def spss_corr_apa_table_excel(mod_raw_data_df, output_df):
# could very easily use the summ_corr_apa_table function here as I pass identical data - seperate is preferred, however, as might be updated/adjusted in the future
correlation_label = mod_raw_data_df.iloc[0, 1]
variables_list = list(mod_raw_data_df.columns)[2:]
if global_vars.corr_table_triangle == "Upper triangle":
header_cols = variables_list[1:]
header_rows = variables_list[:-1]
elif global_vars.corr_table_triangle == "Lower triangle":
header_cols = variables_list[:-1]
header_rows = variables_list[1:]
elif global_vars.corr_table_triangle == "Both":
header_cols, header_rows = variables_list, variables_list
wb = Workbook()
ws = wb.active
ws.append([""] + header_cols)
for ind, var in enumerate(header_rows):
ws.cell(row=ind + 2, column=1).value = var
inside_loop_ind_start = 2
for outside_loop_ind in range(2, len(header_rows) + 2):
if global_vars.corr_table_triangle == "Upper triangle" or global_vars.corr_table_triangle == "Both":
outside_loop_var = ws.cell(row=outside_loop_ind, column=1).value
elif global_vars.corr_table_triangle == "Lower triangle":
outside_loop_var = ws.cell(row=1, column=outside_loop_ind).value
for inside_loop_ind in range(inside_loop_ind_start,
len(header_cols) + 2):
if global_vars.corr_table_triangle == "Upper triangle" or global_vars.corr_table_triangle == "Both":
inside_loop_var = ws.cell(row=1, column=inside_loop_ind).value
elif global_vars.corr_table_triangle == "Lower triangle":
inside_loop_var = ws.cell(row=inside_loop_ind, column=1).value
if outside_loop_var == inside_loop_var:
ws.cell(row=outside_loop_ind, column=inside_loop_ind).value = 1
else:
# here query method is not preferred as it is not only slower as it is much smaller dataset but also cannot refer to two different vaiables (colname and val)
df_filtered = output_df[(
(output_df["var1"] == outside_loop_var) &
(output_df["var2"] == inside_loop_var)) | (
(output_df["var1"] == inside_loop_var) &
(output_df["var2"] == outside_loop_var))].iloc[0]
r = df_filtered[correlation_label]
p = df_filtered["adjusted_pvalues"]
r = helper_funcs.correlations_format_val(r, p)
if global_vars.corr_table_triangle == "Upper triangle" or global_vars.corr_table_triangle == "Both":
ws.cell(row=outside_loop_ind,
column=inside_loop_ind).value = r
elif global_vars.corr_table_triangle == "Lower triangle":
ws.cell(row=inside_loop_ind,
column=outside_loop_ind).value = r
if global_vars.corr_table_triangle != "Both":
inside_loop_ind_start += 1
for row in range(1, len(header_rows) + 2):
for cell in ws[row]:
cell.alignment = global_vars.alignment_center
for cell in ws[1]:
cell.font = global_vars.font_header
for row in range(2, len(header_rows) + 2):
ws.cell(row=row, column=1).font = global_vars.font_header
for cell in ws[len(header_rows) + 1]:
cell.border = Border(bottom=global_vars.border_APA)
for cell in ws[1]:
cell.border = Border(top=global_vars.border_APA,
bottom=global_vars.border_APA)
table_notes = ["**p < 0.01", "*p < {}".format(global_vars.alpha_threshold)]
table_notes.append(
"Correlation Coefficient used: {}".format(correlation_label))
helper_funcs.add_table_notes(ws, table_notes)
helper_funcs.savefile(wb=wb)
def spss_corr_apa_table_word(mod_raw_data_df, output_df):
# could very easily use the summ_corr_apa_table function here as I pass identical data - seperate is preferred, however, as might be updated/adjusted in the future
correlation_label = mod_raw_data_df.iloc[0, 1]
variables_list = list(mod_raw_data_df.columns)[2:]
if global_vars.corr_table_triangle == "Upper triangle":
header_cols = variables_list[1:]
header_rows = variables_list[:-1]
elif global_vars.corr_table_triangle == "Lower triangle":
header_cols = variables_list[:-1]
header_rows = variables_list[1:]
elif global_vars.corr_table_triangle == "Both":
