Skip to content

Commit

Permalink
Merge pull request #10 from msrosenberg/dev-branch
Browse files Browse the repository at this point in the history 
3.0.8 pull
  • Loading branch information
@msrosenberg
msrosenberg authored Nov 11, 2022
2 parents 467c0c0 + a9822e7 commit 96930ad
Show file tree
Hide file tree
Showing 4 changed files with 71 additions and 65 deletions.
4 changes: 2 additions & 2 deletions requirements.txt
View file
Original file line number Diff line number Diff line change
@@ -1,4 +1,4 @@
numpy~=1.23.0
PyQt6~=6.3.0
PyQt6~=6.4.0
scipy~=1.9.0
matplotlib~=3.5.2
matplotlib~=3.6.0
125 changes: 65 additions & 60 deletions src/MetaWinAnalysisFunctions.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -375,15 +375,24 @@ def calc_i2(qt, n, alpha: float = 0.05):

# --- common output ---
def mean_effects_table(effect_label, mean_effect_data, bootstrap_mean, decimal_places: int = 4, alpha: float = 0.05,
log_transformed: bool = False) -> list:
log_transformed: bool = False, inc_median: bool = True) -> list:
"""
create a table of mean effect sizes and their confidence intervals
if log transformed, add rows with the exponentiated (unlogged) results as well
"""
output = []
col_headers = ["", "n", get_text("Mean"), get_text("Median"), "{:0.0%} CI".format(1 - alpha)]
col_formats = ["", "d", "f", "f", ""]
# col_headers = ["", "n", get_text("Mean"), get_text("Median"), "{:0.0%} CI".format(1 - alpha)]
# col_formats = ["", "d", "f", "f", ""]

col_headers = ["", "n", get_text("Mean")]
col_formats = ["", "d", "f"]
if inc_median:
col_headers.append(get_text("Median"))
col_formats.append("f")
col_headers.append("{:0.0%} CI".format(1 - alpha))
col_formats.append("")

if bootstrap_mean is not None:
col_headers.extend(["Bootstrap CI", "Bias-corrected CI"])
col_formats.extend(["", ""])
Expand All @@ -394,7 +403,11 @@ def mean_effects_table(effect_label, mean_effect_data, bootstrap_mean, decimal_p
row_name = data.name + " " + effect_label
else:
row_name = data.name
row_data = [row_name, data.n, data.mean, data.median, interval_str]
# row_data = [row_name, data.n, data.mean, data.median, interval_str]
row_data = [row_name, data.n, data.mean]
if inc_median:
row_data.append(data.median)
row_data.append(interval_str)
if bootstrap_mean is not None:
row_data.append(interval_to_str(data.lower_bs_ci, data.upper_bs_ci, decimal_places))
row_data.append(interval_to_str(data.lower_bias_ci, data.upper_bias_ci, decimal_places))
Expand All @@ -403,12 +416,19 @@ def mean_effects_table(effect_label, mean_effect_data, bootstrap_mean, decimal_p
line_after = [len(mean_effect_data)-1]
for data in mean_effect_data:
tmp_mean = math.exp(data.mean)
tmp_median = math.exp(data.median)
if inc_median:
tmp_median = math.exp(data.median)
else:
tmp_median = None
tmp_lower = math.exp(data.lower_ci)
tmp_upper = math.exp(data.upper_ci)
tmp_interval = interval_to_str(tmp_lower, tmp_upper, decimal_places)
row_data = [str(data.name + " " + format(exponential_label(effect_label))).strip(), data.n, tmp_mean,
tmp_median, tmp_interval]
# row_data = [str(data.name + " " + format(exponential_label(effect_label))).strip(), data.n, tmp_mean,
# tmp_median, tmp_interval]
row_data = [str(data.name + " " + format(exponential_label(effect_label))).strip(), data.n, tmp_mean]
if inc_median:
row_data.append(tmp_median)
row_data.append(tmp_interval)
if bootstrap_mean is not None:
tmp_bs_lower = math.exp(data.lower_bs_ci)
tmp_bs_upper = math.exp(data.upper_bs_ci)
Expand Down Expand Up @@ -485,7 +505,8 @@ def i2_table(i2_data, decimal_places: int = 4, alpha: float = 0.05) -> list:


def create_global_output(output_blocks: list, effect_label: str, mean_data, het_data, pooled_var, i2_data,
bootstrap_mean, decimal_places: int = 4, alpha: float = 0.05, log_transformed: bool = False):
bootstrap_mean, decimal_places: int = 4, alpha: float = 0.05, log_transformed: bool = False,
inc_median: bool = True):
"""
create output for general global calculations (e.g., grand mean, Qtotal)
"""
Expand All @@ -496,7 +517,7 @@ def create_global_output(output_blocks: list, effect_label: str, mean_data, het_

