def latex_table(inner_tabular: str,
position="H",
caption=None,
escape_caption=False,
label=None,
output_filepath=None):
table = Table(position=position)
table.append(NoEscape(inner_tabular))
if caption:
caption_ = caption if escape_caption else NoEscape(caption)
table.add_caption(caption_)
if label:
table.append(NoEscape("\label{{tab:{label}}}".format(label=label)))
if output_filepath:
table.generate_tex(output_filepath)
return table.dumps()
stats.kurtosis(spf_bal_UNEM_1Y.iloc[:, 0].values, axis=0),
stats.kurtosis(spf_bal_UNEM_2Y.iloc[:, 0].values, axis=0),
])
spf_desc_stat.iloc[8, :] = np.array([
stats.skew(spf_bal_RGDP_1Y.iloc[:, 0].values, axis=0),
stats.skew(spf_bal_RGDP_2Y.iloc[:, 0].values, axis=0),
stats.skew(spf_bal_HICP_1Y.iloc[:, 0].values, axis=0),
stats.skew(spf_bal_HICP_2Y.iloc[:, 0].values, axis=0),
stats.skew(spf_bal_UNEM_1Y.iloc[:, 0].values, axis=0),
stats.skew(spf_bal_UNEM_2Y.iloc[:, 0].values, axis=0),
])
# create table object
tabl = Table()
tabl.add_caption(
"Descriptive statistics of the SPF target macroeconomic variables for the euro area"
)
tabl.append(NoEscape('\label{tab: spf_Desc_Stats}'))
# create tabular object
tabr = Tabular(table_spec="l|cc|cc|cc")
tabr.add_hline()
tabr.add_hline()
# header row
tabr.add_row(
(MultiRow(2, data="Statistic"), MultiColumn(2, align='|c|', data="RGDP"),
MultiColumn(2, align='|c|',
data="HICP"), MultiColumn(2, align='|c', data="UNEM")))
tabr.add_hline(start=2, end=3, cmidruleoption="lr")
tabr.add_hline(start=4, end=5, cmidruleoption="lr")
tabr.add_hline(start=6, end=7, cmidruleoption="lr")
proc="multiple",
df_name="spf_bal_UNEM_2Y_" +
str(w))
# concatenate the tables together
spf_multitable = pd.concat([
acc_table_RGDP_1Y, acc_table_RGDP_2Y, acc_table_HICP_1Y,
acc_table_HICP_2Y, acc_table_UNEM_1Y, acc_table_UNEM_2Y
],
axis=1,
join_axes=[acc_table_RGDP_1Y.index])
# create table object
tabl = Table()
tabl.add_caption(
"Performance of forecast combinations of ECB SPF forecasts using the training window of the length: "
+ str(w))
tabl.append(NoEscape('\label{tab: spf_comb_perf_' + str(w) + '}'))
# create tabular object
tabr = Tabular(table_spec="c|l" + 6 * "|ccc")
tabr.add_hline()
tabr.add_hline()
# header row
tabr.add_row(
(MultiRow(3,
data="Class"), MultiRow(3,
data="Forecast Combination Method"),
MultiColumn(6, align='|c|',
data="RGDP"), MultiColumn(6, align='|c|', data="HICP"),
MultiColumn(6, align='|c', data="UNEM")))
tabr.add_hline(start=3, end=20, cmidruleoption="lr")
def build_latex(metrics: List[Metric], invert=False, caption=""):
models, metric_labels, splits = {}, {}, {}
score_map = defaultdict(lambda: "-")
for metric in metrics:
metric_vals = metric()
score, metric_label, model_name, split_name = (
metric_vals[SCORE],
metric_vals[METRIC_LABEL],
metric_vals[MODEL_NAME],
metric_vals[SPLIT_NAME],
)
metric_labels[metric_label] = 1
models[model_name] = 1
splits[split_name] = 1
score_map[(metric_label, model_name,
split_name)] = ("-" if math.isnan(score) else score)
models = list(models.keys())
splits = list(splits.keys())
metric_labels = list(metric_labels.keys())
if not invert:
model_split_cols = "".join(
["c|" for _ in range(len(models) * len(splits))])
table = Tabular("|c|" + model_split_cols)
table.add_hline()
row = (MultiRow(
2, data="Metric"), ) if len(splits) > 1 else ("Metric", )
for model_name in models:
row += (MultiColumn(len(splits), align="|c|",
data=model_name), )
table.add_row(row)
table.add_hline(1 if len(splits) == 1 else 2)
if (len(splits)) > 1:
row = ("", )
for _ in range(len(models)):
for split_name in splits:
row += (split_name, )
table.add_row(row)
table.add_hline()
for metric_label in metric_labels:
row = (metric_label, )
for model_name in models:
for split_name in splits:
row += (score_map[(metric_label, model_name,
split_name)], )
table.add_row(row)
table.add_hline()
else:
metric_split_labels = "".join(
["c|" for _ in range(len(metric_labels))])
table = Tabular("|c|" if len(splits) == 1 else "|c|c|" +
metric_split_labels)
table.add_hline()
row = (MultiColumn(2, align="|c|",
data=""), ) if len(splits) > 1 else ("", )
for metric_label in metric_labels:
row += (metric_label, )
table.add_row(row)
table.add_hline()
for model_name in models:
row = (MultiRow(len(splits), data=model_name), )
for split_name in splits:
row += (split_name, )
for metric_label in metric_labels:
row += (score_map[(metric_label, model_name,
split_name)], )
table.add_row(row)
if len(splits) > 1:
table.add_hline(start=2)
row = ("", )
table.add_hline()
latex_table = Table(position="h")
latex_table.append(table)
latex_table.add_caption(caption)
return latex_table
def gen_tex_table(tbl, cap, file_name, r):
"""
The function creates a tex file to which it export the given table from
python.
