annotate MicroPITA.py @ 3:b4cf8c75305b draft default tip

Pointing to the right root directory
author george-weingart
date Tue, 30 Aug 2016 13:10:34 -0400
parents cd71e90abfab
children
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
0
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1 #!/usr/bin/env python
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
2 """
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
3 Author: Timothy Tickle
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
4 Description: Class to Run analysis for the microPITA paper
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
5 """
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
6
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
7 #####################################################################################
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
8 #Copyright (C) <2012>
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
9 #
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
10 #Permission is hereby granted, free of charge, to any person obtaining a copy of
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
11 #this software and associated documentation files (the "Software"), to deal in the
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
12 #Software without restriction, including without limitation the rights to use, copy,
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
13 #modify, merge, publish, distribute, sublicense, and/or sell copies of the Software,
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
14 #and to permit persons to whom the Software is furnished to do so, subject to
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
15 #the following conditions:
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
16 #
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
17 #The above copyright notice and this permission notice shall be included in all copies
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
18 #or substantial portions of the Software.
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
19 #
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
20 #THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED,
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
21 #INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
22 #PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
23 #HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
24 #OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
25 #SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
26 #####################################################################################
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
27
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
28 __author__ = "Timothy Tickle"
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
29 __copyright__ = "Copyright 2012"
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
30 __credits__ = ["Timothy Tickle"]
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
31 __license__ = "MIT"
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
32 __maintainer__ = "Timothy Tickle"
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
33 __email__ = "ttickle@sph.harvard.edu"
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
34 __status__ = "Development"
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
35
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
36 import sys
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
37 import argparse
1
cd71e90abfab Updated version that suppresses the future warnings option that was causing a problem
george-weingart
parents: 0
diff changeset
38 import warnings
cd71e90abfab Updated version that suppresses the future warnings option that was causing a problem
george-weingart
parents: 0
diff changeset
39 warnings.simplefilter(action = "ignore", category = FutureWarning)
0
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
40 from src.breadcrumbs.src.AbundanceTable import AbundanceTable
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
41 from src.breadcrumbs.src.ConstantsBreadCrumbs import ConstantsBreadCrumbs
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
42 from src.breadcrumbs.src.Metric import Metric
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
43 from src.breadcrumbs.src.KMedoids import Kmedoids
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
44 from src.breadcrumbs.src.MLPYDistanceAdaptor import MLPYDistanceAdaptor
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
45 from src.breadcrumbs.src.SVM import SVM
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
46 from src.breadcrumbs.src.UtilityMath import UtilityMath
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
47
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
48 from src.ConstantsMicropita import ConstantsMicropita
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
49 import csv
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
50 import logging
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
51 import math
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
52 import mlpy
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
53 import numpy as np
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
54 import operator
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
55 import os
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
56 import random
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
57 import scipy.cluster.hierarchy as hcluster
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
58 import scipy.spatial.distance
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
59 from types import *
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
60
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
61 class MicroPITA:
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
62 """
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
63 Selects samples from a first tier of a multi-tiered study to be used in a second tier.
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
64 Different methods can be used for selection.
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
65 The expected input is an abundance table (and potentially a text file of targeted features,
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
66 if using the targeted features option). Output is a list of samples exhibiting the
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
67 characteristics of interest.
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
68 """
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
69
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
70 #Constants
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
71 #Diversity metrics Alpha
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
72 c_strInverseSimpsonDiversity = Metric.c_strInvSimpsonDiversity
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
73 c_strChao1Diversity = Metric.c_strChao1Diversity
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
74
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
75 #Diversity metrics Beta
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
76 c_strBrayCurtisDissimilarity = Metric.c_strBrayCurtisDissimilarity
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
77
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
78 #Additive inverses of diversity metrics beta
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
79 c_strInvBrayCurtisDissimilarity = Metric.c_strInvBrayCurtisDissimilarity
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
80
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
81 #Technique Names
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
82 ConstantsMicropita.c_strDiversity2 = ConstantsMicropita.c_strDiversity+"_C"
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
83
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
84 #Targeted feature settings
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
85 c_strTargetedRanked = ConstantsMicropita.c_strTargetedRanked
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
86 c_strTargetedAbundance = ConstantsMicropita.c_strTargetedAbundance
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
87
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
88 #Technique groupings
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
89 # c_lsDiversityMethods = [ConstantsMicropita.c_strDiversity,ConstantsMicropita.c_strDiversity2]
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
90
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
91 #Converts ecology metrics into standardized method selection names
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
92 dictConvertAMetricDiversity = {c_strInverseSimpsonDiversity:ConstantsMicropita.c_strDiversity, c_strChao1Diversity:ConstantsMicropita.c_strDiversity2}
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
93 # dictConvertMicroPITAToAMetric = {ConstantsMicropita.c_strDiversity:c_strInverseSimpsonDiversity, ConstantsMicropita.c_strDiversity2:c_strChao1Diversity}
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
94 dictConvertBMetricToMethod = {c_strBrayCurtisDissimilarity:ConstantsMicropita.c_strRepresentative}
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
95 dictConvertInvBMetricToMethod = {c_strBrayCurtisDissimilarity:ConstantsMicropita.c_strExtreme}
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
96
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
97 #Linkage used in the Hierarchical clustering
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
98 c_strHierarchicalClusterMethod = 'average'
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
99
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
100 ####Group 1## Diversity
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
101 #Testing: Happy path Testing (8)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
102 def funcGetTopRankedSamples(self, lldMatrix = None, lsSampleNames = None, iTopAmount = None):
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
103 """
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
104 Given a list of lists of measurements, for each list the indices of the highest values are returned. If lsSamplesNames is given
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
105 it is treated as a list of string names that is in the order of the measurements in each list. Indices are returned or the sample
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
106 names associated with the indices.