header_cols = [
x for x in variables_list
] # not straight up variables_list; pass by value vs pass by reference
header_rows = [x for x in variables_list]
# this adds an empty column where the significance signs will be placed for better presentation
# the code looks slightly wrong as it return a None array but it works
[header_cols.insert(x, "") for x in range(1, len(header_cols) * 2, 2)]
doc = Document()
table_rows_len = len(header_rows) + 1
table_cols_len = len(header_cols) + 1
table = doc.add_table(rows=table_rows_len, cols=table_cols_len)
for ind, var in enumerate([""] + header_cols):
table.cell(row_idx=0, col_idx=ind).text = var
for ind, var in enumerate([""] + header_rows):
table.cell(row_idx=ind, col_idx=0).text = var
if global_vars.corr_table_triangle == "Upper triangle":
inside_loop_ind_start = 1
for outside_loop_ind in range(1, table_rows_len):
outside_loop_var = table.cell(row_idx=outside_loop_ind,
col_idx=0).text
for inside_loop_ind in range(inside_loop_ind_start,
table_cols_len):
inside_loop_var = table.cell(row_idx=0,
col_idx=inside_loop_ind).text
if inside_loop_var == "" or outside_loop_var == "": # this allows to skip the columns designed for the significance signs
continue
else:
# here query method is not preferred as it is not only slower as it is much smaller dataset but also cannot refer to two different vaiables (colname and val)
df_filtered = output_df[(
(output_df["var1"] == outside_loop_var) &
(output_df["var2"] == inside_loop_var)) | (
(output_df["var1"] == inside_loop_var) &
(output_df["var2"] == outside_loop_var))].iloc[0]
r = df_filtered[correlation_label]
p = df_filtered["adjusted_pvalues"]
r = helper_funcs.correlations_format_val(r, p)
if "_" in r:
r, sign = r.split("_")
else:
sign = ""
table.cell(row_idx=outside_loop_ind,
col_idx=inside_loop_ind).text = r
table.cell(row_idx=outside_loop_ind,
col_idx=inside_loop_ind + 1).text = sign
inside_loop_ind_start += 2
elif global_vars.corr_table_triangle == "Lower triangle":
inside_loop_ind_start = 1
for outside_loop_ind in range(1, table_cols_len):
outside_loop_var = table.cell(row_idx=0,
col_idx=outside_loop_ind).text
if outside_loop_var == "":
continue
else:
for inside_loop_ind in range(inside_loop_ind_start,
table_rows_len):
inside_loop_var = table.cell(row_idx=inside_loop_ind,
col_idx=0).text
# here query method is not preferred as it is not only slower as it is much smaller dataset but also cannot refer to two different vaiables (colname and val)
df_filtered = output_df[(
(output_df["var1"] == outside_loop_var) &
(output_df["var2"] == inside_loop_var)) | (
(output_df["var1"] == inside_loop_var) &
(output_df["var2"] == outside_loop_var))].iloc[0]
r = df_filtered[correlation_label]
p = df_filtered["adjusted_pvalues"]
r = helper_funcs.correlations_format_val(r, p)
if "_" in r:
r, sign = r.split("_")
else:
sign = ""
table.cell(row_idx=inside_loop_ind,
col_idx=outside_loop_ind).text = r
table.cell(row_idx=inside_loop_ind,
col_idx=outside_loop_ind + 1).text = sign
inside_loop_ind_start += 1
elif global_vars.corr_table_triangle == "Both":
inside_loop_ind_start = 1
for outside_loop_ind in range(1, table_cols_len):
outside_loop_var = table.cell(row_idx=0,
col_idx=outside_loop_ind).text
if outside_loop_var == "":
continue
else:
for inside_loop_ind in range(inside_loop_ind_start,
table_rows_len):
inside_loop_var = table.cell(row_idx=inside_loop_ind,
col_idx=0).text
if inside_loop_var == "": # this allows to skip the columns designed for the significance signs
continue
else:
if outside_loop_var == inside_loop_var:
table.cell(row_idx=inside_loop_ind,
col_idx=outside_loop_ind).text = "1"
else:
# here query method is not preferred as it is not only slower as it is much smaller dataset but also cannot refer to two different vaiables (colname and val)
df_filtered = output_df[(