output_blocks.append(["<h4>Mean Effect Size</h4>"])
output_blocks.append(mean_effects_table(effect_label, [mean_data], bootstrap_mean, decimal_places, alpha,
log_transformed))
log_transformed, inc_median=inc_median))

sqrt_pooled = math.sqrt(pooled_var)
ratio = sqrt_pooled / mean_data.avg_var
Expand Down Expand Up @@ -876,7 +897,7 @@ def cumulative_meta_analysis(data, options, decimal_places: int = 4, alpha: floa
if row.not_filtered():
e = data.check_value(r, effect_sizes.position(), value_type=MetaWinConstants.VALUE_NUMBER)
v = data.check_value(r, variances.position(), value_type=MetaWinConstants.VALUE_NUMBER)
g = data.check_value(r, effect_sizes.position(), value_type=MetaWinConstants.VALUE_ANY)
g = data.check_value(r, order.position(), value_type=MetaWinConstants.VALUE_ANY)
if (e is not None) and (v is not None) and (v > 0) and (g is not None):
tmp_data.append([g, e, v])
else:
Expand Down Expand Up @@ -1957,11 +1978,11 @@ def phylogenetic_meta_analysis(data, options, tree, decimal_places: int = 4, alp
row.append(col[r])

e_data = numpy.array(e_data)
w_data = numpy.array(w_data)
# w_data = numpy.array(w_data)
n = len(e_data)
v_data = numpy.array(v_data)
x_data = numpy.array(x_data)
boot_data = numpy.array(boot_data)
# boot_data = numpy.array(boot_data)

p_matrix = phylogenetic_correlation(tip_names, tree)

Expand All @@ -1972,16 +1993,6 @@ def phylogenetic_meta_analysis(data, options, tree, decimal_places: int = 4, alp
v_matrix[i, j] = p_matrix[i, j]*math.sqrt(v_data[i])*math.sqrt(v_data[j])
v_matrix[j, i] = v_matrix[i, j]
w_matrix = numpy.linalg.inv(v_matrix)
# w_matrix = numpy.diag(w_data) # convert w to n x n matrix with w on the diagonal
# # add phylogenetic covariances to w_matrix
# for i in range(n):
# for j in range(i):
# if p_matrix[i, j] == 0:
# w_matrix[i, j] = 0
# else:
# # w_matrix[i, j] = 1/(p_matrix[i, j]*math.sqrt(v_data[i])*math.sqrt(v_data[j]))
# w_matrix[i, j] = p_matrix[i, j]*math.sqrt(w_data[i])*math.sqrt(w_data[j])
# w_matrix[j, i] = w_matrix[i, j]

if len(missing_from_tree) > 0:
print(missing_from_tree)
Expand All @@ -1995,10 +2006,10 @@ def phylogenetic_meta_analysis(data, options, tree, decimal_places: int = 4, alp



model_het = None
error_het = None
predictor_table_data = None
global_values = None
# model_het = None
# error_het = None
# predictor_table_data = None
# global_values = None
citations = []
if check_data_for_glm(output_blocks, n, tmp_x_data, [c.label for c in categorical_vars]):
output_blocks.append([get_text("{} studies will be included in this analysis").format(n)])
Expand All @@ -2008,13 +2019,7 @@ def phylogenetic_meta_analysis(data, options, tree, decimal_places: int = 4, alp
output_blocks.append([get_text("Estimate of pooled variance") + ": " +
format(pooled_var, inline_float(decimal_places))])
vs_data = v_data + pooled_var
ws_data = numpy.reciprocal(v_data + pooled_var)
# w_matrix = numpy.diag(ws_data)
# # add phylogenetic covariances to w_matrix
# for i in range(n):
# for j in range(i):
# w_matrix[i, j] = 1 / (p_matrix[i, j] * math.sqrt(v_data[i]) * math.sqrt(v_data[j]))
# w_matrix[j, i] = w_matrix[i, j]
# ws_data = numpy.reciprocal(v_data + pooled_var)
v_matrix = numpy.diag(vs_data) # convert w to n x n matrix with w on the diagonal
# add phylogenetic covariances to v_matrix
for i in range(n):
Expand All @@ -2024,9 +2029,9 @@ def phylogenetic_meta_analysis(data, options, tree, decimal_places: int = 4, alp
w_matrix = numpy.linalg.inv(v_matrix)

qm, qe, beta, sigma_b = calculate_glm(e_data, x_data, w_matrix)
else:
ws_data = w_data
median_e = median_effect(e_data, ws_data)
# else:
# ws_data = w_data
# median_e = median_effect(e_data, ws_data)

