Mercurial > repos > bgruening > flexynesis_cbioportal_import

annotate flexynesis_utils.py @ 8:461b8e2d79d5 draft default tip

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit bcee55468e70cd025cc5359bef5e9991e893eca9
author bgruening
date Thu, 16 Oct 2025 20:15:31 +0000
parents 693011647a67
children
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
2
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
1 #!/usr/bin/env python
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
2
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
3 import argparse
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
4 import os
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
5 import sys
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
6 from pathlib import Path
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
7
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
8 import pandas as pd
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
9
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
10
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
11 def read_data(data_input, index=False):
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
12 """Load CSV or TSV data file."""
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
13 try:
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
14 file_ext = Path(data_input).suffix.lower()
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
15 sep = ',' if file_ext == '.csv' else '\t'
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
16 index_col = 0 if index else None
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
17
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
18 if file_ext in ['.csv', '.tsv', '.txt', '.tab', '.tabular']:
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
19 return pd.read_csv(data_input, sep=sep, index_col=index_col)
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
20 else:
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
21 raise ValueError(f"Unsupported file extension: {file_ext}")
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
22 except Exception as e:
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
23 raise ValueError(f"Error loading data from {data_input}: {e}") from e
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
24
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
25
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
26 def binarize_mutations(df, gene_idx=1, sample_idx=2):
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
27 """
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
28 Binarize mutation data by creating a matrix of gene x sample with 1/0 values.
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
29 """
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
30 # galaxy index is 1-based, convert to zero-based
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
31 gene_idx -= 1
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
32 sample_idx -= 1
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
33 # check idx
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
34 if gene_idx >= len(df.columns) or sample_idx >= len(df.columns):
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
35 raise ValueError(f"Column indices out of bounds. DataFrame has {len(df.columns)} columns, "
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
36 f"but requested indices are {gene_idx} and {sample_idx}")
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
37 if gene_idx == sample_idx:
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
38 raise ValueError("Gene and sample column indices must be different")
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
39
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
40 # Get column names by index
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
41 gene_col = df.columns[gene_idx]
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
42 print(f"Using gene column: {gene_col} (index {gene_idx})")
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
43 sample_col = df.columns[sample_idx]
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
44 print(f"Using sample column: {sample_col} (index {sample_idx})")
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
45
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
46 # Check if columns contain data
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
47 if df[gene_col].isna().all():
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
48 raise ValueError(f"Gene column (index {gene_idx}) contains only NaN values.")
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
49 if df[sample_col].isna().all():
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
50 raise ValueError(f"Sample column (index {sample_idx}) contains only NaN values.")
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
51
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
52 # Group by gene and sample, count mutations
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
53 mutation_counts = df.groupby([gene_col, sample_col]).size().reset_index(name='count')
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
54
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
55 # Create pivot table
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
56 mutation_matrix = mutation_counts.pivot(index=gene_col, columns=sample_col, values='count').fillna(0)
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
57
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
58 # Binarize: convert any count > 0 to 1
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
59 mutation_matrix[mutation_matrix > 0] = 1
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
60
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
61 return mutation_matrix
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
62
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
63
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
64 def make_data_dict(clin_path, omics_paths):
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
65 """Read clinical and omics data files into a dictionary."""
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
66 data = {}
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
67
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
68 # Read clinical data
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
69 print(f"Reading clinical data from {clin_path}")
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
70 try:
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
71 clin = read_data(clin_path, index=True)
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
72
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
73 if clin.empty:
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
74 raise ValueError(f"Clinical file {clin_path} is empty")
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
75 data['clin'] = clin
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
76 print(f"Loaded clinical data: {clin.shape[0]} samples, {clin.shape[1]} features")
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
77 except Exception as e:
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
78 raise ValueError(f"Error reading clinical file {clin_path}: {e}")
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
79
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
80 # Read omics data
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
81 print(f"Reading omics data from {', '.join(omics_paths)}")
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
82 for path in omics_paths:
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
83 try:
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
84 name = os.path.splitext(os.path.basename(path))[0]
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
85 df = read_data(path, index=True)
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
86 if df.empty:
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
87 print(f"Warning: Omics file {path} is empty, skipping")
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
88 continue
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
89 data[name] = df
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
90 print(f"Loaded {name}: {df.shape[0]} features, {df.shape[1]} samples")
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
91 except Exception as e:
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
92 print(f"Warning: Error reading omics file {path}: {e}")
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
93 continue
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
94
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
95 if len(data) == 1: # Only clinical data loaded
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
96 raise ValueError("No omics data was successfully loaded")
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
97
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
98 return data
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
99
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
100
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
101 def validate_data_consistency(data):
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
102 """Validate that clinical and omics data have consistent samples."""