(output_df["var1"] == outside_loop_var) &
(output_df["var2"] == inside_loop_var)) | (
(output_df["var1"] == inside_loop_var) &
(output_df["var2"] == outside_loop_var)
)].iloc[0]
r = df_filtered[correlation_label]
p = df_filtered["adjusted_pvalues"]
r = helper_funcs.correlations_format_val(r, p)
if "_" in r:
r, sign = r.split("_")
else:
sign = ""
table.cell(row_idx=inside_loop_ind,
col_idx=outside_loop_ind).text = r
table.cell(row_idx=inside_loop_ind,
col_idx=outside_loop_ind +
1).text = sign
for cell in table.rows[0].cells:
helper_funcs.set_cell_border(cell,
top=global_vars.border_APA_word,
bottom=global_vars.border_APA_word)
for cell in table.rows[table_rows_len - 1].cells:
helper_funcs.set_cell_border(cell, bottom=global_vars.border_APA_word)
for cell in table.rows[0].cells:
helper_funcs.word_style(cell, italic=True)
for cell in table.columns[0].cells:
helper_funcs.word_style(cell, italic=True)
for i in range(1, table_cols_len, 2):
table.cell(row_idx=0,
col_idx=i).merge(table.cell(row_idx=0, col_idx=i + 1))
table.cell(row_idx=0, col_idx=i).text = table.cell(
row_idx=0, col_idx=i
).text[:
-1] # meging cells adds a paragraph break at the end; this gets rid of it
for cell in table.rows[0].cells:
helper_funcs.word_style(cell, italic=True)
cell.paragraphs[0].alignment = WD_ALIGN_PARAGRAPH.CENTER
for cell in table.columns[0].cells:
helper_funcs.word_style(cell, italic=True)
cell.paragraphs[0].alignment = WD_ALIGN_PARAGRAPH.CENTER
for row in range(0, table_rows_len):
for cell in table.rows[row].cells:
cell.vertical_alignment = WD_ALIGN_VERTICAL.CENTER
cell.paragraphs[0].paragraph_format.space_after = Inches(0)
cell.paragraphs[0].paragraph_format.space_before = Inches(0)
cell.paragraphs[0].paragraph_format.line_spacing = 1
for row in range(1, table_rows_len):
for col in range(1, table_cols_len, 2):
table.cell(
row_idx=row, col_idx=col
).paragraphs[0].paragraph_format.right_indent = Inches(-0.08)
table.cell(
row_idx=row, col_idx=col +
1).paragraphs[0].paragraph_format.left_indent = Inches(-0.06)
table.cell(
row_idx=row,
col_idx=col).paragraphs[0].alignment = WD_ALIGN_PARAGRAPH.RIGHT
table.cell(row_idx=row, col_idx=col +
1).paragraphs[0].alignment = WD_ALIGN_PARAGRAPH.LEFT
doc = helper_funcs.set_autofit(doc)
para = doc.add_paragraph("** ")
para.add_run("p").italic = True
para.add_run(" < 0.01")
para.add_run().add_break()
para.add_run("* ")
para.add_run("p").italic = True
para.add_run(" < {}".format(global_vars.alpha_threshold))
doc.add_paragraph(
"Correlation Coefficient used: {}".format(correlation_label))
helper_funcs.add_correction_message_word(doc)
helper_funcs.savefile(doc=doc)
def raw_corr_apa_table_noCIs_excel(mod_raw_data_df, output_df):
variables_list = list(
mod_raw_data_df.columns
) #local var and .columns preferred to minimize use of global scope
wb = Workbook()
ws = wb.active
if global_vars.corr_table_triangle == "Upper triangle":
header_cols = variables_list[1:]
header_rows = variables_list[:-1]
elif global_vars.corr_table_triangle == "Lower triangle":
header_cols = variables_list[:-1]
header_rows = variables_list[1:]
elif global_vars.corr_table_triangle == "Both":
header_cols = [x for x in variables_list]
header_rows = [x for x in variables_list]
ws.append([""] + header_cols)
for ind, var in enumerate(header_rows):
ws.cell(row=ind + 2, column=1).value = var
inside_loop_ind_start = 2
for outside_loop_ind in range(2, len(header_rows) + 2):
if global_vars.corr_table_triangle == "Upper triangle" or global_vars.corr_table_triangle == "Both":
outside_loop_var = ws.cell(row=outside_loop_ind, column=1).value
elif global_vars.corr_table_triangle == "Lower triangle":
outside_loop_var = ws.cell(row=1, column=outside_loop_ind).value
for inside_loop_ind in range(inside_loop_ind_start,
