forked from nf-core/variantbenchmarking
-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathmerge_reports.py
executable file
·274 lines (212 loc) · 10.5 KB
/
merge_reports.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
#!/usr/bin/env python
import pandas as pd
import glob
import re
import os
import sys
import errno
import argparse
def parse_args(args=None):
Description = "Merges svbenchmark or truvari bench reports from multiple samples"
Epilog = "Example usage: python merge_reports.py file1 file2 file3 -o merged_table.csv -b truvari/svbenchmark/wittyer/happy/sompy -v snv/indel -a germline/somatic "
parser = argparse.ArgumentParser(description=Description, epilog=Epilog)
parser.add_argument("inputs", nargs="+", help="List of files to merge")
parser.add_argument("--output", "-o", required=True, help="Output file")
parser.add_argument("--bench", "-b", required=True, help="svbenchmark/truvari/happy/sompy")
parser.add_argument("--vartype", "-v", required=True, help="Variant type: snv,indel,sv,small")
parser.add_argument("--analysis", "-a", required=True, help="Analysis type: germline,somatic")
return parser.parse_args(args)
## SVanalyzer results
def get_svbenchmark_resuls(file_paths):
# Initialize an empty DataFrame to store the merged data
merged_df = pd.DataFrame()
# Define regular expressions to extract the values
DTP_pattern = re.compile(r'Number of detected true variants \(.*\): (\d+)')
FN_pattern = re.compile(r'Number of undetected true variants \(.*\): (\d+)')
PTP_pattern = re.compile(r'Number of predictions that are true \(.*\): (\d+)')
FP_pattern = re.compile(r'Number of false positives \(.*\): (\d+)')
recall_pattern = re.compile(r'Recall \(.*\): (\d+\.\d+)%')
precision_pattern = re.compile(r'Precision \(.*\): (\d+\.\d+)%')
f1_pattern = re.compile(r'F1 \(.*\): ([\d\.]+(?:e[+-]?\d+)?)')
# Iterate over each table file
for file in file_paths:
# Read the table into a DataFrame
filename = os.path.basename(file)
with open(file, 'r') as f:
text = f.read()
# Search for matches in the text
DTP_match = DTP_pattern.search(text)
FN_match = FN_pattern.search(text)
PTP_match = PTP_pattern.search(text)
FP_match = FP_pattern.search(text)
recall_match = recall_pattern.search(text)
precision_match = precision_pattern.search(text)
f1_match = f1_pattern.search(text)
# Initialize a dictionary to store the data
data = {
'Tool': [filename.split(".")[0]],
'TP_base': [int(DTP_match.group(1)) if DTP_match else 'NA'],
'FP': [int(FP_match.group(1)) if FP_match else 'NA'],
'TP_comp': [int(DTP_match.group(1)) if DTP_match else 'NA'],
'FN': [int(FN_match.group(1)) if FN_match else 'NA'],
'Recall': [float(recall_match.group(1))/100 if recall_match else 'NA'],
'Precision': [float(precision_match.group(1))/100 if precision_match else 'NA'],
'F1': [float(f1_match.group(1)) if f1_match else 'NA']}
df = pd.DataFrame(data)
merged_df = pd.concat([merged_df, df], ignore_index=True)
return merged_df
## Truvari results
def get_truvari_resuls(file_paths):
# Initialize an empty DataFrame to store the merged data
merged_df = pd.DataFrame()
# Iterate over each table file
for file in file_paths:
# Read the json into a DataFrame
filename = os.path.basename(file)
with open(file, 'r') as f:
data = pd.read_json(f)
relevant_data = {
"Tool": filename.split(".")[0],
"TP_base": int(data["TP-base"].iloc[0]),
"TP_comp": int(data["TP-comp"].iloc[0]),
"FP": int(data["FP"].iloc[0]),
"FN": int(data["FN"].iloc[0]),
"Precision": float(data["precision"].iloc[0]),
