Mercurial > repos > bgruening > keras_batch_models
annotate association_rules.py @ 16:967647f7bbaa draft
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit 9981e25b00de29ed881b2229a173a8c812ded9bb
| author | bgruening | 
|---|---|
| date | Wed, 09 Aug 2023 12:42:22 +0000 | 
| parents | 70846a2dd227 | 
| children | 
| rev | line source | 
|---|---|
| 11 
508cf7278239
"planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
 bgruening parents: diff
changeset | 1 import argparse | 
| 
508cf7278239
"planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
 bgruening parents: diff
changeset | 2 import json | 
| 
508cf7278239
"planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
 bgruening parents: diff
changeset | 3 import warnings | 
| 
508cf7278239
"planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
 bgruening parents: diff
changeset | 4 | 
| 
508cf7278239
"planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
 bgruening parents: diff
changeset | 5 import pandas as pd | 
| 
508cf7278239
"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 | 
| 
508cf7278239
"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 | 
| 
508cf7278239
"planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
 bgruening parents: diff
changeset | 8 | 
| 
508cf7278239
"planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
 bgruening parents: diff
changeset | 9 | 
| 15 
70846a2dd227
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit 9981e25b00de29ed881b2229a173a8c812ded9bb
 bgruening parents: 
11diff
changeset | 10 def main( | 
| 
70846a2dd227
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit 9981e25b00de29ed881b2229a173a8c812ded9bb
 bgruening parents: 
11diff
changeset | 11 inputs, | 
| 
70846a2dd227
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit 9981e25b00de29ed881b2229a173a8c812ded9bb
 bgruening parents: 
11diff
changeset | 12 infile, | 
| 
70846a2dd227
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit 9981e25b00de29ed881b2229a173a8c812ded9bb
 bgruening parents: 
11diff
changeset | 13 outfile, | 
| 
70846a2dd227
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit 9981e25b00de29ed881b2229a173a8c812ded9bb
 bgruening parents: 
11diff
changeset | 14 min_support=0.5, | 
| 
70846a2dd227
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit 9981e25b00de29ed881b2229a173a8c812ded9bb
 bgruening parents: 
11diff
changeset | 15 min_confidence=0.5, | 
| 
70846a2dd227
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit 9981e25b00de29ed881b2229a173a8c812ded9bb
 bgruening parents: 
11diff
changeset | 16 min_lift=1.0, | 
| 
70846a2dd227
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit 9981e25b00de29ed881b2229a173a8c812ded9bb
 bgruening parents: 
11diff
changeset | 17 min_conviction=1.0, | 
| 
70846a2dd227
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit 9981e25b00de29ed881b2229a173a8c812ded9bb
 bgruening parents: 
11diff
changeset | 18 max_length=None, | 
| 
70846a2dd227
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit 9981e25b00de29ed881b2229a173a8c812ded9bb
 bgruening parents: 
11diff
changeset | 19 ): | 
| 11 
508cf7278239
"planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
 bgruening parents: diff
changeset | 20 """ | 
| 
508cf7278239
"planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
 bgruening parents: diff
changeset | 21 Parameter | 
| 
508cf7278239
"planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
 bgruening parents: diff
changeset | 22 --------- | 
| 
508cf7278239
"planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
 bgruening parents: diff
changeset | 23 input : str | 
| 
508cf7278239
"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 | 
| 
508cf7278239
"planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
 bgruening parents: diff
changeset | 25 | 
| 
508cf7278239
"planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
 bgruening parents: diff
changeset | 26 infile : str | 
| 
508cf7278239
"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 | 
| 
508cf7278239
"planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
 bgruening parents: diff
changeset | 28 | 
| 
508cf7278239
"planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
 bgruening parents: diff
changeset | 29 outfile : str | 
| 
508cf7278239
"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 | 
| 
508cf7278239
"planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
 bgruening parents: diff
changeset | 31 | 
| 
508cf7278239
"planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
 bgruening parents: diff
changeset | 32 min_support: float | 
| 
508cf7278239
"planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
 bgruening parents: diff
changeset | 33 Minimum support | 
| 
508cf7278239
"planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
 bgruening parents: diff
changeset | 34 | 
| 
508cf7278239
"planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
 bgruening parents: diff
changeset | 35 min_confidence: float | 
| 
508cf7278239
"planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
 bgruening parents: diff
changeset | 36 Minimum confidence | 
| 
508cf7278239
"planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
 bgruening parents: diff
changeset | 37 | 
| 
508cf7278239
"planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
 bgruening parents: diff
changeset | 38 min_lift: float | 
| 
508cf7278239
"planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
 bgruening parents: diff
changeset | 39 Minimum lift | 
| 
508cf7278239
"planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
 bgruening parents: diff
changeset | 40 | 
| 
508cf7278239
"planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
 bgruening parents: diff
changeset | 41 min_conviction: float | 
| 
508cf7278239
"planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
 bgruening parents: diff
changeset | 42 Minimum conviction | 
| 
508cf7278239
"planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
 bgruening parents: diff
changeset | 43 | 
| 
508cf7278239
"planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
 bgruening parents: diff
changeset | 44 max_length: int | 
| 
508cf7278239
"planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
 bgruening parents: diff
changeset | 45 Maximum length | 
| 
508cf7278239
"planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
 bgruening parents: diff
changeset | 46 | 
| 
508cf7278239
"planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
 bgruening parents: diff
changeset | 47 """ | 
| 15 
70846a2dd227
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit 9981e25b00de29ed881b2229a173a8c812ded9bb
 bgruening parents: 
11diff
changeset | 48 warnings.simplefilter("ignore") | 
| 11 
508cf7278239
"planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
 bgruening parents: diff
changeset | 49 | 
| 15 
70846a2dd227
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit 9981e25b00de29ed881b2229a173a8c812ded9bb
 bgruening parents: 
11diff
changeset | 50 with open(inputs, "r") as param_handler: | 
| 11 
508cf7278239
"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) | 
| 
508cf7278239
"planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
 bgruening parents: diff
changeset | 52 | 
| 15 
70846a2dd227
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit 9981e25b00de29ed881b2229a173a8c812ded9bb
 bgruening parents: 
11diff
changeset | 53 input_header = params["header0"] | 
| 
70846a2dd227
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit 9981e25b00de29ed881b2229a173a8c812ded9bb
 bgruening parents: 
11diff
changeset | 54 header = "infer" if input_header else None | 
| 11 
508cf7278239
"planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
 bgruening parents: diff
changeset | 55 | 
| 
508cf7278239
"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: | 
| 
508cf7278239
"planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
 bgruening parents: diff
changeset | 57 lines = fp.read().splitlines() | 
| 
508cf7278239
"planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
 bgruening parents: diff
changeset | 58 | 
| 
508cf7278239
"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: | 
| 
508cf7278239
"planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
 bgruening parents: diff
changeset | 60 lines = lines[1:] | 
| 
508cf7278239
"planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
 bgruening parents: diff
changeset | 61 | 
| 
508cf7278239
"planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
 bgruening parents: diff
changeset | 62 dataset = [] | 
| 
508cf7278239
"planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
 bgruening parents: diff
changeset | 63 for line in lines: | 
| 
508cf7278239
"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") | 
| 
508cf7278239
"planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
 bgruening parents: diff
changeset | 65 dataset.append(line_items) | 
| 
508cf7278239
"planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
 bgruening parents: diff
changeset | 66 | 
| 
508cf7278239
"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 | 
| 
508cf7278239
"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 | 
| 
508cf7278239
"planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
