Performance issues

[brianmay]

Just need somewhere to record my data on performance issues. I think a issue might be the best place.

In my tests if I use:

            t0 = time.time()
            cls._compiled_workbook = ExcelCompiler(file=...)
            t1 = time.time()
            cls.calculator = cls._compiled_workbook.gen_graph()
            t2 = time.time()
            print("loaded", t1-t0, t2-t1, t2=t0)

I get the following times for 3 runs:

loaded 1.2965116500854492 8.90232515335083 10.19883680343628
loaded 1.3343768119812012 9.017873525619507 10.352250337600708
loaded 1.342972755432129 8.79265022277832 10.13562297821045

For the following:

        t0 = time.time()
        data = self.calculator.asdict()

        t1 = time.time()
        calculator = Spreadsheet.from_dict(data)

        t2 = time.time()
        print("done", t1-t0, t2-t1, t2-t0)

I get the following times for 3 runs:

done 0.2672147750854492 311.735009431839 312.00222420692444
done 0.18346762657165527 252.20295310020447 252.38642072677612
done 0.29434633255004883 322.1760091781616 322.47035551071167

For the following:

        t0 = time.time()
        self.calculator.dump("dump.txt.gz")

        t1 = time.time()
        calculator = calculator = Spreadsheet.load("dump.txt.gz")

        t2 = time.time()
        print("done", t1 - t0, t2 - t1, t2 - t0)

I get the following times for three runs:

done 0.7115657329559326 1.6663601398468018 2.3779258728027344
done 0.7441580295562744 1.4909374713897705 2.235095500946045
done 0.7094802856445312 1.4782490730285645 2.1877293586730957

While none of these times are anything special, I find it weird that from_dict() is much slower then anything else.

[danielsjf]

@brianmay, could you have another look at this performance after the latest fixes in #213? This could drastically improve your results.