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
107
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
108 :param lldMatrix: List of lists [[value,value,value,value],[value,value,value,value]].
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
109 :type: List of lists List of measurements. Each list is a different measurement. Each measurement in positionally related to a sample.
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
110 :param lsSampleNames: List of sample names positionally related (the same) to each list (Optional).
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
111 :type: List of strings List of strings.
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
112 :param iTopAmount: The amount of top measured samples (assumes the higher measurements are better).
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
113 :type: integer Integer amount of sample names/ indices to return.
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
114 :return List: List of samples to be selected.
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
115 """
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
116 topRankListRet = []
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
117 for rowMetrics in lldMatrix:
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
118 #Create 2 d array to hold value and index and sort
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
119 liIndexX = [rowMetrics,range(len(rowMetrics))]
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
120 liIndexX[1].sort(key = liIndexX[0].__getitem__,reverse = True)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
121
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
122 if lsSampleNames:
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
123 topRankListRet.append([lsSampleNames[iIndex] for iIndex in liIndexX[1][:iTopAmount]])
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
124 else:
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
125 topRankListRet.append(liIndexX[1][:iTopAmount])
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
126
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
127 return topRankListRet
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
128
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
129 ####Group 2## Representative Dissimilarity
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
130 #Testing: Happy path tested 1
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
131 def funcGetCentralSamplesByKMedoids(self, npaMatrix=None, sMetric=None, lsSampleNames=None, iNumberSamplesReturned=0, istmBetaMatrix=None, istrmTree=None, istrmEnvr=None):
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
132 """
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
133 Gets centroid samples by k-medoids clustering of a given matrix.
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
134
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
135 :param npaMatrix: Numpy array where row=features and columns=samples
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
136 :type: Numpy array Abundance Data.
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
137 :param sMetric: String name of beta metric used as the distance metric.
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
138 :type: String String name of beta metric.
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
139 :param lsSampleNames: The names of the sample
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
140 :type: List List of strings
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
141 :param iNumberSamplesReturned: Number of samples to return, each will be a centroid of a sample.
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
142 :type: Integer Number of samples to return
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
143 :return List: List of selected samples.
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
144 :param istmBetaMatrix: File with beta-diversity matrix
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
145 :type: File stream or file path string
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
146 """
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
147
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
148 #Count of how many rows
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
149 sampleCount = npaMatrix.shape[0]
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
150 if iNumberSamplesReturned > sampleCount:
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
151 logging.error("MicroPITA.funcGetCentralSamplesByKMedoids:: There are not enough samples to return the amount of samples specified. Return sample count = "+str(iNumberSamplesReturned)+". Sample number = "+str(sampleCount)+".")
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
152 return False
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
153
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
154 #If the cluster count is equal to the sample count return all samples
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
155 if sampleCount == iNumberSamplesReturned:
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
156 return list(lsSampleNames)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
157
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
158 #Get distance matrix
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
159 distanceMatrix=scipy.spatial.distance.squareform(Metric.funcReadMatrixFile(istmMatrixFile=istmBetaMatrix,lsSampleOrder=lsSampleNames)[0]) if istmBetaMatrix else Metric.funcGetBetaMetric(npadAbundancies=npaMatrix, sMetric=sMetric, istrmTree=istrmTree, istrmEnvr=istrmEnvr, lsSampleOrder=lsSampleNames)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
160 if type(distanceMatrix) is BooleanType:
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
161 logging.error("MicroPITA.funcGetCentralSamplesByKMedoids:: Could not read in the supplied distance matrix, returning false.")