qt = qm + qe
df = n-1
Expand Down Expand Up @@ -2090,30 +2095,30 @@ def phylogenetic_meta_analysis(data, options, tree, decimal_places: int = 4, alp
output_blocks.append(predictor_table(predictor_table_data, decimal_places))

global_het_data = heterogeneity_test_tuple(get_text("Total"), qt, df, pqt, "")
# if has_model:
# output_blocks.append(["<h4>{}</h4>".format(get_text("Heterogeneity"))])
# model_het = heterogeneity_test_tuple(get_text("Model"), qm, dfm, pqm, p_random_str)
# error_het = heterogeneity_test_tuple(get_text("Error"), qe, dfe, pqe, "")
# model_table = [model_het, error_het, global_het_data]
# output_blocks.append(heterogeneity_table(model_table, decimal_places, total_line=True,
# randomization=options.randomize_model))
# else:
# """ need to pull the values from already calculated model """
# mean_e = beta[0]
# var_e = sigma_b[0, 0]
# mean_v = numpy.sum(v_data) / n
# lower_ci, upper_ci = scipy.stats.t.interval(alpha=1 - alpha, df=df, loc=mean_e, scale=math.sqrt(var_e))
# mean_data = mean_data_tuple(get_text("Global"), 0, n, mean_e, None, var_e, mean_v, lower_ci, upper_ci,
# 0, 0, 0, 0)
#
# i2, i2_lower, i2_upper = calc_i2(qt, n, alpha)
# i2_data = [i2_values(get_text("Total"), i2, i2_lower, i2_upper)]
#
# output_blocks.append(["<h3>{}</h3>".format(get_text("Global Results"))])
# new_cites = create_global_output(output_blocks, effect_sizes.label, mean_data, global_het_data, pooled_var,
# i2_data, options.bootstrap_mean, decimal_places, alpha,
# options.log_transformed)
# citations.extend(new_cites)
if has_model:
output_blocks.append(["<h4>{}</h4>".format(get_text("Heterogeneity"))])
model_het = heterogeneity_test_tuple(get_text("Model"), qm, dfm, pqm, p_random_str)
error_het = heterogeneity_test_tuple(get_text("Error"), qe, dfe, pqe, "")
model_table = [model_het, error_het, global_het_data]
output_blocks.append(heterogeneity_table(model_table, decimal_places, total_line=True,
randomization=options.randomize_model))
else:
""" need to pull the values from already calculated model """
mean_e = beta[0]
var_e = sigma_b[0, 0]
mean_v = numpy.sum(v_data) / n
lower_ci, upper_ci = scipy.stats.t.interval(alpha=1 - alpha, df=df, loc=mean_e, scale=math.sqrt(var_e))
mean_data = mean_data_tuple(get_text("Global"), 0, n, mean_e, None, var_e, mean_v, lower_ci, upper_ci,
0, 0, 0, 0)

i2, i2_lower, i2_upper = calc_i2(qt, n, alpha)
i2_data = [i2_values(get_text("Total"), i2, i2_lower, i2_upper)]

output_blocks.append(["<h3>{}</h3>".format(get_text("Global Results"))])
new_cites = create_global_output(output_blocks, effect_sizes.label, mean_data, global_het_data, pooled_var,
i2_data, options.bootstrap_mean, decimal_places, alpha,
options.log_transformed, inc_median=False)
citations.extend(new_cites)

return output_blocks, citations

Expand Down
2 changes: 1 addition & 1 deletion src/MetaWinConstants.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -13,7 +13,7 @@

MAJOR_VERSION = 3
MINOR_VERSION = 0
PATCH_VERSION = 7
PATCH_VERSION = 8

# validity check when fetching value from data matrix
VALUE_NUMBER = 0
Expand Down
5 changes: 3 additions & 2 deletions src/MetaWinUtils.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -95,7 +95,8 @@ def create_output_table(output_text: list, table_data: list, col_headers: list,
cols = [format(h, "^{}".format(max_width[i])) for i, h in enumerate(col_headers)]
header = col_spacer.join(cols)
output_text.append(header)
output_text.append("-"*(sum(max_width) + (len(cols)-1)*sbc))
header_line_width = sum(max_width) + (len(cols)-1)*sbc
output_text.append("-"*header_line_width)

# create table data template
cols = []
Expand Down Expand Up @@ -126,7 +127,7 @@ def create_output_table(output_text: list, table_data: list, col_headers: list,
output_text.append(template.format(*row))
if line_after is not None:
if r in line_after:
output_text.append("-"*len(header))
output_text.append("-"*header_line_width)

# add format strings
output_text[0] = "<code><pre>" + output_text[0]
Expand Down

0 comments on commit 96930ad

Please sign in to comment.