xls2metametrik

#!/usr/bin/env python

import sys
import xlrd
import json

def null_to_none(s):
    if s == "":
        return None
    return s

def independent_variables(worksheet):
    for row_num in range(1, worksheet.nrows):
        row = worksheet.row_values(row_num)
        assert len(row) == 4
        yield (row)

def rrs_metadata(worksheet):
    keys = ["dependent_variable_name",
            "dependent_variable_description",
            "r_squared",
            "sample_size",
            "model_name",
            "regression_type"
            ]

    def answers():
        for i, row_num in enumerate(range(2, worksheet.nrows)):
            key = keys[i]
            row = worksheet.row_values(row_num)
            value = row[1]
            yield (key, value)

    return dict([ kv for kv in answers() ])

def get_effects(model, ivs):
    for iv in ivs:
        yield {
            "model_name": model["model_name"],
            "independent_variable": iv[0],
            "significance": null_to_none(iv[2]),
            "standard_error": null_to_none(iv[3])
            }

def run():
    _, filename = sys.argv

    workbook = xlrd.open_workbook(filename)
    sheet_names = workbook.sheet_names()

    required_sheet_names = [u'RegressionResultSet', u'IndependentVariables']

    assert sorted(sheet_names) == sorted(required_sheet_names)

    ws1 = workbook.sheet_by_name('IndependentVariables')

    ivs = [i for i in independent_variables(ws1) ]
    
    def var_data(v):
        return {
            "variable_name": v[0],
            "variable_description": v[1]
            }
    def iv_metadata():
        for iv in ivs:
            yield var_data(iv)

    iv_metadata = [ i for i in iv_metadata() ]

    ws2 = workbook.sheet_by_name('RegressionResultSet')

    rrs = rrs_metadata(ws2)
    dv = (rrs["dependent_variable_name"], rrs["dependent_variable_description"])

    model = {
        "model_name": rrs["model_name"],
        "model_type": rrs["regression_type"],
        "goodness_of_fit": rrs["r_squared"],
        "sample_size": rrs["sample_size"]
        }
    
    effects = [e for e in get_effects(model, ivs)]

    data = {
        "regression_result_set": { 
            "dependent_variable": var_data(dv),
            "independent_variables": iv_metadata,
            "effects": effects,
            "models": [ model ]
            }
        }
    print json.dumps(data, indent=2)


if __name__ == '__main__':
    run()