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
162 return False
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
163
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
164 # Handle unifrac output
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
165 if sMetric in [Metric.c_strUnifracUnweighted,Metric.c_strUnifracWeighted]:
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
166 distanceMatrix = distanceMatrix[0]
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
167
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
168 #Log distance matrix
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
169 logging.debug("MicroPITA.funcGetCentralSamplesByKMedoids:: Distance matrix for representative selection using metric="+str(sMetric))
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
170
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
171 distance = MLPYDistanceAdaptor(npaDistanceMatrix=distanceMatrix, fIsCondensedMatrix=True)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
172
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
173 #Create object to determine clusters/medoids
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
174 medoidsMaker = Kmedoids(k=iNumberSamplesReturned, dist=distance)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
175 #medoidsData includes(1d numpy array, medoids indexes;
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
176 # 1d numpy array, non-medoids indexes;
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
177 # 1d numpy array, cluster membership for non-medoids;
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
178 # double, cost of configuration)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
179 #npaMatrix is samples x rows
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
180 #Build a matrix of lists of indicies to pass to the distance matrix
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
181 lliIndicesMatrix = [[iIndexPosition] for iIndexPosition in xrange(0,len(npaMatrix))]
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
182 medoidsData = medoidsMaker.compute(np.array(lliIndicesMatrix))
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
183 logging.debug("MicroPITA.funcGetCentralSamplesByKMedoids:: Results from the kmedoid method in representative selection:")
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
184 logging.debug(str(medoidsData))
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
185
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
186 #If returning the same amount of clusters and samples
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
187 #Return centroids
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
188 selectedIndexes = medoidsData[0]
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
189 return [lsSampleNames[selectedIndexes[index]] for index in xrange(0,iNumberSamplesReturned)]
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
190
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
191 ####Group 3## Highest Dissimilarity
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
192 #Testing: Happy path tested
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
193 def funcSelectExtremeSamplesFromHClust(self, strBetaMetric, npaAbundanceMatrix, lsSampleNames, iSelectSampleCount, istmBetaMatrix=None, istrmTree=None, istrmEnvr=None):
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
194 """
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
195 Select extreme samples from HClustering.
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
196
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
197 :param strBetaMetric: The beta metric to use for distance matrix generation.
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
198 :type: String The name of the beta metric to use.
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
199 :param npaAbundanceMatrix: Numpy array where row=samples and columns=features.
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
200 :type: Numpy Array Abundance data.
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
201 :param lsSampleNames: The names of the sample.
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
202 :type: List List of strings.
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
203 :param iSelectSampleCount: Number of samples to select (return).
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
204 :type: Integer Integer number of samples returned.
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
205 :return Samples: List of samples.
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
206 :param istmBetaMatrix: File with beta-diversity matrix
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
207 :type: File stream or file path string
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
208 """
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
209
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
210 #If they want all the sample count, return all sample names
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
211 iSampleCount=len(npaAbundanceMatrix[:,0])
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
212 if iSelectSampleCount==iSampleCount:
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
213 return lsSampleNames
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
214
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
215 #Holds the samples to be returned
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
216 lsReturnSamplesRet = []
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
217
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
218 #Generate beta matrix
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
219 #Returns condensed matrix
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
220 tempDistanceMatrix = scipy.spatial.distance.squareform(Metric.funcReadMatrixFile(istmMatrixFile=istmBetaMatrix,lsSampleOrder=lsSampleNames)[0]) if istmBetaMatrix else Metric.funcGetBetaMetric(npadAbundancies=npaAbundanceMatrix, sMetric=strBetaMetric, istrmTree=istrmTree, istrmEnvr=istrmEnvr, lsSampleOrder=lsSampleNames, fAdditiveInverse = True)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
221
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
222 if strBetaMetric in [Metric.c_strUnifracUnweighted,Metric.c_strUnifracWeighted]:
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
223 tempDistanceMatrix = tempDistanceMatrix[0]
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
224
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
225 if type(tempDistanceMatrix) is BooleanType:
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
226 logging.error("MicroPITA.funcSelectExtremeSamplesFromHClust:: Could not read in the supplied distance matrix, returning false.")
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
227 return False
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
228
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
229 if istmBetaMatrix:
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
230 tempDistanceMatrix = 1-tempDistanceMatrix
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
231
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
232 #Feed beta matrix to linkage to cluster
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
233 #Send condensed matrix
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
234 linkageMatrix = hcluster.linkage(tempDistanceMatrix, method=self.c_strHierarchicalClusterMethod)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
235
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
236 #Extract cluster information from dendrogram
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
237 #The linakge matrix is of the form
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
238 #[[int1 int2 doube int3],...]
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
239 #int1 and int1 are the paired samples indexed at 0 and up.
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
240 #each list is an entry for a branch that is number starting with the first
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
241 #list being sample count index + 1
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
242 #each list is then named by an increment as they appear
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
243 #this means that if a number is in the list and is = sample count or greater it is not
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
244 #terminal and is instead a branch.