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
103 clin_samples = set(data['clin'].index)
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
104
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
105 for name, df in data.items():
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
106 if name == 'clin':
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
107 continue
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
108
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
109 omics_samples = set(df.columns)
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
110
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
111 # Check for sample overlap
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
112 common_samples = clin_samples.intersection(omics_samples)
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
113 if len(common_samples) == 0:
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
114 raise ValueError(f"No common samples between clinical data and {name}")
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
115
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
116 missing_in_omics = clin_samples - omics_samples
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
117 missing_in_clin = omics_samples - clin_samples
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
118
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
119 if missing_in_omics:
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
120 print(f"Warning: {len(missing_in_omics)} clinical samples not found in {name}")
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
121 if missing_in_clin:
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
122 print(f"Warning: {len(missing_in_clin)} samples in {name} not found in clinical data")
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
123
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
124 return True
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
125
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
126
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
127 def split_and_save_data(data, ratio=0.7, output_dir='.'):
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
128 """Split data into train/test sets and save to files."""
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
129 # Validate data consistency first
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
130 validate_data_consistency(data)
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
131
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
132 samples = data['clin'].index.tolist()
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
133
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
134 train_samples = list(pd.Series(samples).sample(frac=ratio, random_state=42))
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
135 test_samples = list(set(samples) - set(train_samples))
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
136
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
137 train_data = {}
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
138 test_data = {}
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
139
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
140 for key, df in data.items():
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
141 try:
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
142 if key == 'clin':
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
143 train_data[key] = df.loc[df.index.intersection(train_samples)]
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
144 test_data[key] = df.loc[df.index.intersection(test_samples)]
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
145 else:
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
146 train_data[key] = df.loc[:, df.columns.intersection(train_samples)]
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
147 test_data[key] = df.loc[:, df.columns.intersection(test_samples)]
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
148 except Exception as e:
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
149 print(f"Error splitting data {key}: {e}")
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
150 continue
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
151
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
152 # Create output directories
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
153 os.makedirs(os.path.join(output_dir, 'train'), exist_ok=True)
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
154 os.makedirs(os.path.join(output_dir, 'test'), exist_ok=True)
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
155
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
156 # Save train and test data
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
157 for key in data.keys():
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
158 try:
6
3f7481deaff1 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit 1afbaf45449e25238935e222f983da62392c067a
bgruening
parents: 2
diff changeset
159 train_data[key].to_csv(os.path.join(output_dir, 'train', f'{key}.tabular'), sep='\t')
3f7481deaff1 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit 1afbaf45449e25238935e222f983da62392c067a
bgruening
parents: 2
diff changeset
160 test_data[key].to_csv(os.path.join(output_dir, 'test', f'{key}.tabular'), sep='\t')
2
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
161 except Exception as e:
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
162 print(f"Error saving {key}: {e}")
