Mercurial > repos > bgruening > sklearn_searchcv

annotate association_rules.py @ 23:bc3b489825b2 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/sklearn commit 80417bf0158a9b596e485dd66408f738f405145a
author bgruening
date Mon, 02 Oct 2023 07:59:32 +0000
parents 006db575e1f3
children
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
19
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
1 import argparse
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
2 import json
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
3 import warnings
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
4
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
5 import pandas as pd
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
6 from mlxtend.frequent_patterns import association_rules, fpgrowth
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
7 from mlxtend.preprocessing import TransactionEncoder
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
8
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
9
23
bc3b489825b2 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit 80417bf0158a9b596e485dd66408f738f405145a
bgruening
parents: 19
diff changeset
10 def main(
bc3b489825b2 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit 80417bf0158a9b596e485dd66408f738f405145a
bgruening
parents: 19
diff changeset
11 inputs,
bc3b489825b2 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit 80417bf0158a9b596e485dd66408f738f405145a
bgruening
parents: 19
diff changeset
12 infile,
bc3b489825b2 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit 80417bf0158a9b596e485dd66408f738f405145a
bgruening
parents: 19
diff changeset
13 outfile,
bc3b489825b2 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit 80417bf0158a9b596e485dd66408f738f405145a
bgruening
parents: 19
diff changeset
14 min_support=0.5,
bc3b489825b2 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit 80417bf0158a9b596e485dd66408f738f405145a
bgruening
parents: 19
diff changeset
15 min_confidence=0.5,
bc3b489825b2 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit 80417bf0158a9b596e485dd66408f738f405145a
bgruening
parents: 19
diff changeset
16 min_lift=1.0,
bc3b489825b2 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit 80417bf0158a9b596e485dd66408f738f405145a
bgruening
parents: 19
diff changeset
17 min_conviction=1.0,
bc3b489825b2 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit 80417bf0158a9b596e485dd66408f738f405145a
bgruening
parents: 19
diff changeset
18 max_length=None,
bc3b489825b2 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit 80417bf0158a9b596e485dd66408f738f405145a
bgruening
parents: 19
diff changeset
19 ):
19
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
20 """
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
21 Parameter
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
22 ---------
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
23 input : str
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
24 File path to galaxy tool parameter
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
25
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
26 infile : str
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
27 File paths of input vector
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
28
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
29 outfile : str
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
30 File path to output matrix
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
31
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
32 min_support: float
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
33 Minimum support
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
34
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
35 min_confidence: float
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
36 Minimum confidence
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
37
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
38 min_lift: float
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
39 Minimum lift
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
40
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
41 min_conviction: float
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
42 Minimum conviction
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
43
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
44 max_length: int
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
45 Maximum length
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
46
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
47 """
23
bc3b489825b2 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit 80417bf0158a9b596e485dd66408f738f405145a
bgruening
parents: 19
diff changeset
48 warnings.simplefilter("ignore")
19
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
49
23
bc3b489825b2 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit 80417bf0158a9b596e485dd66408f738f405145a
bgruening
parents: 19
diff changeset
50 with open(inputs, "r") as param_handler:
19
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
51 params = json.load(param_handler)
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
52
23
bc3b489825b2 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit 80417bf0158a9b596e485dd66408f738f405145a
bgruening
parents: 19
diff changeset
53 input_header = params["header0"]
bc3b489825b2 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit 80417bf0158a9b596e485dd66408f738f405145a
bgruening
parents: 19
diff changeset
54 header = "infer" if input_header else None
19
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
55
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
56 with open(infile) as fp:
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
57 lines = fp.read().splitlines()
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
58
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
59 if header is not None:
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
60 lines = lines[1:]
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
61
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
62 dataset = []
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
63 for line in lines:
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
64 line_items = line.split("\t")
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
65 dataset.append(line_items)
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
66
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
67 # TransactionEncoder learns the unique labels in the dataset and transforms the
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
68 # input dataset (a Python list of lists) into a one-hot encoded NumPy boolean array
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
69 te = TransactionEncoder()
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
70 te_ary = te.fit_transform(dataset)
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
71
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
72 # Turn the encoded NumPy array into a DataFrame
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
73 df = pd.DataFrame(te_ary, columns=te.columns_)
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
74
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
75 # Extract frequent itemsets for association rule mining
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
76 # use_colnames: Use DataFrames' column names in the returned DataFrame instead of column indices
23
bc3b489825b2 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit 80417bf0158a9b596e485dd66408f738f405145a
bgruening
parents: 19
diff changeset
77 frequent_itemsets = fpgrowth(
bc3b489825b2 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit 80417bf0158a9b596e485dd66408f738f405145a
bgruening
parents: 19
diff changeset
78 df, min_support=min_support, use_colnames=True, max_len=max_length
bc3b489825b2 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit 80417bf0158a9b596e485dd66408f738f405145a