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
245 #This method just takes the lowest metric measurement (highest distance pairs/clusters)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
246 #Works much better than the original technique
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
247 #get total number of samples
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
248
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
249 iCurrentSelectCount = 0
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
250 for row in linkageMatrix:
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
251 #Get nodes ofthe lowest pairing (so the furthest apart pair)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
252 iNode1 = int(row[0])
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
253 iNode2 = int(row[1])
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
254 #Make sure the nodes are a terminal node (sample) and not a branch in the dendrogram
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
255 #The branching in the dendrogram will start at the number of samples and increment higher.
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
256 #Add each of the pair one at a time breaking when enough samples are selected.
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
257 if iNode1<iSampleCount:
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
258 lsReturnSamplesRet.append(lsSampleNames[iNode1])
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
259 iCurrentSelectCount = iCurrentSelectCount + 1
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
260 if iCurrentSelectCount == iSelectSampleCount:
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
261 break
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
262 if iNode2<iSampleCount:
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
263 lsReturnSamplesRet.append(lsSampleNames[iNode2])
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
264 iCurrentSelectCount = iCurrentSelectCount + 1
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
265 if iCurrentSelectCount == iSelectSampleCount:
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
266 break
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
267
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
268 #Return selected samples
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
269 return lsReturnSamplesRet
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
270
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
271 ####Group 4## Rank Average of user Defined Taxa
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
272 #Testing: Happy Path Tested
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
273 def funcGetAverageAbundanceSamples(self, abndTable, lsTargetedFeature, fRank=False):
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
274 """
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
275 Averages feature abundance or ranked abundance. Expects a column 0 of taxa id that is skipped.
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
276
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
277 :param abndTable: Abundance Table to analyse
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
278 :type: AbundanceTable Abundance Table
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
279 :param lsTargetedFeature: String names
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
280 :type: list list of string names of features (bugs) which are measured after ranking against the full sample
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
281 :param fRank: Indicates to rank the abundance before getting the average abundance of the features (default false)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
282 :type: boolean Flag indicating ranking abundance before calculating average feature measurement (false= no ranking)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
283 :return List of lists or boolean: List of lists or False on error. One internal list per sample indicating the sample,
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
284 feature average abundance or ranked abundance. Lists will already be sorted.
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
285 For not Ranked [[sample,average abundance of selected feature,1]]
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
286 For Ranked [[sample,average ranked abundance, average abundance of selected feature]]
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
287 Error Returns false
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
288 """
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
289
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
290 llAbundance = abndTable.funcGetAverageAbundancePerSample(lsTargetedFeature)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
291 if not llAbundance:
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
292 logging.error("MicroPITA.funcGetAverageAbundanceSamples:: Could not get average abundance, returned false. Make sure the features (bugs) are spelled correctly and in the abundance table.")
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
293 return False
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
294 #Add a space for ranking if needed
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
295 #Not ranked will be [[sSample,average abundance,1]]
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
296 #(where 1 will not discriminant ties if used in later functions, so this generalizes)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
297 #Ranked will be [[sSample, average rank, average abundance]]
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
298 llRetAbundance = [[llist[0],-1,llist[1]] for llist in llAbundance]
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
299 #Rank if needed
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
300 if fRank:
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
301 abndRanked = abndTable.funcRankAbundance()
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
302 if abndRanked == None:
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
303 logging.error("MicroPITA.funcGetAverageAbundanceSamples:: Could not rank the abundance table, returned false.")
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
304 return False
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
305 llRetRank = abndRanked.funcGetAverageAbundancePerSample(lsTargetedFeature)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
306 if not llRetRank:
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
307 logging.error("MicroPITA.funcGetAverageAbundanceSamples:: Could not get average ranked abundance, returned false. Make sure the features (bugs) are spelled correctly and in the abundance table.")
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
308 return False
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
309 dictRanks = dict(llRetRank)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
310 llRetAbundance = [[a[0],dictRanks[a[0]],a[2]] for a in llRetAbundance]
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
311
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
312 #Sort first for ties and then for the main feature
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
313 if not fRank or ConstantsMicropita.c_fBreakRankTiesByDiversity:
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
314 llRetAbundance = sorted(llRetAbundance, key = lambda sampleData: sampleData[2], reverse = not fRank)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
315 if fRank:
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
316 llRetAbundance = sorted(llRetAbundance, key = lambda sampleData: sampleData[1], reverse = not fRank)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
317 return llRetAbundance
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
318
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
319 #Testing: Happy Path Tested
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
320 def funcSelectTargetedTaxaSamples(self, abndMatrix, lsTargetedTaxa, iSampleSelectionCount, sMethod = ConstantsMicropita.lsTargetedFeatureMethodValues[0]):
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
321 """