 bgruening parents: diff
changeset | 69 te = TransactionEncoder() | 
| 
508cf7278239
"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) | 
| 
508cf7278239
"planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
 bgruening parents: diff
changeset | 71 | 
| 
508cf7278239
"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 | 
| 
508cf7278239
"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_) | 
| 
508cf7278239
"planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
 bgruening parents: diff
changeset | 74 | 
| 
508cf7278239
"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 | 
| 
508cf7278239
"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 | 
| 15 
70846a2dd227
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit 9981e25b00de29ed881b2229a173a8c812ded9bb
 bgruening parents: 
11diff
changeset | 77 frequent_itemsets = fpgrowth( | 
| 
70846a2dd227
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit 9981e25b00de29ed881b2229a173a8c812ded9bb
 bgruening parents: 
11diff
changeset | 78 df, min_support=min_support, use_colnames=True, max_len=max_length | 
| 
70846a2dd227
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit 9981e25b00de29ed881b2229a173a8c812ded9bb
 bgruening parents: 
11diff
changeset | 79 ) | 
| 11 
508cf7278239
"planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
 bgruening parents: diff
changeset | 80 | 
| 
508cf7278239
"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 | 
| 15 
70846a2dd227
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit 9981e25b00de29ed881b2229a173a8c812ded9bb
 bgruening parents: 
11diff
changeset | 82 rules = association_rules( | 
| 
70846a2dd227
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit 9981e25b00de29ed881b2229a173a8c812ded9bb
 bgruening parents: 
11diff
changeset | 83 frequent_itemsets, metric="confidence", min_threshold=min_confidence | 
| 
70846a2dd227
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit 9981e25b00de29ed881b2229a173a8c812ded9bb
 bgruening parents: 
11diff
changeset | 84 ) | 
| 11 
508cf7278239
"planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
 bgruening parents: diff
changeset | 85 | 
| 
508cf7278239
"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 | 
| 15 
70846a2dd227
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit 9981e25b00de29ed881b2229a173a8c812ded9bb
 bgruening parents: 
11diff
changeset | 87 rules = rules[(rules["lift"] >= min_lift) & (rules["conviction"] >= min_conviction)] | 
| 11 
508cf7278239
"planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
 bgruening parents: diff
changeset | 88 | 
| 
508cf7278239
"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) | 
| 15 
70846a2dd227
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit 9981e25b00de29ed881b2229a173a8c812ded9bb
 bgruening parents: 
11diff
changeset | 90 rules["antecedents"] = rules["antecedents"].apply(list) | 
| 
70846a2dd227
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit 9981e25b00de29ed881b2229a173a8c812ded9bb
 bgruening parents: 
11diff
changeset | 91 rules["consequents"] = rules["consequents"].apply(list) | 
| 11 
508cf7278239
"planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
 bgruening parents: diff
changeset | 92 | 
| 
508cf7278239
"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 | 
| 
508cf7278239
"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 | 
| 
508cf7278239
"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 | 
| 
508cf7278239
"planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
 bgruening parents: diff
changeset | 96 | 
| 
508cf7278239
"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' | 
| 15 
70846a2dd227
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit 9981e25b00de29ed881b2229a173a8c812ded9bb
 bgruening parents: 
11diff
changeset | 98 rules["antecedents"] = rules["antecedents"].apply(lambda row: sorted(row)) | 
| 
70846a2dd227
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit 9981e25b00de29ed881b2229a173a8c812ded9bb
 bgruening parents: 
11diff
changeset | 99 rules["consequents"] = rules["consequents"].apply(lambda row: sorted(row)) | 
| 11 
508cf7278239
"planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
 bgruening parents: diff
changeset | 100 | 
| 
508cf7278239
"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 | 
| 15 
70846a2dd227
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit 9981e25b00de29ed881b2229a173a8c812ded9bb
 bgruening parents: 
11diff
changeset | 102 rules["ant_str"] = rules["antecedents"].apply(lambda row: " ".join(row)) | 
| 
70846a2dd227