"Recall": float(data["recall"].iloc[0]),
"F1": float(data["f1"].iloc[0])}
df = pd.DataFrame([relevant_data])
merged_df = pd.concat([merged_df, df], ignore_index=True)
return merged_df
def get_wittyer_resuls(file_paths):
# Initialize an empty DataFrame to store the merged data
merged_df = pd.DataFrame()
for file in file_paths:
# Read the json into a DataFrame
filename = os.path.basename(file)
with open(file, 'r') as f:
data = pd.read_json(f)
relevant_data = []
for sample in data['PerSampleStats']:
for stats in sample['OverallStats']:
relevant_data.append({
"Tool": filename.split(".")[0],
"StatsType": stats["StatsType"],
"TP_base": int(stats["TruthTpCount"]) if pd.notna(stats["TruthTpCount"]) else 0,
"TP_comp": int(stats["QueryTpCount"]) if pd.notna(stats["QueryTpCount"]) else 0,
"FP": int(stats["QueryFpCount"]) if pd.notna(stats["QueryFpCount"]) else 0,
"FN": int(stats["TruthFnCount"]) if pd.notna(stats["TruthFnCount"]) else 0,
"Precision": float(stats["Precision"]) if pd.notna(stats["Precision"]) else float('nan'),
"Recall": float(stats["Recall"]) if pd.notna(stats["Recall"]) else float('nan'),
"F1": float(stats["Fscore"]) if pd.notna(stats["Fscore"]) else float('nan')
})
df = pd.DataFrame(relevant_data)
merged_df = pd.concat([merged_df, df], ignore_index=True)
return merged_df
def get_rtgtools_resuls(file_paths):
# Initialize an empty DataFrame to store the merged data
merged_df = pd.DataFrame()
# Iterate over each table file
for file in file_paths:
filename = os.path.basename(file)
with open(file, 'r') as f:
lines = f.readlines()
# Extract header
header = lines[0].strip().split()
# Extract data
data = []
for line in lines[2:]:
data.append(line.strip().split())
# Create DataFrame
df = pd.DataFrame(data, columns=header)
df['Tool'] = filename.split(".")[0]
df_redesigned = df[['Tool', 'Threshold','True-pos-baseline','True-pos-call','False-pos','False-neg','Precision','Sensitivity','F-measure']]
df_redesigned.columns = ['Tool', 'Threshold','TP_base','TP_call','FP','FN','Precision','Recall','F1']
# Convert relevant columns to integers, handling potential NaN values
int_columns = ['TP_base', 'FN', 'TP_call', 'FP']
float_columns = ['Recall','Precision','F1']
df_redesigned[int_columns] = df_redesigned[int_columns].fillna(0).astype(int)
df_redesigned[float_columns] = df_redesigned[float_columns].fillna(0).astype(float)
merged_df = pd.concat([merged_df, df_redesigned], ignore_index=True)
return merged_df
def get_happy_resuls(file_paths):
# Initialize an empty DataFrame to store the merged data
merged_df = pd.DataFrame()
# Iterate over each table file
for file in file_paths:
filename = os.path.basename(file)
df = pd.read_csv(file)
df['Tool'] = filename.split(".")[0]
df_redesigned = df[['Tool', 'Type','Filter','TRUTH.TOTAL','TRUTH.TP','TRUTH.FN','QUERY.TOTAL','QUERY.FP','QUERY.UNK','FP.gt','FP.al','METRIC.Recall','METRIC.Precision','METRIC.Frac_NA','METRIC.F1_Score','TRUTH.TOTAL.TiTv_ratio','QUERY.TOTAL.TiTv_ratio','TRUTH.TOTAL.het_hom_ratio','QUERY.TOTAL.het_hom_ratio']]
df_redesigned.columns = ['Tool', 'Type','Filter','TP_base','TP','FN','TP_call','FP','UNK','FP_gt','FP_al','Recall','Precision','Frac_NA','F1','TRUTH_TiTv_ratio','QUERY_TiTv_ratio','TRUTH_het_hom_ratio','QUERY_het_hom_ratio']
# Convert relevant columns to integers, handling potential NaN values
int_columns = ['TP_base', 'TP', 'FN', 'TP_call', 'FP', 'UNK', 'FP_gt', 'FP_al']
float_columns = ['Recall','Precision','Frac_NA','F1','TRUTH_TiTv_ratio','QUERY_TiTv_ratio','TRUTH_het_hom_ratio','QUERY_het_hom_ratio']