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
322 Selects samples with the highest ranks or abundance of targeted features.
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
323 If ranked, select the highest abundance for tie breaking
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
324
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
325 :param abndMatrix: Abundance table to analyse
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
326 :type: AbundanceTable Abundance table
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
327 :param lsTargetedTaxa: List of features
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
328 :type: list list of strings
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
329 :param iSampleSelectionCount: Number of samples to select
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
330 :type: integer integer
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
331 :param sMethod: Method to select targeted features
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
332 :type: string String (Can be values found in ConstantsMicropita.lsTargetedFeatureMethodValues)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
333 :return List of strings: List of sample names which were selected
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
334 List of strings Empty list is returned on an error.
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
335 """
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
336
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
337 #Check data
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
338 if(len(lsTargetedTaxa) < 1):
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
339 logging.error("MicroPITA.funcSelectTargetedTaxaSamples. Taxa defined selection was requested but no features were given.")
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
340 return []
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
341
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
342 lsTargetedSamples = self.funcGetAverageAbundanceSamples(abndTable=abndMatrix, lsTargetedFeature=lsTargetedTaxa,
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
343 fRank=sMethod.lower() == self.c_strTargetedRanked.lower())
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
344 #If an error occured or the key word for the method was not recognized
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
345 if lsTargetedSamples == False:
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
346 logging.error("MicroPITA.funcSelectTargetedTaxaSamples:: Was not able to select for the features given. So targeted feature selection was performed. Check to make sure the features are spelled correctly and exist in the abundance file.")
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
347 return []
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
348
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
349 #Select from results
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
350 return [sSample[0] for sSample in lsTargetedSamples[:iSampleSelectionCount]]
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
351
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
352 ####Group 5## Random
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
353 #Testing: Happy path Tested
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
354 def funcGetRandomSamples(self, lsSamples=None, iNumberOfSamplesToReturn=0):
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
355 """
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
356 Returns random sample names of the number given. No replacement.
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
357
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
358 :param lsSamples: List of sample names
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
359 :type: list list of strings
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
360 :param iNumberOfSamplesToReturn: Number of samples to select
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
361 :type: integer integer.
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
362 :return List: List of selected samples (strings).
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
363 """
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
364
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
365 #Input matrix sample count
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
366 sampleCount = len(lsSamples)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
367
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
368 #Return the full matrix if they ask for a return matrix where length == original
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
369 if(iNumberOfSamplesToReturn >= sampleCount):
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
370 return lsSamples
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
371
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
372 #Get the random indices for the sample (without replacement)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
373 liRandomIndices = random.sample(range(sampleCount), iNumberOfSamplesToReturn)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
374
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
375 #Create a boolean array of if indexes are to be included in the reduced array
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
376 return [sSample for iIndex, sSample in enumerate(lsSamples) if iIndex in liRandomIndices]
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
377
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
378 #Happy path tested (case 3)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
379 def funcGetAveragePopulation(self, abndTable, lfCompress):
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
380 """
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
381 Get the average row per column in the abndtable.
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
382
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
383 :param abndTable: AbundanceTable of data to be averaged
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
384 :type: AbudanceTable
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
385 :param lfCompress: List of boolean flags (false means to remove sample before averaging
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
386 :type: List of floats
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
387 :return List of doubles:
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
388 """
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
389 if sum(lfCompress) == 0:
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
390 return []
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
391
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
392 #Get the average populations
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
393 lAverageRet = []
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
394
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
395 for sFeature in abndTable.funcGetAbundanceCopy():
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
396 sFeature = list(sFeature)[1:]
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
397 sFeature=np.compress(lfCompress,sFeature,axis=0)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
398 lAverageRet.append(sum(sFeature)/float(len(sFeature)))
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
399 return lAverageRet
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
400
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
401 #Happy path tested (2 cases)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
402 def funcGetDistanceFromAverage(self, abndTable,ldAverage,lsSamples,lfSelected):
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
403 """
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
404 Given an abundance table and an average sample, this returns the distance of each sample
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
405 (measured using brays-curtis dissimilarity) from the average.
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
406 The distances are reduced by needing to be in the lsSamples and being a true in the lfSelected
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
407 (which is associated with the samples in the order of the samples in the abundance table;
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
408 use abundancetable.funcGetSampleNames() to see the order if needed).
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
409
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
410 :param abndTable: Abundance table holding the data to be analyzed.
