def update_hist(selection, zoom_range, n_click, url, selected_col):
#blank before selection
if (selection is None) or (selection == []):
return blank_fig()
#get triggered dictionary
trigger_list = dash.callback_context.triggered[0]["prop_id"]
#is button been triggered?
reset_df = True if 'reset-button.n_clicks' in trigger_list else False
df = import_dataset()
#list of key to check if 'auto' is in any key
is_auto_zoom = any(['auto' in x for x in zoom_range.keys()])
if (not is_auto_zoom) and (not reset_df):
name_range_list = ['x', 'y']
for i, col in enumerate(selected_col):
axis_string = f'{name_range_list[i]}axis'
df = df[(df[col] >= zoom_range[f'{axis_string}.range[0]'])
& (df[col] <= zoom_range[f'{axis_string}.range[1]'])]
df[selection] = df[selection].astype(str)
fig = get_histogram(df, selection)
return fig
def update_scatter(selection, clicked, n_click, url):
#blank before selection
if (selection is None) or (selection == []):
return blank_fig()
#get triggered dictionary
trigger_list = dash.callback_context.triggered[0]["prop_id"]
#is button been triggered?
reset_df = True if 'reset-button.n_clicks' in trigger_list else False
df = import_dataset()
#filter data index based on clicked hist
if (clicked is not None) and (not reset_df):
clicked_index = clicked['points'][0]['pointNumbers']
df = df[df.index.isin(clicked_index)]
#assign x
x_selection = selection[0]
#return histogram
if (len(selection) == 1):
fig = get_histogram(df, x_selection)
else:
y_selection = selection[1]
fig = get_scatter(df, x_selection, y_selection)
return fig
def update_dataset(url):
df = import_dataset()
#need to import something
if df is None:
return [None, None]
columns = [{"name": i, "id": i} for i in df.columns]
return [columns, df.to_dict('records')]
def update_info_rows(rows, url):
df = import_dataset()
#need to import something
if df is None:
return [None]
dff = df if rows is None else pd.DataFrame(rows)
num_rows, num_cols = dff.shape
return [html.P(f'The dataset has {num_rows} rows and {num_cols} columns.')]
def update_table(rows, url):
_, metadata_type = import_metadata(get_type_colname=True)
df = import_dataset()
#need to import something
if (df is None) or (metadata_type is None):
return blank_fig()
dff = df if rows is None else pd.DataFrame(rows)
numeric_col = [x for x in metadata_type['numeric'] if x in dff.columns]
numeric_info = [[
col,
np.round(dff[col].mean(), 2),
np.round(dff[col].std(), 2),
np.round(dff[col].median(), 2),
np.round(dff[col].min(), 2),
np.round(dff[col].max(), 2),
] for col in numeric_col]
#reshape
numeric_info = np.array(numeric_info).T.tolist()
fig_table = go.Figure(data=[
go.Table(header=dict(
values=['Columns', 'Mean', 'Std', 'Median', 'Min', 'Max'],
align=['left', 'center']),
cells=dict(values=numeric_info, align=['left', 'center']))
],
layout=go.Layout(template='plotly_white',
margin={'l': 0}))
return fig_table
def update_graphs(string_col, rows, url):
if string_col is None:
return blank_fig()
df = import_dataset()
dff = df if rows is None else pd.DataFrame(rows)
# if you filter and have empty dataframe
if (string_col not in dff.columns):
return blank_fig()
#coherce to string... integer will be displayed as str
dff_string_unique = [str(x) for x in dff[string_col].unique().tolist()]
dff_string_count = [(dff[string_col] == cat).sum()
for cat in dff[string_col].unique()]
fig = go.Figure(data=go.Bar(x=dff_string_unique, y=dff_string_count),
layout=go.Layout(template='plotly_white',
title=string_col,
margin={'l': 0}))
return fig