df_redesigned[int_columns] = df_redesigned[int_columns].fillna(0).astype(int)
df_redesigned[float_columns] = df_redesigned[float_columns].fillna(0).astype(float)
# Concatenate with the merged DataFrame
merged_df = pd.concat([merged_df, df_redesigned], ignore_index=True)
return merged_df
def get_sompy_resuls(file_paths, vartype):
# Initialize an empty DataFrame to store the merged data
merged_df = pd.DataFrame()
# Iterate over each table file
for file in file_paths:
filename = os.path.basename(file)
df = pd.read_csv(file)
df['Tool'] = filename.split(".")[0]
df_redesigned = df[['Tool','type','total.truth','tp','fn','total.query','fp','unk','recall','precision','recall_lower','recall_upper','recall2','precision_lower','precision_upper','na','ambiguous','fp.region.size','fp.rate']]
df_redesigned.columns = ['Tool','Type','TP_base','TP','FN','TP_call','FP','UNK','Recall','Precision','recall_lower','recall_upper','recall2','precision_lower','precision_upper','na','ambiguous','fp.region.size','fp.rate']
# Convert relevant columns to integers, handling potential NaN values
int_columns = ['TP_base', 'TP', 'FN', 'TP_call', 'FP', 'UNK']
float_columns = ['Recall','Precision','recall_lower','recall_upper','recall2','precision_lower','precision_upper','na','ambiguous','fp.region.size','fp.rate']
df_redesigned[int_columns] = df_redesigned[int_columns].fillna(0).astype(int)
df_redesigned[float_columns] = df_redesigned[float_columns].fillna(0).astype(float)
merged_df = pd.concat([merged_df, df_redesigned], ignore_index=True)
if vartype == "snv":
merged_df1 = merged_df[merged_df["Type"] == 'SNVs']
elif vartype == "indel":
merged_df1 = merged_df[merged_df["Type"] == "indels"]
else:
merged_df1 = merged_df[merged_df["Type"] == "records"]
if vartype == "snv":
merged_df2 = merged_df[merged_df["Type"].str.contains(r'SNVs.')]
elif vartype == "indel":
merged_df2 = merged_df[merged_df["Type"].str.contains(r"indels.")]
else:
merged_df2 = merged_df[merged_df["Type"].str.contains(r"records.")]
return merged_df1,merged_df2
def main(args=None):
args = parse_args(args)
#check if the files are from svanalyzer or truvari
if args.analysis == "germline":
if args.bench == "truvari":
summ_table = get_truvari_resuls(args.inputs)
elif args.bench == "svbenchmark":
summ_table = get_svbenchmark_resuls(args.inputs)
elif args.bench == "wittyer":
summ_table = get_wittyer_resuls(args.inputs)
elif args.bench == "rtgtools":
summ_table = get_rtgtools_resuls(args.inputs)
elif args.bench == "happy":
summ_table = get_happy_resuls(args.inputs)
else:
raise ValueError('Only truvari/svbenchmark/wittyer/rtgtools/happy results can be merged for germline analysis!!')
summ_table.reset_index(drop=True, inplace=True)
summ_table.to_csv(args.output + ".summary.csv", index=False)
elif args.analysis == "somatic":
if args.bench == "sompy":
summ_table,summ_table2 = get_sompy_resuls(args.inputs,args.vartype)
summ_table2.reset_index(drop=True, inplace=True)
summ_table2.to_csv(args.output + ".regions.csv", index=False)
elif args.bench == "truvari":
summ_table = get_truvari_resuls(args.inputs)
elif args.bench == "svbenchmark":
summ_table = get_svbenchmark_resuls(args.inputs)
else:
raise ValueError('Only truvari/svbenchmark/sompy results can be merged for somatic analysis!!')
summ_table.reset_index(drop=True, inplace=True)
summ_table.to_csv(args.output + ".summary.csv", index=False)
else:
raise ValueError('Analysis must be germline or somatic')
if __name__ == "__main__":
sys.exit(main())