def __repr__(self):
''' Return a string representation for a particular DataFrame '''
buf = six.StringIO('')
# Calling a private DataFrame method here, so protect it.
if getattr(self, '_info_repr', lambda: False)():
self.info(buf=buf)
return buf.getvalue()
if self.label:
buf.write('%s\n\n' % self.label)
kwargs = {}
try:
kwargs['min_rows'] = pd.get_option('display.min_rows')
except Exception: # noqa: E722
pass
max_rows = pd.get_option('display.max_rows')
max_cols = pd.get_option('display.max_columns')
show_dimensions = pd.get_option('display.show_dimensions')
if pd.get_option('display.expand_frame_repr'):
width, _ = get_console_size()
else:
width = None
if not hasattr(pdfmt, 'DataFrameRenderer'):
kwargs['line_width'] = width
if get_option('display.apply_formats'):
kwargs['formatters'] = self._get_formatters()
if 'na_rep' not in kwargs:
kwargs['na_rep'] = '.'
formatter = pdfmt.DataFrameFormatter(self,
max_rows=max_rows,
max_cols=max_cols,
show_dimensions=show_dimensions,
**kwargs)
# NOTE: Patch for bug in pandas DataFrameFormatter when using
# formatters on a DataFrame that is truncated in the console.
formatter.columns = formatter.tr_frame.columns
# pandas 1.2.0 uses a renderer instead of just a formatter
if hasattr(pdfmt, 'DataFrameRenderer'):
formatter = pdfmt.DataFrameRenderer(formatter)
txt = formatter.to_string(line_width=width)
else:
txt = formatter.to_string()
if txt is None:
if getattr(formatter, 'buf', None) is not None:
buf.write(formatter.buf.getvalue())
else:
buf.write(txt)
return buf.getvalue()
def patchPandasHTMLrepr(self, **kwargs):
'''
Patched default escaping of HTML control characters to allow molecule image rendering dataframes
'''
formatter = fmt.DataFrameFormatter(
self, buf=None, columns=None, col_space=None, colSpace=None, header=True, index=True,
na_rep='NaN', formatters=None, float_format=None, sparsify=None, index_names=True, justify=None,
force_unicode=None, bold_rows=True, classes=None, escape=False)
formatter.to_html()
html = formatter.buf.getvalue()
return html
def to_string(self, apply_formats=None, **kwargs):
'''
Return a string representation of a DataFrame
Parameters
----------
apply_formats : bool or None, optional
Should SAS formats be applied to the data values in the
rendered output? If None, the `display.apply_formats`
option value will be used.
**kwargs : keyword-parameters, optional
All keyword parameters for the `pandas.DataFrame.to_string`
method are accepted here as well.
See Also
--------
:meth:`pandas.DataFrame.to_string`
Returns
-------
string
'''
buf = six.StringIO('')
if self.label:
buf.write('%s\n\n' % self.label)
formatters = kwargs.get('formatters', None)
if apply_formats or (apply_formats is None
and get_option('display.apply_formats')):
kwargs['formatters'] = self._get_formatters(formatters)
if 'na_rep' not in kwargs:
kwargs['na_rep'] = '.'
formatter = pdfmt.DataFrameFormatter(self, **kwargs)
# NOTE: Patch for bug in pandas DataFrameFormatter when using
# formatters on a DataFrame that is truncated in the console.
formatter.columns = formatter.tr_frame.columns
# pandas 1.2.0 uses a renderer instead of just a formatter
if hasattr(pdfmt, 'DataFrameRenderer'):
formatter = pdfmt.DataFrameRenderer(formatter)
txt = formatter.to_string()
if txt is None:
if getattr(formatter, 'buf', None) is not None:
buf.write(formatter.buf.getvalue())
else:
buf.write(txt)
return buf.getvalue()
def _repr_html_(self):
''' Return a html representation for a particular DataFrame '''
# Calling a private DataFrame method here, so protect it.
if getattr(self, '_info_repr', lambda: False)():
buf = six.StringIO('')
self.info(buf=buf)
# need to escape the <class>, should be the first line.
val = buf.getvalue().replace('<', r'<', 1)
val = val.replace('>', r'>', 1)
return '<pre>' + val + '</pre>'
kwargs = {}
if get_option('display.apply_formats'):
kwargs['formatters'] = self._get_formatters()
kwargs['na_rep'] = '.'
if pd.get_option('display.notebook_repr_html'):
try:
kwargs['min_rows'] = pd.get_option('display.min_rows')
except: # noqa: E722
pass
max_rows = pd.get_option('display.max_rows')
max_cols = pd.get_option('display.max_columns')
show_dimensions = pd.get_option('display.show_dimensions')
formatter = pdfmt.DataFrameFormatter(
self,
max_rows=max_rows,
max_cols=max_cols,
show_dimensions=show_dimensions,
**kwargs)
# NOTE: Patch for bug in pandas DataFrameFormatter when using
# formatters on a DataFrame that is truncated in the console.
formatter.columns = formatter.tr_frame.columns
# pandas 1.2.0 uses a renderer instead of just a formatter
if hasattr(pdfmt, 'DataFrameRenderer'):
formatter = pdfmt.DataFrameRenderer(formatter)
html = formatter.to_html(**notebook_opts)
if html is None:
if getattr(formatter, 'buf', None) is not None:
html = formatter.buf.getvalue()
else:
return None
return self._post_process_html(html)
return None
# out_topics_html_dir = outdir + f'topics-central-docs-abstracts-{datafile_date}-html/'
out_topics_html_dir = web_out_dir
os.makedirs(out_topics_html_dir, mode=out_path_mode, exist_ok=True)
for c in predominant_doc_dfd.keys():
for n in predominant_doc_dfd[c].keys():
ofdir = out_topics_html_dir + f'{c}-{n}/'
os.makedirs(ofdir, mode=out_path_mode, exist_ok=True)
print(ofdir)
for i in predominant_doc_dfd[c][n].keys():
ofname = ofdir + f'Topic_{i+1:02d}.html'
with open(ofname, 'w') as ofp:
html_df = (predominant_doc_dfd[c][n][i].drop(columns=[
'sha', 'major topics', 'abstract clean',
'predominant topic', 'predominant topic num'
] + [' '.join(c.split('_'))
for c in sims_columns]).copy().set_index(
np.arange(1,
len(predominant_doc_dfd[c][n][i]) + 1)))
# html_table = html_df.to_html(escape=False)
df_formatter = fmt.DataFrameFormatter(escape=False,
frame=html_df,
index=True,
bold_rows=True)
html_formatter = MyHTMLFormatter('cord uid',
formatter=df_formatter)
# html_formatter = HTMLFormatter(formatter=df_formatter)
html_table = html_formatter.get_result()
html_str = html_template.format(f'Topic {i+1:02d}', html_style,
html_table)
ofp.write(html_str)