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
411 :type: AbundanceTable
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
412 :param ldAverage: Average population (Average features of the abundance table of samples)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
413 :type: List of doubles which represent the average population
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
414 :param lsSamples: These are the only samples used in the analysis
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
415 :type: List of strings (sample ids)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
416 :param lfSelected: Samples to be included in the analysis
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
417 :type: List of boolean (true means include)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
418 :return: List of distances (doubles)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
419 """
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
420 #Get the distance from label 1 of all samples in label0 splitting into selected and not selected lists
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
421 ldSelectedDistances = []
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
422
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
423 for sSampleName in [sSample for iindex, sSample in enumerate(lsSamples) if lfSelected[iindex]]:
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
424 #Get the sample measurements
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
425 ldSelectedDistances.append(Metric.funcGetBrayCurtisDissimilarity(np.array([abndTable.funcGetSample(sSampleName),ldAverage]))[0])
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
426 return ldSelectedDistances
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
427
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
428 #Happy path tested (1 case)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
429 def funcMeasureDistanceFromLabelToAverageOtherLabel(self, abndTable, lfGroupOfInterest, lfGroupOther):
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
430 """
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
431 Get the distance of samples from one label from the average sample of not the label.
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
432 Note: This assumes 2 classes.
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
433
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
434 :param abndTable: Table of data to work out of.
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
435 :type: Abundace Table
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
436 :param lfGroupOfInterest: Boolean indicator of the sample being in the first group.
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
437 :type: List of floats, true indicating an individual in the group of interest.
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
438 :param lfGroupOther: Boolean indicator of the sample being in the other group.
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
439 :type: List of floats, true indicating an individual in the
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
440 :return List of List of doubles: [list of tuples (string sample name,double distance) for the selected population, list of tuples for the not selected population]
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
441 """
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
442 #Get all sample names
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
443 lsAllSamples = abndTable.funcGetSampleNames()
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
444
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
445 #Get average populations
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
446 lAverageOther = self.funcGetAveragePopulation(abndTable=abndTable, lfCompress=lfGroupOther)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
447
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
448 #Get the distance from the average of the other label (label 1)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
449 ldSelectedDistances = self.funcGetDistanceFromAverage(abndTable=abndTable, ldAverage=lAverageOther,
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
450 lsSamples=lsAllSamples, lfSelected=lfGroupOfInterest)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
451
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
452 return zip([lsAllSamples[iindex] for iindex, fGroup in enumerate(lfGroupOfInterest) if fGroup],ldSelectedDistances)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
453
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
454 #Happy path tested (1 test case)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
455 def funcPerformDistanceSelection(self, abndTable, iSelectionCount, sLabel, sValueOfInterest):
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
456 """
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
457 Given metadata, metadata of one value (sValueOfInterest) is measured from the average (centroid) value of another label group.
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
458 An iSelectionCount of samples is selected from the group of interest closest to and furthest from the centroid of the other group.
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
459
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
460 :params abndTable: Abundance of measurements
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
461 :type: AbundanceTable
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
462 :params iSelectionCount: The number of samples selected per sample.
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
463 :type: Integer Integer greater than 0
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
464 :params sLabel: ID of the metadata which is the supervised label
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
465 :type: String
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
466 :params sValueOfInterest: Metadata value in the sLabel metadta row of the abundance table which defines the group of interest.
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
467 :type: String found in the abundance table metadata row indicated by sLabel.
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
468 :return list list of tuples (samplename, distance) [[iSelectionCount of tuples closest to the other centroid], [iSelectionCount of tuples farthest from the other centroid], [all tuples of samples not selected]]
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
469 """
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
470
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
471 lsMetadata = abndTable.funcGetMetadata(sLabel)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
472 #Other metadata values
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
473 lsUniqueOtherValues = list(set(lsMetadata)-set(sValueOfInterest))
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
474
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
475 #Get boolean indicator of values of interest
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
476 lfLabelsInterested = [sValueOfInterest == sValue for sValue in lsMetadata]
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
477
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
478 #Get the distances of the items of interest from the other metadata values
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
479 dictDistanceAverages = {}
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
480 for sOtherLabel in lsUniqueOtherValues:
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
481 #Get boolean indicator of labels not of interest
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
482 lfLabelsOther = [sOtherLabel == sValue for sValue in lsMetadata]
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
483
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
484 #Get the distances of data from two different groups to the average of the other
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
485 ldValueDistances = dict(self.funcMeasureDistanceFromLabelToAverageOtherLabel(abndTable, lfLabelsInterested, lfLabelsOther))
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
486
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
487 for sKey in ldValueDistances:
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
488 dictDistanceAverages[sKey] = ldValueDistances[sKey] + dictDistanceAverages[sKey] if sKey in dictDistanceAverages else ldValueDistances[sKey]
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
489
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
490 #Finish average by dividing by length of lsUniqueOtherValues
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
491 ltpleAverageDistances = [(sKey, dictDistanceAverages[sKey]/float(len(lsUniqueOtherValues))) for sKey in dictDistanceAverages]
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
492
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
493 #Sort to extract extremes
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
494 ltpleAverageDistances = sorted(ltpleAverageDistances,key=operator.itemgetter(1))
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
495
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
496 #Get the closest and farthest distances
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
497 ltupleDiscriminantSamples = ltpleAverageDistances[:iSelectionCount]
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
498 ltupleDistinctSamples = ltpleAverageDistances[iSelectionCount*-1:]
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
499
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
500 #Remove the selected samples from the larger population of distances (better visualization)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
501 ldSelected = [tpleSelected[0] for tpleSelected in ltupleDiscriminantSamples+ltupleDistinctSamples]
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
502
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
503 #Return discriminant tuples, distinct tuples, other tuples
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
504 return [ltupleDiscriminantSamples, ltupleDistinctSamples,
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
505 [tplData for tplData in ltpleAverageDistances if tplData[0] not in ldSelected]]
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
506
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
507 #Run the supervised method surrounding distance from centroids
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
508 #Happy path tested (3 test cases)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
509 def funcRunSupervisedDistancesFromCentroids(self, abundanceTable, fRunDistinct, fRunDiscriminant,
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
510 xOutputSupFile, xPredictSupFile, strSupervisedMetadata,
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
511 iSampleSupSelectionCount, lsOriginalSampleNames, lsOriginalLabels, fAppendFiles = False):
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
512 """
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
513 Runs supervised methods based on measuring distances of one label from the centroid of another. NAs are evaluated as theirown group.
