Mercurial > repos > bgruening > stacking_ensemble_models

comparison association_rules.py @ 3:0a1812986bc3 draft

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/sklearn commit 9981e25b00de29ed881b2229a173a8c812ded9bb
author bgruening
date Wed, 09 Aug 2023 11:10:37 +0000
parents
children
comparison
equal deleted inserted replaced
2:38c4f8a98038 3:0a1812986bc3
1 import argparse
2 import json
3 import warnings
4
5 import pandas as pd
6 from mlxtend.frequent_patterns import association_rules, fpgrowth
7 from mlxtend.preprocessing import TransactionEncoder
8
9
10 def main(
11 inputs,
12 infile,
13 outfile,
14 min_support=0.5,
15 min_confidence=0.5,
16 min_lift=1.0,
17 min_conviction=1.0,
18 max_length=None,
19 ):
20 """
21 Parameter
22 ---------
23 input : str
24 File path to galaxy tool parameter
25
26 infile : str
27 File paths of input vector
28
29 outfile : str
30 File path to output matrix
31
32 min_support: float
33 Minimum support
34
35 min_confidence: float
36 Minimum confidence
37
38 min_lift: float
39 Minimum lift
40
41 min_conviction: float
42 Minimum conviction
43
44 max_length: int
45 Maximum length
46
47 """
48 warnings.simplefilter("ignore")
49
50 with open(inputs, "r") as param_handler:
51 params = json.load(param_handler)
52
53 input_header = params["header0"]
54 header = "infer" if input_header else None
55
56 with open(infile) as fp:
57 lines = fp.read().splitlines()
58
59 if header is not None:
60 lines = lines[1:]
61
62 dataset = []
63 for line in lines:
64 line_items = line.split("\t")
65 dataset.append(line_items)
66
67 # TransactionEncoder learns the unique labels in the dataset and transforms the
68 # input dataset (a Python list of lists) into a one-hot encoded NumPy boolean array
69 te = TransactionEncoder()
70 te_ary = te.fit_transform(dataset)
71
72 # Turn the encoded NumPy array into a DataFrame
73 df = pd.DataFrame(te_ary, columns=te.columns_)
74
75 # Extract frequent itemsets for association rule mining
76 # use_colnames: Use DataFrames' column names in the returned DataFrame instead of column indices
77 frequent_itemsets = fpgrowth(
78 df, min_support=min_support, use_colnames=True, max_len=max_length
79 )
80
81 # Get association rules, with confidence larger than min_confidence
82 rules = association_rules(
83 frequent_itemsets, metric="confidence", min_threshold=min_confidence
84 )
85
86 # Filter association rules, keeping rules with lift and conviction larger than min_liftand and min_conviction
87 rules = rules[(rules["lift"] >= min_lift) & (rules["conviction"] >= min_conviction)]
88
89 # Convert columns from frozenset to list (more readable)
90 rules["antecedents"] = rules["antecedents"].apply(list)
91 rules["consequents"] = rules["consequents"].apply(list)
92
93 # The next 3 steps are intended to fix the order of the association
94 # rules generated, so tests that rely on diff'ing a desired output
95 # with an expected output can pass
96
97 # 1) Sort entry in every row/column for columns 'antecedents' and 'consequents'
98 rules["antecedents"] = rules["antecedents"].apply(lambda row: sorted(row))
99 rules["consequents"] = rules["consequents"].apply(lambda row: sorted(row))
100
101 # 2) Create two temporary string columns to sort on
102 rules["ant_str"] = rules["antecedents"].apply(lambda row: " ".join(row))
103 rules["con_str"] = rules["consequents"].apply(lambda row: " ".join(row))
104
105 # 3) Sort results so they are re-producable
106 rules.sort_values(by=["ant_str", "con_str"], inplace=True)
107 del rules["ant_str"]
108 del rules["con_str"]
109 rules.reset_index(drop=True, inplace=True)
110
111 # Write association rules and metrics to file
112 rules.to_csv(outfile, sep="\t", index=False)
113
114
115 if __name__ == "__main__":
116 aparser = argparse.ArgumentParser()
117 aparser.add_argument("-i", "--inputs", dest="inputs", required=True)
118 aparser.add_argument("-y", "--infile", dest="infile", required=True)
119 aparser.add_argument("-o", "--outfile", dest="outfile", required=True)
120 aparser.add_argument("-s", "--support", dest="support", default=0.5)
121 aparser.add_argument("-c", "--confidence", dest="confidence", default=0.5)
122 aparser.add_argument("-l", "--lift", dest="lift", default=1.0)
123 aparser.add_argument("-v", "--conviction", dest="conviction", default=1.0)
124 aparser.add_argument("-t", "--length", dest="length", default=5)
125 args = aparser.parse_args()
126
127 main(
128 args.inputs,
129 args.infile,
130 args.outfile,
131 min_support=float(args.support),
132 min_confidence=float(args.confidence),
133 min_lift=float(args.lift),
134 min_conviction=float(args.conviction),
135 max_length=int(args.length),
136 )