def db(
meta: Metadata, folder: str, rank: str = "genus", threads: int = 1
) -> JSONDirectory:
"""Create a model database from a set of SBML files."""
meta = meta.to_dataframe()
meta.columns = meta.columns.str.lower()
if not REQ_FIELDS.isin(meta.columns).all():
raise ValueError(
"Metadata File needs to have the following "
"columns %s." % ", ".join(REQ_FIELDS)
)
meta["id"] = meta.index
files = os.listdir(folder)
meta["file"] = meta.id + ".xml"
bad = meta.file.apply(lambda x: x not in files)
if any(bad):
raise ValueError(
"The following models are in the Metadata but not "
"in the folder: %s" % meta.file[bad]
)
meta = meta.groupby(rank).apply(reduce_group).reset_index(drop=True)
meta.index = meta[rank]
json_dir = JSONDirectory()
args = [
(tid, row, str(json_dir.json_files.path_maker(model_id=tid)), folder)
for tid, row in meta.iterrows()
]
workflow(_summarize_models, args, threads)
meta["file"] = meta.index + ".json"
meta["id"] = meta.index
meta["summary_rank"] = rank
meta.to_csv(json_dir.manifest.path_maker(), index=False)
return json_dir
def denoise_stats(
output_dir: str,
stats: qiime2.
Metadata, #stats type is not a metadata but this is the transformer used by DADA2 plugin to make DADA2Stats into pd.dataframe
plot_type: str = 'line',
labels: str = None,
style: str = 'whitegrid',
context: str = 'talk') -> None:
if not labels:
stats = load_df(stats)
else:
stats = load_df_labels(stats, labels)
stats = pd.concat(stats)
numeric = ['denoised', 'filtered', 'input', 'non-chimeric']
stats[numeric] = stats[numeric].apply(pd.to_numeric)
#makes into a df
stats = stats.groupby('id').sum()
stats = stats.drop(columns=[
'percentage of input passed filter', 'percentage of input non-chimeric'
])
df = pd.melt(stats.reset_index(),
id_vars='id',
var_name='step',
value_name='read_number')
input_read_number = df['read_number'].max()
df['% of Reads Remaining'] = df['read_number'] / input_read_number * 100
step_order = {'input': 0, 'filtered': 1, 'denoised': 2, 'non-chimeric': 3}
df['order'] = df['step'].apply(lambda x: step_order[x])
df = df.reset_index()
df['Run Number'] = 'Run ' + df['id']
hue_order = df.query('step == "non-chimeric"').sort_values(
'% of Reads Remaining', ascending=False)['id']
sns.set_style('whitegrid')
sns.set_context("talk")
line_graph = sns.lineplot(data=df,
y='% of Reads Remaining',
x='order',
hue='Run Number')
plt.ylim(0, 100)
plt.xlim(0, 4)
plt.xticks([x / 2 for x in range(0, 9)],
['Input', '', 'Filtered', '', 'Denoised', "", 'Non-chimeric'])
plt.xlabel('Processing Steps')
plt.legend(loc='center left', bbox_to_anchor=(1.25, 0.5), ncol=1)
# plt.title('allow to give any title or default one')
line_graph.figure.savefig(os.path.join(output_dir, 'line_graph.png'),
bbox_inches='tight')
line_graph.figure.savefig(os.path.join(output_dir, 'line_graph.pdf'),
bbox_inches='tight')
plt.gcf().clear()
#maybe the bargraph
r = range(len(stats))
vars_to_plot = df['id'].values
print(df.shape)
colors = ['darkorange', 'orange', 'sandybrown', 'navajowhite']
Step = ['Input', 'Filtered', 'Denoised', 'Non-chimeric']
plt.bar(r,
df[df['step'] == 'input']['read_number'],
color=colors[0],
edgecolor='white',
width=1)
plt.bar(r,
df[df['step'] == 'filtered']['read_number'],
color=colors[1],
edgecolor='white',
width=1)
plt.bar(r,
df[df['step'] == 'denoised']['read_number'],
color=colors[2],
edgecolor='white',
width=1)
#plt.bar(r, df[df['step']=='merged']['read_number'], color = colors[3], edgecolor = 'white', width = 1)
plt.bar(r,
df[df['step'] == 'non-chimeric']['read_number'],
color=colors[3],
edgecolor='white',
width=1)
plt.xticks(r, vars_to_plot, fontweight='bold')
plt.xlabel('Method')
plt.ylabel('Sequencing Depth')
plt.legend(Step, bbox_to_anchor=(1.05, 1), loc=2)
plt.savefig(os.path.join(output_dir, 'bar_plot.png'), bbox_inches='tight')
plt.savefig(os.path.join(output_dir, 'bar_plot.pdf'), bbox_inches='tight')
plt.gcf().clear()
index = os.path.join(TEMPLATES, 'denoise_assets', 'index.html')
q2templates.render(index, output_dir)