len(header_cols) + 2):
if global_vars.corr_table_triangle == "Upper triangle" or global_vars.corr_table_triangle == "Both":
inside_loop_var = ws.cell(row=1, column=inside_loop_ind).value
elif global_vars.corr_table_triangle == "Lower triangle":
inside_loop_var = ws.cell(row=inside_loop_ind, column=1).value
if outside_loop_var == inside_loop_var:
ws.cell(row=outside_loop_ind, column=inside_loop_ind).value = 1
else:
# here query method is not preferred as it is not only slower as it is much smaller dataset but also cannot refer to two different vaiables (colname and val)
df_filtered = output_df[(
(output_df["Variable1"] == outside_loop_var) &
(output_df["Variable2"] == inside_loop_var)) | (
(output_df["Variable1"] == inside_loop_var) &
(output_df["Variable2"] == outside_loop_var))].iloc[0]
r = df_filtered["Correlation_Coefficient"]
p = df_filtered["adjusted_pvalues"]
r = helper_funcs.correlations_format_val(r, p)
if global_vars.corr_table_triangle == "Upper triangle" or global_vars.corr_table_triangle == "Both":
ws.cell(row=outside_loop_ind,
column=inside_loop_ind).value = r
elif global_vars.corr_table_triangle == "Lower triangle":
ws.cell(row=inside_loop_ind,
column=outside_loop_ind).value = r
if global_vars.corr_table_triangle != "Both":
inside_loop_ind_start += 1
for row in range(1, len(header_rows) + 2):
for cell in ws[row]:
cell.alignment = global_vars.alignment_center
for cell in ws[1] + ws["A"]:
cell.font = global_vars.font_header
for cell in ws[1]:
cell.border = Border(top=global_vars.border_APA,
bottom=global_vars.border_APA)
for cell in ws[len(header_rows) + 1]:
cell.border = Border(bottom=global_vars.border_APA)
table_notes = [
"Correlation coefficient used: {}".format(
list(global_vars.master_dict.keys())[list(
global_vars.master_dict.values()).index(
global_vars.raw_corr_type)])
]
table_notes.append("**p < 0.01")
table_notes.append("*p < {}".format(global_vars.alpha_threshold))
helper_funcs.add_table_notes(ws, table_notes)
helper_funcs.savefile(wb=wb)
def raw_corr_apa_table_withCIs_word(mod_raw_data_df, output_df):
variables_list = list(mod_raw_data_df.columns)
if global_vars.corr_table_triangle == "Upper triangle":
header_cols = variables_list[1:]
header_rows = variables_list[:-1]
elif global_vars.corr_table_triangle == "Lower triangle":
header_cols = variables_list[:-1]
header_rows = variables_list[1:]
elif global_vars.corr_table_triangle == "Both":
header_cols = [x for x in variables_list]
header_rows = [x for x in variables_list]
# this adds an empty column where the significance signs will be placed for better presentation
# then adds empty rows for the CIs
# the code looks slightly wrong as it return a None array but it works
[header_cols.insert(x, "") for x in range(1, len(header_cols) * 2, 2)]
[header_rows.insert(x, "") for x in range(1, len(header_rows) * 2, 2)]
doc = Document()
table_rows_len = len(header_rows) + 1
table_cols_len = len(header_cols) + 1
table = doc.add_table(rows=table_rows_len, cols=table_cols_len)
for ind, var in enumerate([""] + header_cols):
table.cell(row_idx=0, col_idx=ind).text = var
for ind, var in enumerate([""] + header_rows):
table.cell(row_idx=ind, col_idx=0).text = var
if global_vars.corr_table_triangle == "Upper triangle":
inside_loop_ind_start = 1
for outside_loop_ind in range(1, table_rows_len, 2):
outside_loop_var = table.cell(row_idx=outside_loop_ind,
col_idx=0).text
for inside_loop_ind in range(inside_loop_ind_start, table_cols_len,
2):
inside_loop_var = table.cell(row_idx=0,
col_idx=inside_loop_ind).text
# here query method is not preferred as it is not only slower as it is much smaller dataset but also cannot refer to two different vaiables (colname and val)
df_filtered = output_df[(
(output_df["Variable1"] == outside_loop_var) &