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit 9981e25b00de29ed881b2229a173a8c812ded9bb
 bgruening parents: 
11diff
changeset | 103 rules["con_str"] = rules["consequents"].apply(lambda row: " ".join(row)) | 
| 11 
508cf7278239
"planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
 bgruening parents: diff
changeset | 104 | 
| 
508cf7278239
"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 | 
| 15 
70846a2dd227
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit 9981e25b00de29ed881b2229a173a8c812ded9bb
 bgruening parents: 
11diff
changeset | 106 rules.sort_values(by=["ant_str", "con_str"], inplace=True) | 
| 
70846a2dd227
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit 9981e25b00de29ed881b2229a173a8c812ded9bb
 bgruening parents: 
11diff
changeset | 107 del rules["ant_str"] | 
| 
70846a2dd227
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit 9981e25b00de29ed881b2229a173a8c812ded9bb
 bgruening parents: 
11diff
changeset | 108 del rules["con_str"] | 
| 11 
508cf7278239
"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) | 
| 
508cf7278239
"planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
 bgruening parents: diff
changeset | 110 | 
| 
508cf7278239
"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 | 
| 
508cf7278239
"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) | 
| 
508cf7278239
"planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
 bgruening parents: diff
changeset | 113 | 
| 
508cf7278239
"planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
 bgruening parents: diff
changeset | 114 | 
| 15 
70846a2dd227
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit 9981e25b00de29ed881b2229a173a8c812ded9bb
 bgruening parents: 
11diff
changeset | 115 if __name__ == "__main__": | 
| 11 
508cf7278239
"planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
 bgruening parents: diff
changeset | 116 aparser = argparse.ArgumentParser() | 
| 
508cf7278239
"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) | 
| 
508cf7278239
"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) | 
| 
508cf7278239
"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) | 
| 
508cf7278239
"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) | 
| 
508cf7278239
"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) | 
| 
508cf7278239
"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) | 
| 
508cf7278239
"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) | 
| 
508cf7278239
"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) | 
| 
508cf7278239
"planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
 bgruening parents: diff
changeset | 125 args = aparser.parse_args() | 
| 
508cf7278239
"planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ea12f973df4b97a2691d9e4ce6bf6fae59d57717"
 bgruening parents: diff
changeset | 126 | 
| 15 
70846a2dd227
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit 9981e25b00de29ed881b2229a173a8c812ded9bb
 bgruening parents: 
11diff
changeset | 127 main( | 
| 
70846a2dd227
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit 9981e25b00de29ed881b2229a173a8c812ded9bb
 bgruening parents: 
11diff
changeset | 128 args.inputs, | 
| 
70846a2dd227
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit 9981e25b00de29ed881b2229a173a8c812ded9bb
 bgruening parents: 
11diff
changeset | 129 args.infile, | 
| 
70846a2dd227
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit 9981e25b00de29ed881b2229a173a8c812ded9bb
 bgruening parents: 
11diff
changeset | 130 args.outfile, | 
| 
70846a2dd227
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit 9981e25b00de29ed881b2229a173a8c812ded9bb
 bgruening parents: 
11diff
changeset | 131 min_support=float(args.support), | 
| 
70846a2dd227
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit 9981e25b00de29ed881b2229a173a8c812ded9bb
 bgruening parents: 
11diff
changeset | 132 min_confidence=float(args.confidence), | 
| 
70846a2dd227
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit 9981e25b00de29ed881b2229a173a8c812ded9bb
 bgruening parents: 
11diff
changeset | 133 min_lift=float(args.lift), | 
| 
70846a2dd227
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit 9981e25b00de29ed881b2229a173a8c812ded9bb
 bgruening parents: 
11diff
changeset | 134 min_conviction=float(args.conviction), | 
| 
70846a2dd227
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit 9981e25b00de29ed881b2229a173a8c812ded9bb
 bgruening parents: 
11diff
changeset | 135 max_length=int(args.length), | 
| 
70846a2dd227
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit 9981e25b00de29ed881b2229a173a8c812ded9bb
 bgruening parents: 
11diff
changeset | 136 ) | 