bgruening
parents: 19
diff changeset
79 )
19
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
80
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
81 # Get association rules, with confidence larger than min_confidence
23
bc3b489825b2 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit 80417bf0158a9b596e485dd66408f738f405145a
bgruening
parents: 19
diff changeset
82 rules = association_rules(
bc3b489825b2 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit 80417bf0158a9b596e485dd66408f738f405145a
bgruening
parents: 19
diff changeset
83 frequent_itemsets, metric="confidence", min_threshold=min_confidence
bc3b489825b2 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit 80417bf0158a9b596e485dd66408f738f405145a
bgruening
parents: 19
diff changeset
84 )
19
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
85
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
86 # Filter association rules, keeping rules with lift and conviction larger than min_liftand and min_conviction
23
bc3b489825b2 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit 80417bf0158a9b596e485dd66408f738f405145a
bgruening
parents: 19
diff changeset
87 rules = rules[(rules["lift"] >= min_lift) & (rules["conviction"] >= min_conviction)]
19
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
88
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
89 # Convert columns from frozenset to list (more readable)
23
bc3b489825b2 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit 80417bf0158a9b596e485dd66408f738f405145a
bgruening
parents: 19
diff changeset
90 rules["antecedents"] = rules["antecedents"].apply(list)
bc3b489825b2 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit 80417bf0158a9b596e485dd66408f738f405145a
bgruening
parents: 19
diff changeset
91 rules["consequents"] = rules["consequents"].apply(list)
19
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
92
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
93 # The next 3 steps are intended to fix the order of the association
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
94 # rules generated, so tests that rely on diff'ing a desired output
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
95 # with an expected output can pass
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
96
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
97 # 1) Sort entry in every row/column for columns 'antecedents' and 'consequents'
23
bc3b489825b2 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit 80417bf0158a9b596e485dd66408f738f405145a
bgruening
parents: 19
diff changeset
98 rules["antecedents"] = rules["antecedents"].apply(lambda row: sorted(row))
bc3b489825b2 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit 80417bf0158a9b596e485dd66408f738f405145a
bgruening
parents: 19
diff changeset
99 rules["consequents"] = rules["consequents"].apply(lambda row: sorted(row))
19
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
100
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
101 # 2) Create two temporary string columns to sort on
23
bc3b489825b2 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit 80417bf0158a9b596e485dd66408f738f405145a
bgruening
parents: 19
diff changeset
102 rules["ant_str"] = rules["antecedents"].apply(lambda row: " ".join(row))
bc3b489825b2 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit 80417bf0158a9b596e485dd66408f738f405145a
bgruening
parents: 19
diff changeset
103 rules["con_str"] = rules["consequents"].apply(lambda row: " ".join(row))
19
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
104
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
105 # 3) Sort results so they are re-producable
23
bc3b489825b2 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit 80417bf0158a9b596e485dd66408f738f405145a
bgruening
parents: 19
diff changeset
106 rules.sort_values(by=["ant_str", "con_str"], inplace=True)
bc3b489825b2 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit 80417bf0158a9b596e485dd66408f738f405145a
bgruening
parents: 19
diff changeset
107 del rules["ant_str"]
bc3b489825b2 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit 80417bf0158a9b596e485dd66408f738f405145a
bgruening
parents: 19
diff changeset
108 del rules["con_str"]
19
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
109 rules.reset_index(drop=True, inplace=True)
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
110
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
111 # Write association rules and metrics to file
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
112 rules.to_csv(outfile, sep="\t", index=False)
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
113
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
114
23
bc3b489825b2 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit 80417bf0158a9b596e485dd66408f738f405145a
bgruening
parents: 19
diff changeset
115 if __name__ == "__main__":
19
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
116 aparser = argparse.ArgumentParser()
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
117 aparser.add_argument("-i", "--inputs", dest="inputs", required=True)
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
118 aparser.add_argument("-y", "--infile", dest="infile", required=True)
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
119 aparser.add_argument("-o", "--outfile", dest="outfile", required=True)
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
120 aparser.add_argument("-s", "--support", dest="support", default=0.5)
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
121 aparser.add_argument("-c", "--confidence", dest="confidence", default=0.5)
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
122 aparser.add_argument("-l", "--lift", dest="lift", default=1.0)
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
123 aparser.add_argument("-v", "--conviction", dest="conviction", default=1.0)
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
124 aparser.add_argument("-t", "--length", dest="length", default=5)
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
125 args = aparser.parse_args()
006db575e1f3 "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
bgruening
parents:
diff changeset
126
23
bc3b489825b2 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit 80417bf0158a9b596e485dd66408f738f405145a
bgruening
parents: 19
diff changeset
127 main(
bc3b489825b2 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit 80417bf0158a9b596e485dd66408f738f405145a
bgruening
parents: 19
diff changeset
128 args.inputs,
bc3b489825b2 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit 80417bf0158a9b596e485dd66408f738f405145a
bgruening
parents: 19
diff changeset
129 args.infile,
bc3b489825b2 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit 80417bf0158a9b596e485dd66408f738f405145a
bgruening
parents: 19
diff changeset
130 args.outfile,
bc3b489825b2 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit 80417bf0158a9b596e485dd66408f738f405145a
bgruening
parents: 19
diff changeset
131 min_support=float(args.support),
bc3b489825b2 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit 80417bf0158a9b596e485dd66408f738f405145a
bgruening
parents: 19
diff changeset
132 min_confidence=float(args.confidence),
bc3b489825b2 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit 80417bf0158a9b596e485dd66408f738f405145a
bgruening
parents: 19
diff changeset
133 min_lift=float(args.lift),
bc3b489825b2 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit 80417bf0158a9b596e485dd66408f738f405145a
bgruening
parents: 19
diff changeset
134 min_conviction=float(args.conviction),
bc3b489825b2 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit 80417bf0158a9b596e485dd66408f738f405145a
bgruening
parents: 19
diff changeset
135 max_length=int(args.length),
bc3b489825b2 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit 80417bf0158a9b596e485dd66408f738f405145a
bgruening
parents: 19
diff changeset
136 )