(output_df["Variable2"] == inside_loop_var)) | (
(output_df["Variable1"] == inside_loop_var) &
(output_df["Variable2"] == outside_loop_var))].iloc[0]
r = df_filtered["Correlation_Coefficient"]
p = df_filtered["adjusted_pvalues"]
ci_low, ci_high = helper_funcs.correlations_format_val(
df_filtered["CI_low"]
), helper_funcs.correlations_format_val(df_filtered["CI_high"])
r = helper_funcs.correlations_format_val(r, p)
if "_" in r:
r, sign = r.split("_")
else:
sign = ""
table.cell(row_idx=outside_loop_ind,
col_idx=inside_loop_ind).text = r
table.cell(row_idx=outside_loop_ind,
col_idx=inside_loop_ind + 1).text = sign
table.cell(row_idx=outside_loop_ind + 1,
col_idx=inside_loop_ind
).text = "[" + ci_low + ", " + ci_high + "]"
inside_loop_ind_start += 2
elif global_vars.corr_table_triangle == "Lower triangle":
inside_loop_ind_start = 1
for outside_loop_ind in range(1, table_cols_len, 2):
outside_loop_var = table.cell(row_idx=0,
col_idx=outside_loop_ind).text
for inside_loop_ind in range(inside_loop_ind_start, table_rows_len,
2):
inside_loop_var = table.cell(row_idx=inside_loop_ind,
col_idx=0).text
# here query method is not preferred as it is not only slower as it is much smaller dataset but also cannot refer to two different vaiables (colname and val)
df_filtered = output_df[(
(output_df["Variable1"] == outside_loop_var) &
(output_df["Variable2"] == inside_loop_var)) | (
(output_df["Variable1"] == inside_loop_var) &
(output_df["Variable2"] == outside_loop_var))].iloc[0]
r = df_filtered["Correlation_Coefficient"]
p = df_filtered["adjusted_pvalues"]
ci_low, ci_high = helper_funcs.correlations_format_val(
df_filtered["CI_low"]
), helper_funcs.correlations_format_val(df_filtered["CI_high"])
r = helper_funcs.correlations_format_val(r, p)
if "_" in r:
r, sign = r.split("_")
else:
sign = ""
table.cell(row_idx=inside_loop_ind,
col_idx=outside_loop_ind).text = r
table.cell(row_idx=inside_loop_ind,
col_idx=outside_loop_ind + 1).text = sign
table.cell(row_idx=inside_loop_ind + 1,
col_idx=outside_loop_ind
).text = "[" + ci_low + ", " + ci_high + "]"
inside_loop_ind_start += 2
elif global_vars.corr_table_triangle == "Both":
inside_loop_ind_start = 1
for outside_loop_ind in range(1, table_cols_len, 2):
outside_loop_var = table.cell(row_idx=0,
col_idx=outside_loop_ind).text
for inside_loop_ind in range(inside_loop_ind_start, table_rows_len,
2):
inside_loop_var = table.cell(row_idx=inside_loop_ind,
col_idx=0).text
if outside_loop_var == inside_loop_var:
table.cell(row_idx=inside_loop_ind,
col_idx=outside_loop_ind).text = "1"
else:
# here query method is not preferred as it is not only slower as it is much smaller dataset but also cannot refer to two different vaiables (colname and val)
df_filtered = output_df[
((output_df["Variable1"] == outside_loop_var) &
(output_df["Variable2"] == inside_loop_var)) |
((output_df["Variable1"] == inside_loop_var) &
(output_df["Variable2"] == outside_loop_var))].iloc[0]
r = df_filtered["Correlation_Coefficient"]
p = df_filtered["adjusted_pvalues"]
ci_low, ci_high = helper_funcs.correlations_format_val(
df_filtered["CI_low"]
), helper_funcs.correlations_format_val(
df_filtered["CI_high"])
r = helper_funcs.correlations_format_val(r, p)
if "_" in r:
r, sign = r.split("_")
else:
sign = ""
table.cell(row_idx=inside_loop_ind,
col_idx=outside_loop_ind).text = r
table.cell(row_idx=inside_loop_ind,
col_idx=outside_loop_ind + 1).text = sign
table.cell(row_idx=inside_loop_ind + 1,
col_idx=outside_loop_ind
).text = "[" + ci_low + ", " + ci_high + "]"
for cell in table.rows[0].cells:
helper_funcs.set_cell_border(cell,
top=global_vars.border_APA_word,
bottom=global_vars.border_APA_word)
for cell in table.rows[table_rows_len - 1].cells:
helper_funcs.set_cell_border(cell, bottom=global_vars.border_APA_word)
for i in range(1, table_cols_len, 2):