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
514
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
515 :param abundanceTable: AbundanceTable
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
516 :type: AbudanceTable Data to analyze
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
517 :param fRunDistinct: Run distinct selection method
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
518 :type: Boolean boolean (true runs method)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
519 :param fRunDiscriminant: Run discriminant method
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
520 :type: Boolean boolean (true runs method)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
521 :param xOutputSupFile: File output from supervised methods detailing data going into the method.
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
522 :type: String or FileStream
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
523 :param xPredictSupFile: File output from supervised methods distance results from supervised methods.
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
524 :type: String or FileStream
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
525 :param strSupervisedMetadata: The metadata that will be used to group samples.
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
526 :type: String
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
527 :param iSampleSupSelectionCount: Number of samples to select
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
528 :type: Integer int sample selection count
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
529 :param lsOriginalSampleNames: List of the sample names, order is important and should be preserved from the abundanceTable.
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
530 :type: List of samples
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
531 :param fAppendFiles: Indicates that output files already exist and appending is occuring.
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
532 :type: Boolean
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
533 :return Selected Samples: A dictionary of selected samples by selection ID
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
534 Dictionary {"Selection Method":["SampleID","SampleID"...]}
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
535 """
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
536 #Get labels and run one label against many
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
537 lstrMetadata = abundanceTable.funcGetMetadata(strSupervisedMetadata)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
538 dictlltpleDistanceMeasurements = {}
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
539 for sMetadataValue in set(lstrMetadata):
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
540
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
541 #For now perform the selection here for the label of interest against the other labels
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
542 dictlltpleDistanceMeasurements.setdefault(sMetadataValue,[]).extend(self.funcPerformDistanceSelection(abndTable=abundanceTable,
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
543 iSelectionCount=iSampleSupSelectionCount, sLabel=strSupervisedMetadata, sValueOfInterest=sMetadataValue))
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
544
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
545 #Make expected output files for supervised methods
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
546 #1. Output file which is similar to an input file for SVMs
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
547 #2. Output file that is similar to the probabilitic output of a SVM (LibSVM)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
548 #Manly for making output of supervised methods (Distance from Centroid) similar
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
549 #MicropitaVis needs some of these files
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
550 if xOutputSupFile:
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
551 if fAppendFiles:
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
552 SVM.funcUpdateSVMFileWithAbundanceTable(abndAbundanceTable=abundanceTable, xOutputSVMFile=xOutputSupFile,
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
553 lsOriginalLabels=lsOriginalLabels, lsSampleOrdering=lsOriginalSampleNames)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
554 else:
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
555 SVM.funcConvertAbundanceTableToSVMFile(abndAbundanceTable=abundanceTable, xOutputSVMFile=xOutputSupFile,
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
556 sMetadataLabel=strSupervisedMetadata, lsOriginalLabels=lsOriginalLabels, lsSampleOrdering=lsOriginalSampleNames)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
557
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
558 #Will contain the samples selected to return
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
559 #One or more of the methods may be active so this is why I am extending instead of just returning the result of each method type
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
560 dictSelectedSamplesRet = dict()
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
561 for sKey, ltplDistances in dictlltpleDistanceMeasurements.items():
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
562 if fRunDistinct:
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
563 dictSelectedSamplesRet.setdefault(ConstantsMicropita.c_strDistinct,[]).extend([ltple[0] for ltple in ltplDistances[1]])
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
564 if fRunDiscriminant:
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
565 dictSelectedSamplesRet.setdefault(ConstantsMicropita.c_strDiscriminant,[]).extend([ltple[0] for ltple in ltplDistances[0]])
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
566
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
567 if xPredictSupFile:
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
568 dictFlattenedDistances = dict()
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
569 [dictFlattenedDistances.setdefault(sKey, []).append(tple)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
570 for sKey, lltple in dictlltpleDistanceMeasurements.items()
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
571 for ltple in lltple for tple in ltple]
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
572 if fAppendFiles:
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
