an evaluation value of the conversion path candidate based on the second signature of the conversion program and the conversion example, by using evaluation value calculation unit, the conversion program having associated names of input data and output data with a name of another piece of data;
• and
• selecting a conversion path candidate in a descending order of the evaluation values, converting the received conversion-source data in an order of conversion programs determined by the selected conversion paths, by using conversion path identification unit and, when a result of the conversion is the received post-conversion data, identifying the selected conversion path candidate as the conversion path matching the received conversion-source data name, post-conversion data name, and the conversion example, by using the conversion path identification unit.

A non-transitory computer-readable storage medium recording a conversion program search program for causing a computer to, according to a name of conversion-source data, a name of post-conversion data, and a conversion example including the conversion-source data and the post-conversion data, identify a conversion path which is information indicating a relationship among the conversion-source data, the post-conversion data, input data of one or more conversion programs, and output data outputted by the one or more conversion programs, the computer including program information storage unit storing a names of input data and output data of a conversion program and execution definition information for executing the conversion program, the conversion program search program causing the computer to perform:

• a first signature generation process of generating a first signature with respect to each of the conversion-source data and the post-conversion data, the first signature being information indicating, in a binary manner, whether one piece of data meets each of a predetermined plurality of description formats;
• a second signature generation process of generating a second signature based on a difference between the first signature of the conversion-source data and the first signature of the post-conversion data generated in the first-signature generation process, the second signature being metadata about the conversion program;
• a second signature storage process of storing the second signature in program information storage unit;
• a program candidate selection process of, when receiving the name of the conversion-source data, the name of the post-conversion data, and the conversion example including the conversion-source data and the post-conversion data, selecting candidates for conversion programs to be used in a data conversion process, based on a similarity between the received conversion-source data name and a name of input data of the conversion program and a similarity between the received post-conversion data name and a name of output data of the conversion program;
• a conversion path candidate generation process of, with respect to combinations of the candidates selected in the program candidate selection process, associating the name of the conversion-source data with the name of the input data of the conversion program and associating the name of the output data of the conversion program with the name of the post-conversion data and, when a combination of the selected candidates contains a plurality of conversion programs, generating conversion path candidates by associating the name of the output data of the conversion program with a name of input data of another candidate conversion program;
• an evaluation value calculation process of calculating evaluation values of the conversion path candidates based on the second signature of the conversion program and the conversion example, the conversion program having associated names of input data and output data with a name of another piece of data;
• and
• a conversion path identification process of selecting a conversion path candidate in a descending order of the evaluation values, converting the received conversion-source data in an order of conversion programs determined by the selected conversion path, and, when a result of the conversion is the received post-conversion data, identifying the selected conversion path candidate as the conversion path matching the received conversion-source data name, post-conversion data name, and the conversion example.

A conversion program search system that, according to a name of conversion-source data, a name of post-conversion data, and a conversion example including the conversion-source data and the post-conversion data, identifies a conversion path which is information indicating a relationship among the conversion-source data, the post-conversion data, input data of one or more conversion programs, and output data outputted by the one or more conversion programs, the conversion program search system comprising:

• hardware including a processor;
• first-signature generation means implemented at least by the hardware and that generates a first signature which is information indicating, in a binary manner, whether one piece of data meets each of a predetermined plurality of description formats;
• second-signature generation means implemented at least by the hardware and that generates a second signature based on a difference between the first signature of the conversion-source data and the first signature of the post-conversion data generated by the first-signature generation means, the second signature being metadata about the conversion program;
• program information storage means implemented at least by the hardware and that stores names of input data and output data of the conversion program, execution definition information for executing the conversion program, and the second signature generated by the second-signature generation means;
• program candidate selection means implemented at least by the hardware and that, when receiving the name of the conversion-source data, the name of the post-conversion data, and the conversion example including the conversion-source data and the post-conversion data, selects candidates for conversion programs to be used in a data conversion process, based on a similarity between the received conversion-source data name and the name of the input data of the conversion program and a similarity between the received post-conversion data name and the name of output data of the conversion program;
• conversion path candidate generation means implemented at least by the hardware and that, with respect to combinations of the candidates selected by the program candidate selection means, associates the name of the conversion-source data with the name of input data of the conversion program and associates the name of output data of the conversion program with the name of the post-conversion data and that, when a combination of the selected candidates includes a plurality of conversion programs, generates a conversion path candidate by associating the name of output data of the conversion program with a name of input data of another candidate conversion program;
• evaluation value calculation means implemented at least by the hardware and that calculates evaluation values of conversion path candidates based on the second signature of the conversion program and the conversion example, the conversion program having associated names of input data and output data with a name of another piece of data;
• and
• conversion path identification means implemented at least by the hardware and that selects the conversion path candidates in a descending order of the evaluation values, converts the received conversion-source data in an order of conversion programs determined by the selected conversion path and, when a result of the conversion is the received post-conversion data, identifies the selected conversion path candidate as the conversion path matching the received conversion-source data name, post-conversion data name, and the conversion example.