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
163 continue
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
164
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
165
7
693011647a67 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit 6b520305ec30e6dc37eba92c67a5368cea0fc5ad
bgruening
parents: 6
diff changeset
166 def validate_numeric_column(df, column_names, require_integer=False):
693011647a67 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit 6b520305ec30e6dc37eba92c67a5368cea0fc5ad
bgruening
parents: 6
diff changeset
167 """ Validate that a column(s) in the DataFrame contains numeric values. """
693011647a67 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit 6b520305ec30e6dc37eba92c67a5368cea0fc5ad
bgruening
parents: 6
diff changeset
168 if isinstance(column_names, str):
693011647a67 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit 6b520305ec30e6dc37eba92c67a5368cea0fc5ad
bgruening
parents: 6
diff changeset
169 # Handle comma-separated string: "col1,col2,col3"
693011647a67 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit 6b520305ec30e6dc37eba92c67a5368cea0fc5ad
bgruening
parents: 6
diff changeset
170 if ',' in column_names:
693011647a67 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit 6b520305ec30e6dc37eba92c67a5368cea0fc5ad
bgruening
parents: 6
diff changeset
171 column_names = [col.strip() for col in column_names.split(',')]
693011647a67 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit 6b520305ec30e6dc37eba92c67a5368cea0fc5ad
bgruening
parents: 6
diff changeset
172 else:
693011647a67 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit 6b520305ec30e6dc37eba92c67a5368cea0fc5ad
bgruening
parents: 6
diff changeset
173 # Single column name
693011647a67 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit 6b520305ec30e6dc37eba92c67a5368cea0fc5ad
bgruening
parents: 6
diff changeset
174 column_names = [column_names]
693011647a67 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit 6b520305ec30e6dc37eba92c67a5368cea0fc5ad
bgruening
parents: 6
diff changeset
175
693011647a67 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit 6b520305ec30e6dc37eba92c67a5368cea0fc5ad
bgruening
parents: 6
diff changeset
176 # Validate each column
693011647a67 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit 6b520305ec30e6dc37eba92c67a5368cea0fc5ad
bgruening
parents: 6
diff changeset
177 for column_name in column_names:
693011647a67 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit 6b520305ec30e6dc37eba92c67a5368cea0fc5ad
bgruening
parents: 6
diff changeset
178 if column_name not in df.columns:
693011647a67 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit 6b520305ec30e6dc37eba92c67a5368cea0fc5ad
bgruening
parents: 6
diff changeset
179 raise ValueError(f"Column '{column_name}' not found in DataFrame.")
693011647a67 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit 6b520305ec30e6dc37eba92c67a5368cea0fc5ad
bgruening
parents: 6
diff changeset
180
693011647a67 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit 6b520305ec30e6dc37eba92c67a5368cea0fc5ad
bgruening
parents: 6
diff changeset
181 try:
693011647a67 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit 6b520305ec30e6dc37eba92c67a5368cea0fc5ad
bgruening
parents: 6
diff changeset
182 numeric_col = pd.to_numeric(df[column_name], errors='raise')
693011647a67 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit 6b520305ec30e6dc37eba92c67a5368cea0fc5ad
bgruening
parents: 6
diff changeset
183 except Exception as e:
693011647a67 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit 6b520305ec30e6dc37eba92c67a5368cea0fc5ad
bgruening
parents: 6
diff changeset
184 raise ValueError(f"Non-numeric values found in column '{column_name}': {e}")
693011647a67 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit 6b520305ec30e6dc37eba92c67a5368cea0fc5ad
bgruening
parents: 6
diff changeset
185
693011647a67 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit 6b520305ec30e6dc37eba92c67a5368cea0fc5ad
bgruening
parents: 6
diff changeset
186 if require_integer:
693011647a67 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit 6b520305ec30e6dc37eba92c67a5368cea0fc5ad
bgruening
parents: 6
diff changeset
187 # Check if all non-null values are equivalent to integers
693011647a67 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit 6b520305ec30e6dc37eba92c67a5368cea0fc5ad
bgruening
parents: 6
diff changeset
188 non_null_values = numeric_col.dropna()
693011647a67 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit 6b520305ec30e6dc37eba92c67a5368cea0fc5ad
bgruening
parents: 6
diff changeset
189 if not (non_null_values == non_null_values.round()).all():
693011647a67 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit 6b520305ec30e6dc37eba92c67a5368cea0fc5ad
bgruening
parents: 6
diff changeset
190 raise ValueError(f"Column '{column_name}' contains non-integer numeric values.")
693011647a67 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit 6b520305ec30e6dc37eba92c67a5368cea0fc5ad
bgruening
parents: 6
diff changeset
191 print(f"Column '{column_name}': All values are integers or integer-equivalent floats.")
693011647a67 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit 6b520305ec30e6dc37eba92c67a5368cea0fc5ad
bgruening
parents: 6
diff changeset
192 else:
693011647a67 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit 6b520305ec30e6dc37eba92c67a5368cea0fc5ad
bgruening
parents: 6
diff changeset
193 print(f"Column '{column_name}': All values are numeric (integers and floats accepted).")