table.cell(row_idx=0,
col_idx=i).merge(table.cell(row_idx=0, col_idx=i + 1))
table.cell(row_idx=0, col_idx=i).text = table.cell(
row_idx=0, col_idx=i
).text[:
-1] # meging cells adds a paragraph break at the end; this gets rid of it
for i in range(1, table_rows_len, 2):
table.cell(row_idx=i,
col_idx=0).merge(table.cell(row_idx=i + 1, col_idx=0))
table.cell(row_idx=i, col_idx=0).text = table.cell(
row_idx=i, col_idx=0
).text[:
-1] # meging cells adds a paragraph break at the end; this gets rid of it
for row in range(2, table_rows_len, 2):
for col in range(1, table_cols_len, 2):
table.cell(row_idx=row, col_idx=col).merge(
table.cell(row_idx=row, col_idx=col + 1))
table.cell(row_idx=row,
col_idx=col).alignment = WD_ALIGN_PARAGRAPH.CENTER
helper_funcs.word_style(table.cell(row_idx=row, col_idx=col),
size=9)
for cell in table.rows[0].cells:
helper_funcs.word_style(cell, italic=True)
cell.paragraphs[0].alignment = WD_ALIGN_PARAGRAPH.CENTER
cell.vertical_alignment = WD_ALIGN_VERTICAL.TOP
for cell in table.columns[0].cells:
helper_funcs.word_style(cell, italic=True)
cell.paragraphs[0].alignment = WD_ALIGN_PARAGRAPH.CENTER
for row in range(0, table_rows_len):
for col in range(1, table_cols_len):
cell = table.cell(row_idx=row, col_idx=col)
if col >= 1:
cell.vertical_alignment = WD_ALIGN_VERTICAL.CENTER
cell.paragraphs[0].paragraph_format.space_after = Inches(0)
cell.paragraphs[0].paragraph_format.space_before = Inches(0)
cell.paragraphs[0].paragraph_format.line_spacing = 1
for row in range(1, table_rows_len, 2):
for col in range(1, table_cols_len, 2):
table.cell(
row_idx=row, col_idx=col
).paragraphs[0].paragraph_format.right_indent = Inches(-0.08)
table.cell(
row_idx=row, col_idx=col +
1).paragraphs[0].paragraph_format.left_indent = Inches(-0.06)
table.cell(
row_idx=row,
col_idx=col).paragraphs[0].alignment = WD_ALIGN_PARAGRAPH.RIGHT
table.cell(row_idx=row, col_idx=col +
1).paragraphs[0].alignment = WD_ALIGN_PARAGRAPH.LEFT
doc = helper_funcs.set_autofit(doc)
para = doc.add_paragraph("** ")
para.add_run("p").italic = True
para.add_run(" < 0.01")
para.add_run().add_break()
para.add_run("* ")
para.add_run("p").italic = True
para.add_run(" < {}".format(global_vars.alpha_threshold))
doc.add_paragraph("Correlation coefficient used: {}".format(
list(global_vars.master_dict.keys())[list(
global_vars.master_dict.values()).index(
global_vars.raw_corr_type)]))
helper_funcs.add_correction_message_word(doc)
helper_funcs.savefile(doc=doc)
def raw_corr_apa_table_withCIs_excel(mod_raw_data_df, output_df):
variables_list = list(
mod_raw_data_df.columns
) # local var and .columns preferred to minimize use of global scope
wb = Workbook()
ws = wb.active
if global_vars.corr_table_triangle == "Upper triangle":
header_cols = variables_list[1:]
header_rows = variables_list[:-1]
elif global_vars.corr_table_triangle == "Lower triangle":
header_cols = variables_list[:-1]
header_rows = variables_list[1:]
elif global_vars.corr_table_triangle == "Both":
header_cols = [x for x in variables_list]
header_rows = [x for x in variables_list]
# this adds an empty row where the CIs will be
# the code looks slightly wrong as it return a None array but it works
[header_rows.insert(x, "") for x in range(1, len(header_rows) * 2, 2)]
ws.append([""] + header_cols)
for ind, var in enumerate(header_rows):
ws.cell(row=ind + 2, column=1).value = var
if global_vars.corr_table_triangle == "Upper triangle":
inside_loop_ind_start = 2
for outside_loop_ind in range(2, len(header_rows) + 2, 2):
outside_loop_var = ws.cell(row=outside_loop_ind, column=1).value
for inside_loop_ind in range(inside_loop_ind_start,
len(header_cols) + 2):
inside_loop_var = ws.cell(row=1, column=inside_loop_ind).value