Parameters
----------
tbl : DataFrame
DataFrame containing the table to be exported.
cap : Str
Table caption.
file_name : Str
Name of the tex file in the "Tables" directory.
r : Int
Number of decimals to round up the metrics.
"""
# create tabule object
tabl = Table()
tabl.add_caption(cap)
tabl.append(NoEscape("\label{tab: " + cap + "}"))
# create tabular object
tabr = Tabular(table_spec="lccc")
tabr.add_hline()
tabr.add_hline()
# header row
tabr.add_row(["Forecast Combination Method"] + list(tbl))
tabr.add_hline()
# number of combination methods + additional rows
R = tbl.shape[0]
# specify format
fmt = "{:." + str(r) + "f}"
# fill in the rows for each combination method (-3 for individuals)
for i in range(R-3):
tabr.add_row([tbl.index[i]] + [
fmt.format(item) for item in tbl.iloc[i, :]])
tabr.add_hline()
# additional rows
for i in range(R-3, R):
tabr.add_row([tbl.index[i]] + [
fmt.format(item) for item in tbl.iloc[i, :]])
# end of table
tabr.add_hline()
tabr.add_hline()
# add tabular to table
tabl.append(tabr)
# export the table
tabl.generate_tex("C:/Users/Marek/Dropbox/Master_Thesis/Latex/Tables/" +
file_name)
return
ret_desc_stat = pd.DataFrame(data=np.full((9, 4), np.nan, dtype=np.float),
columns=tickers,
index=ds_stats)
ret_desc_stat.iloc[0, :] = np.mean(ret.values, axis=0)
ret_desc_stat.iloc[1, :] = np.median(ret.values, axis=0)
ret_desc_stat.iloc[2, :] = stats.mode(ret.values, axis=0)[0]
ret_desc_stat.iloc[3, :] = np.std(ret.values, axis=0)
ret_desc_stat.iloc[4, :] = np.var(ret.values, axis=0)
ret_desc_stat.iloc[5, :] = np.min(ret.values, axis=0)
ret_desc_stat.iloc[6, :] = np.max(ret.values, axis=0)
ret_desc_stat.iloc[7, :] = stats.kurtosis(ret.values, axis=0)
ret_desc_stat.iloc[8, :] = stats.skew(ret.values, axis=0)
# create table object
tabl = Table()
tabl.add_caption(
"Descriptive statistics of log-returns of U.S. Treasury futures")
tabl.append(NoEscape('\label{tab: Bond_ret_Desc_Stats}'))
# create tabular object
tabr = Tabular(table_spec="lcccc")
tabr.add_hline()
tabr.add_hline()
# header row
tabr.add_row(["Statistic"] + tickers)
tabr.add_hline()
# fill in the rows of tabular
for i in range(9):
tabr.add_row([ds_stats[i]] +
["{:.7f}".format(item) for item in ret_desc_stat.iloc[i, :]])
# end of table
tabr.add_hline()
tabr.add_hline()
np.transpose(X), y])[-2:]
# randomization inference p-values
p = np.mean(abs(beta_original) <= abs(beta_dist), axis=0)
# fill the output table
output_table.iloc[:, [4, 6]] = beta_original.reshape((3, 2))
output_table.iloc[:, [5, 7]] = p.reshape((3, 2))
########################
# Printing output table#
########################
# create tabule object
tabl = Table()
tabl.add_caption(
"Estimates of the effects of changing price rules on differents variables for both categories along with the p-values from the randomization inference."
)
tabl.append(NoEscape("\label{tab: output_table}"))
# create tabular object
tabr = Tabular(table_spec="l|cc|cc|cc|cc")
tabr.add_hline()
tabr.add_hline()
# header row
tabr.add_row((MultiRow(3, data="Dependent Variable"),
MultiColumn(4, align='|c|', data="Mobile Phones"),
MultiColumn(4, align='|c', data="Water Heaters")))
tabr.add_hline(start=2, end=9, cmidruleoption="lr")
tabr.add_row(
("", MultiColumn(2, align='|c|', data="Pilot 1"),