693011647a67 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit 6b520305ec30e6dc37eba92c67a5368cea0fc5ad
bgruening
parents: 6
diff changeset
194
693011647a67 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit 6b520305ec30e6dc37eba92c67a5368cea0fc5ad
bgruening
parents: 6
diff changeset
195
693011647a67 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit 6b520305ec30e6dc37eba92c67a5368cea0fc5ad
bgruening
parents: 6
diff changeset
196 def validate_survival(df, column_names):
693011647a67 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit 6b520305ec30e6dc37eba92c67a5368cea0fc5ad
bgruening
parents: 6
diff changeset
197 """Validate survival column(s) (integer)."""
693011647a67 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit 6b520305ec30e6dc37eba92c67a5368cea0fc5ad
bgruening
parents: 6
diff changeset
198 validate_numeric_column(df, column_names, require_integer=True)
693011647a67 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit 6b520305ec30e6dc37eba92c67a5368cea0fc5ad
bgruening
parents: 6
diff changeset
199
693011647a67 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit 6b520305ec30e6dc37eba92c67a5368cea0fc5ad
bgruening
parents: 6
diff changeset
200
693011647a67 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit 6b520305ec30e6dc37eba92c67a5368cea0fc5ad
bgruening
parents: 6
diff changeset
201 def validate_covariate(df, column_names):
693011647a67 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit 6b520305ec30e6dc37eba92c67a5368cea0fc5ad
bgruening
parents: 6
diff changeset
202 """Validate covariate column(s) (numeric)."""
693011647a67 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit 6b520305ec30e6dc37eba92c67a5368cea0fc5ad
bgruening
parents: 6
diff changeset
203 validate_numeric_column(df, column_names, require_integer=False)
693011647a67 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit 6b520305ec30e6dc37eba92c67a5368cea0fc5ad
bgruening
parents: 6
diff changeset
204
693011647a67 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit 6b520305ec30e6dc37eba92c67a5368cea0fc5ad
bgruening
parents: 6
diff changeset
205
2
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
206 def main():
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
207 parser = argparse.ArgumentParser(description='Flexynesis extra utilities')
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
208
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
209 parser.add_argument("--util", type=str, required=True,
7
693011647a67 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit 6b520305ec30e6dc37eba92c67a5368cea0fc5ad
bgruening
parents: 6
diff changeset
210 choices=['split', 'binarize', 'validate_survival', 'validate_covariate'],
693011647a67 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit 6b520305ec30e6dc37eba92c67a5368cea0fc5ad
bgruening
parents: 6
diff changeset
211 help="Utility function: 'split' for spiting data to train and test, 'binarize' for creating a binarized matrix from a mutation data, 'validate_survival' for validating survival data.")
2
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
212
7
693011647a67 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit 6b520305ec30e6dc37eba92c67a5368cea0fc5ad
bgruening
parents: 6
diff changeset
213 # Arguments for split (clin also for validate_survival and validate_covariate)
2
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
214 parser.add_argument('--clin', required=False,
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
215 help='Path to clinical data CSV file (samples in rows)')
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
216 parser.add_argument('--omics', required=False,
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
217 help='Comma-separated list of omics CSV files (samples in columns)')
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
218 parser.add_argument('--split', type=float, default=0.7,
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
219 help='Train split ratio (default: 0.7)')
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
220
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
221 # Arguments for binarize
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
222 parser.add_argument('--mutations', type=str, required=False,
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
223 help='Path to mutation data CSV file (samples in rows, genes in columns)')
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
224 parser.add_argument('--gene_idx', type=int, default=0,
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
225 help='Column index for genes in mutation data (default: 0)')
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
226 parser.add_argument('--sample_idx', type=int, default=1,
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
227 help='Column index for samples in mutation data (default: 1)')
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
228
7
693011647a67 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit 6b520305ec30e6dc37eba92c67a5368cea0fc5ad
bgruening
parents: 6
diff changeset
229 # Arguments for validate_survival and validate_covariate
693011647a67 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit 6b520305ec30e6dc37eba92c67a5368cea0fc5ad
bgruening
parents: 6
diff changeset
230 parser.add_argument('--clin_variable', type=str, required=False,
693011647a67 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit 6b520305ec30e6dc37eba92c67a5368cea0fc5ad
bgruening
parents: 6
diff changeset
231 help='Column name for clinical variable (e.g., death, SEX, ...)')