# here query method is not preferred as it is not only slower as it is much smaller dataset but also cannot refer to two different vaiables (colname and val)
df_filtered = output_df[(
(output_df["Variable1"] == outside_loop_var) &
(output_df["Variable2"] == inside_loop_var)) | (
(output_df["Variable1"] == inside_loop_var) &
(output_df["Variable2"] == outside_loop_var))].iloc[0]
r = df_filtered["Correlation_Coefficient"]
p = df_filtered["adjusted_pvalues"]
r = helper_funcs.correlations_format_val(r, p)
ci_low, ci_high = helper_funcs.correlations_format_val(
df_filtered["CI_low"]
), helper_funcs.correlations_format_val(df_filtered["CI_high"])
ws.cell(row=outside_loop_ind, column=inside_loop_ind).value = r
ws.cell(row=outside_loop_ind + 1, column=inside_loop_ind
).value = "[" + ci_low + ", " + ci_high + "]"
inside_loop_ind_start += 1
elif global_vars.corr_table_triangle == "Lower triangle":
inside_loop_ind_start = 2
for outside_loop_ind in range(2, len(header_cols) + 2):
outside_loop_var = ws.cell(row=1, column=outside_loop_ind).value
for inside_loop_ind in range(inside_loop_ind_start,
len(header_rows) + 2, 2):
inside_loop_var = ws.cell(row=inside_loop_ind, column=1).value
# here query method is not preferred as it is not only slower as it is much smaller dataset but also cannot refer to two different vaiables (colname and val)
df_filtered = output_df[(
(output_df["Variable1"] == outside_loop_var) &
(output_df["Variable2"] == inside_loop_var)) | (
(output_df["Variable1"] == inside_loop_var) &
(output_df["Variable2"] == outside_loop_var))].iloc[0]
r = df_filtered["Correlation_Coefficient"]
p = df_filtered["adjusted_pvalues"]
r = helper_funcs.correlations_format_val(r, p)
ci_low, ci_high = helper_funcs.correlations_format_val(
df_filtered["CI_low"]
), helper_funcs.correlations_format_val(df_filtered["CI_high"])
ws.cell(row=inside_loop_ind, column=outside_loop_ind).value = r
ws.cell(row=inside_loop_ind + 1, column=outside_loop_ind
).value = "[" + ci_low + ", " + ci_high + "]"
inside_loop_ind_start += 2
elif global_vars.corr_table_triangle == "Both":
inside_loop_ind_start = 2
for outside_loop_ind in range(2, len(header_cols) + 2):
outside_loop_var = ws.cell(row=1, column=outside_loop_ind).value
for inside_loop_ind in range(inside_loop_ind_start,
len(header_rows) + 2, 2):
inside_loop_var = ws.cell(row=inside_loop_ind, column=1).value
if outside_loop_var == inside_loop_var:
ws.cell(row=inside_loop_ind,
column=outside_loop_ind).value = "1"
else:
# here query method is not preferred as it is not only slower as it is much smaller dataset but also cannot refer to two different vaiables (colname and val)
df_filtered = output_df[
((output_df["Variable1"] == outside_loop_var) &
(output_df["Variable2"] == inside_loop_var)) |
((output_df["Variable1"] == inside_loop_var) &
(output_df["Variable2"] == outside_loop_var))].iloc[0]
r = df_filtered["Correlation_Coefficient"]
p = df_filtered["adjusted_pvalues"]
r = helper_funcs.correlations_format_val(r, p)
ci_low, ci_high = helper_funcs.correlations_format_val(
df_filtered["CI_low"]
), helper_funcs.correlations_format_val(
df_filtered["CI_high"])
ws.cell(row=inside_loop_ind,
column=outside_loop_ind).value = r
ws.cell(row=inside_loop_ind + 1, column=outside_loop_ind
).value = "[" + ci_low + ", " + ci_high + "]"
for row in range(2, len(header_rows) + 1, 2):
ws.merge_cells(start_row=row,
start_column=1,
end_row=row + 1,
end_column=1)
if global_vars.corr_table_triangle == "Both":
for col in range(2, len(header_cols) + 2):
ws.merge_cells(start_row=col * 2 - 2,
start_column=col,
end_row=col * 2 - 1,
end_column=col)
for row in range(1, len(header_rows) + 2):
for cell in ws[row]:
cell.alignment = global_vars.alignment_center
for cell in ws[1] + ws["A"]:
cell.font = global_vars.font_header
for row in range(3, len(header_rows) + 2, 2):
for cell in ws[row]:
cell.font = Font(size=9)
for cell in ws[1]:
cell.border = Border(top=global_vars.border_APA,
bottom=global_vars.border_APA)
for cell in ws[len(header_rows) + 1]:
cell.border = Border(bottom=global_vars.border_APA)
table_notes = [
"Correlation coefficient used: {}".format(
list(global_vars.master_dict.keys())[list(
global_vars.master_dict.values()).index(
global_vars.raw_corr_type)])
]
table_notes.append("**p < 0.01")
table_notes.append("*p < {}".format(global_vars.alpha_threshold))
helper_funcs.add_table_notes(ws, table_notes)
helper_funcs.savefile(wb=wb)
def summ_corr_apa_table_excel(mod_raw_data_df, output_df):
# unique vars in var1 given by user's order and here ONLY adding vars from col 2 that are NOT in var 1
variables_list = list(output_df[global_vars.summ_corr_varOne].unique()) + list(set(output_df[global_vars.summ_corr_varTwo].unique()) - set(output_df[global_vars.summ_corr_varOne].unique()))
if global_vars.corr_table_triangle == "Upper triangle":
header_cols = variables_list[1:]
header_rows = variables_list[:-1]
elif global_vars.corr_table_triangle == "Lower triangle":
header_cols = variables_list[:-1]
header_rows = variables_list[1:]
elif global_vars.corr_table_triangle == "Both":
header_cols, header_rows = variables_list, variables_list
wb = Workbook()
ws = wb.active
ws.append([""] + header_cols)
for ind,var in enumerate(header_rows):
ws.cell(row=ind+2, column=1).value = var
inside_loop_ind_start = 2
for outside_loop_ind in range(2, len(header_rows) + 2):
if global_vars.corr_table_triangle == "Upper triangle" or global_vars.corr_table_triangle == "Both":
outside_loop_var = ws.cell(row=outside_loop_ind, column=1).value
elif global_vars.corr_table_triangle == "Lower triangle":
outside_loop_var = ws.cell(row=1, column=outside_loop_ind).value
for inside_loop_ind in range(inside_loop_ind_start, len(header_cols) + 2):
if global_vars.corr_table_triangle == "Upper triangle" or global_vars.corr_table_triangle == "Both":
inside_loop_var = ws.cell(row=1, column=inside_loop_ind).value
elif global_vars.corr_table_triangle == "Lower triangle":
inside_loop_var = ws.cell(row=inside_loop_ind, column=1).value
if outside_loop_var == inside_loop_var:
ws.cell(row=outside_loop_ind, column=inside_loop_ind).value = 1
else:
# here query method is not preferred as it is not only slower as it is much smaller dataset but also cannot refer to two different vaiables (colname and val)
df_filtered = output_df[((output_df[global_vars.summ_corr_varOne]==outside_loop_var) & (output_df[global_vars.summ_corr_varTwo]==inside_loop_var)) | ((output_df[global_vars.summ_corr_varOne]==inside_loop_var) & (output_df[global_vars.summ_corr_varTwo]==outside_loop_var))].iloc[0]
r = df_filtered[global_vars.summ_corr_coeff]
p = df_filtered["adjusted_pvalues"]
r = helper_funcs.correlations_format_val(r, p)
if global_vars.corr_table_triangle == "Upper triangle" or global_vars.corr_table_triangle == "Both":
ws.cell(row=outside_loop_ind, column=inside_loop_ind).value = r
elif global_vars.corr_table_triangle == "Lower triangle":
ws.cell(row=inside_loop_ind, column=outside_loop_ind).value = r
if global_vars.corr_table_triangle != "Both":
inside_loop_ind_start += 1
for row in range(1, len(header_rows) + 2):
for cell in ws[row]:
cell.alignment = global_vars.alignment_center
for cell in ws[1]:
cell.font = global_vars.font_header
for row in range(2, len(header_rows) + 2):
ws.cell(row=row, column=1).font = global_vars.font_header
for cell in ws[len(header_rows) + 1]:
cell.border = Border(bottom=global_vars.border_APA)
for cell in ws[1]:
cell.border = Border(top=global_vars.border_APA, bottom=global_vars.border_APA)
table_notes = ["**p < 0.01", "*p < {}".format(global_vars.alpha_threshold)]
helper_funcs.add_table_notes(ws, table_notes)
helper_funcs.savefile(wb=wb)