693011647a67 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit 6b520305ec30e6dc37eba92c67a5368cea0fc5ad
bgruening
parents: 6
diff changeset
232
693011647a67 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit 6b520305ec30e6dc37eba92c67a5368cea0fc5ad
bgruening
parents: 6
diff changeset
233 # common arguments (binarize and split)
2
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
234 parser.add_argument('--out', default='.',
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
235 help='Output directory (default: current directory)')
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
236
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
237 args = parser.parse_args()
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
238
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
239 try:
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
240 # validate utility function
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
241 if not args.util:
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
242 raise ValueError("Utility function must be specified")
7
693011647a67 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit 6b520305ec30e6dc37eba92c67a5368cea0fc5ad
bgruening
parents: 6
diff changeset
243 if args.util not in ['split', 'binarize', 'validate_survival', 'validate_covariate']:
2
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
244 raise ValueError(f"Invalid utility function: {args.util}")
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
245
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
246 if args.util == 'split':
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
247 # Validate inputs
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
248 if not args.clin:
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
249 raise ValueError("Clinical data file must be provided")
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
250 if not args.omics:
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
251 raise ValueError("At least one omics file must be provided")
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
252 if not os.path.isfile(args.clin):
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
253 raise FileNotFoundError(f"Clinical file not found: {args.clin}")
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
254 # Validate split ratio
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
255 if not 0 < args.split < 1:
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
256 raise ValueError(f"Split ratio must be between 0 and 1, got {args.split}")
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
257
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
258 elif args.util == 'binarize':
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
259 # Validate mutation data file
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
260 if not args.mutations:
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
261 raise ValueError("Mutation data file must be provided")
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
262 if not os.path.isfile(args.mutations):
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
263 raise FileNotFoundError(f"Mutation data file not found: {args.mutations}")
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
264 # Validate gene and sample indices
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
265 if args.gene_idx < 0 or args.sample_idx < 0:
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
266 raise ValueError("Gene and sample indices must be non-negative integers")
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
267
7
693011647a67 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit 6b520305ec30e6dc37eba92c67a5368cea0fc5ad
bgruening
parents: 6
diff changeset
268 elif args.util == 'validate_survival' or args.util == 'validate_covariate':
693011647a67 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit 6b520305ec30e6dc37eba92c67a5368cea0fc5ad
bgruening
parents: 6
diff changeset
269 # Validate clinical data file
693011647a67 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit 6b520305ec30e6dc37eba92c67a5368cea0fc5ad
bgruening
parents: 6
diff changeset
270 if not args.clin:
693011647a67 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit 6b520305ec30e6dc37eba92c67a5368cea0fc5ad
bgruening
parents: 6
diff changeset
271 raise ValueError("Clinical data file must be provided")
693011647a67 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit 6b520305ec30e6dc37eba92c67a5368cea0fc5ad
bgruening
parents: 6
diff changeset
272 if not os.path.isfile(args.clin):
693011647a67 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit 6b520305ec30e6dc37eba92c67a5368cea0fc5ad
bgruening
parents: 6
diff changeset
273 raise FileNotFoundError(f"Clinical file not found: {args.clin}")
693011647a67 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit 6b520305ec30e6dc37eba92c67a5368cea0fc5ad
bgruening
parents: 6
diff changeset
274 # Validate survival event variable
693011647a67 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit 6b520305ec30e6dc37eba92c67a5368cea0fc5ad
bgruening
parents: 6
diff changeset
275 if not args.clin_variable:
693011647a67 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit 6b520305ec30e6dc37eba92c67a5368cea0fc5ad
bgruening
parents: 6
diff changeset
276 raise ValueError("Survival event variable must be specified")
693011647a67 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit 6b520305ec30e6dc37eba92c67a5368cea0fc5ad
bgruening
parents: 6
diff changeset
277
2
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
278 # Create output directory if it doesn't exist
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
279 if not os.path.exists(args.out):
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
280 os.makedirs(args.out)
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
281
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
282 if args.util == 'split':
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
283 # Parse omics files
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
284 omics_files = [f.strip() for f in args.omics.split(',') if f.strip()]
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
285 if not omics_files:
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
286 raise ValueError("At least one omics file must be provided")
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
287 # Check omics files exist
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
288 for f in omics_files:
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
289 if not os.path.isfile(f):
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
290 raise FileNotFoundError(f"Omics file not found: {f}")
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
291 data = make_data_dict(args.clin, omics_files)
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
292 split_and_save_data(data, ratio=args.split, output_dir=args.out)
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
293
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
294 elif args.util == 'binarize':
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
295 mutations_df = read_data(args.mutations, index=False)
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
296 if mutations_df.empty:
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
297 raise ValueError("Mutation data file is empty")
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
298
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
299 binarized_matrix = binarize_mutations(mutations_df, gene_idx=args.gene_idx, sample_idx=args.sample_idx)
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
300 # Save binarized matrix
6
3f7481deaff1 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit 1afbaf45449e25238935e222f983da62392c067a
bgruening
parents: 2
diff changeset
301 output_file = os.path.join(args.out, 'binarized_mutations.tabular')
3f7481deaff1 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit 1afbaf45449e25238935e222f983da62392c067a
bgruening
parents: 2
diff changeset
302 binarized_matrix.to_csv(output_file, sep='\t')
2
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
303 print(f"Binarized mutation matrix saved to {output_file}")
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
304
7
693011647a67 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit 6b520305ec30e6dc37eba92c67a5368cea0fc5ad
bgruening
parents: 6
diff changeset
305 elif args.util == 'validate_survival':
693011647a67 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit 6b520305ec30e6dc37eba92c67a5368cea0fc5ad
bgruening
parents: 6
diff changeset
306 clin_df = read_data(args.clin, index=False)
693011647a67 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit 6b520305ec30e6dc37eba92c67a5368cea0fc5ad
bgruening
parents: 6
diff changeset
307 if clin_df.empty:
693011647a67 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit 6b520305ec30e6dc37eba92c67a5368cea0fc5ad
bgruening
parents: 6
diff changeset
308 raise ValueError("Clinical data file is empty")
693011647a67 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit 6b520305ec30e6dc37eba92c67a5368cea0fc5ad
bgruening
parents: 6
diff changeset
309
693011647a67 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit 6b520305ec30e6dc37eba92c67a5368cea0fc5ad
bgruening
parents: 6
diff changeset
310 # Validate survival event variable
693011647a67 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit 6b520305ec30e6dc37eba92c67a5368cea0fc5ad
bgruening
parents: 6
diff changeset
311 validate_survival(clin_df, args.clin_variable)
693011647a67 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit 6b520305ec30e6dc37eba92c67a5368cea0fc5ad
bgruening
parents: 6
diff changeset
312
693011647a67 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit 6b520305ec30e6dc37eba92c67a5368cea0fc5ad
bgruening
parents: 6
diff changeset
313 elif args.util == 'validate_covariate':
693011647a67 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit 6b520305ec30e6dc37eba92c67a5368cea0fc5ad
bgruening
parents: 6
diff changeset
314 clin_df = read_data(args.clin, index=False)
693011647a67 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit 6b520305ec30e6dc37eba92c67a5368cea0fc5ad
bgruening
parents: 6
diff changeset
315 if clin_df.empty:
693011647a67 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit 6b520305ec30e6dc37eba92c67a5368cea0fc5ad
bgruening
parents: 6
diff changeset
316 raise ValueError("Clinical data file is empty")
693011647a67 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit 6b520305ec30e6dc37eba92c67a5368cea0fc5ad
bgruening
parents: 6
diff changeset
317
693011647a67 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit 6b520305ec30e6dc37eba92c67a5368cea0fc5ad
bgruening
parents: 6
diff changeset
318 # Validate clinical variable
693011647a67 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit 6b520305ec30e6dc37eba92c67a5368cea0fc5ad
bgruening
parents: 6
diff changeset
319 validate_covariate(clin_df, args.clin_variable)
693011647a67 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit 6b520305ec30e6dc37eba92c67a5368cea0fc5ad
bgruening
parents: 6
diff changeset
320
2
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
321 except Exception as e:
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
322 print(f"Error: {e}", file=sys.stderr)
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
323 sys.exit(1)
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
324
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
325
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
326 if __name__ == "__main__":
ee7e61ab554d planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/flexynesis commit b2463fb68d0ae54864d87718ee72f5e063aa4587
bgruening
parents:
diff changeset
327 main()