573 self._updatePredictFile(xPredictSupFile=xPredictSupFile, xInputLabelsFile=xOutputSupFile,
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
574 dictltpleDistanceMeasurements=dictFlattenedDistances, abundanceTable=abundanceTable, lsOriginalSampleNames=lsOriginalSampleNames)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
575 else:
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
576 self._writeToPredictFile(xPredictSupFile=xPredictSupFile, xInputLabelsFile=xOutputSupFile,
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
577 dictltpleDistanceMeasurements=dictFlattenedDistances, abundanceTable=abundanceTable, lsOriginalSampleNames=lsOriginalSampleNames)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
578 return dictSelectedSamplesRet
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
579
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
580 #Two happy path test cases
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
581 def _updatePredictFile(self, xPredictSupFile, xInputLabelsFile, dictltpleDistanceMeasurements, abundanceTable, lsOriginalSampleNames):
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
582 """
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
583 Manages updating the predict file.
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
584
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
585 :param xPredictSupFile: File that has predictions (distances) from the supervised method.
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
586 :type: FileStream or String file path
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
587 :param xInputLabelsFile: File that as input to the supervised methods.
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
588 :type: FileStream or String file path
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
589 :param dictltpleDistanceMeasurements:
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
590 :type: Dictionary of lists of tuples {"labelgroup":[("SampleName",dDistance)...], "labelgroup":[("SampleName",dDistance)...]}
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
591 """
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
592
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
593 if not isinstance(xPredictSupFile, str):
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
594 xPredictSupFile.close()
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
595 xPredictSupFile = xPredictSupFile.name
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
596 csvr = open(xPredictSupFile,'r')
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
597
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
598 f = csv.reader(csvr,delimiter=ConstantsBreadCrumbs.c_strBreadCrumbsSVMSpace)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
599 lsHeader = f.next()[1:]
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
600 dictlltpleRead = dict([(sHeader,[]) for sHeader in lsHeader])
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
601
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
602 #Read data in
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
603 iSampleIndex = 0
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
604 for sRow in f:
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
605 sLabel = sRow[0]
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
606 [dictlltpleRead[lsHeader[iDistanceIndex]].append((lsOriginalSampleNames[iSampleIndex],dDistance)) for iDistanceIndex, dDistance in enumerate(sRow[1:])
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
607 if not dDistance == ConstantsMicropita.c_sEmptyPredictFileValue]
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
608 iSampleIndex += 1
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
609
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
610 #Combine dictltpleDistanceMeasurements with new data
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
611 #If they share a key then merge keeping parameter data
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
612 #If they do not share the key, keep the full data
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
613 dictNew = {}
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
614 for sKey in dictltpleDistanceMeasurements.keys():
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
615 lsSamples = [tple[0] for tple in dictltpleDistanceMeasurements[sKey]]
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
616 dictNew[sKey] = dictltpleDistanceMeasurements[sKey]+[tple for tple in dictlltpleRead[sKey] if tple[0] not in lsSamples] if sKey in dictlltpleRead.keys() else dictltpleDistanceMeasurements[sKey]
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
617 for sKey in dictlltpleRead:
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
618 if sKey not in dictltpleDistanceMeasurements.keys():
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
619 dictNew[sKey] = dictlltpleRead[sKey]
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
620
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
621 #Call writer
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
622 self._writeToPredictFile(xPredictSupFile=xPredictSupFile, xInputLabelsFile=xInputLabelsFile,
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
623 dictltpleDistanceMeasurements=dictNew, abundanceTable=abundanceTable,
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
624 lsOriginalSampleNames=lsOriginalSampleNames, fFromUpdate=True)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
625
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
626 #2 happy path test cases
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
627 def _writeToPredictFile(self, xPredictSupFile, xInputLabelsFile, dictltpleDistanceMeasurements, abundanceTable, lsOriginalSampleNames, fFromUpdate=False):
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
628 """
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
629 Write to the predict file.
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
630
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
631 :param xPredictSupFile: File that has predictions (distances) from the supervised method.
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
632 :type: FileStream or String file path
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
633 :param xInputLabelsFile: File that as input to the supervised methods.
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
634 :type: FileStream or String file path
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
635 :param dictltpleDistanceMeasurements:
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
636 :type: Dictionary of lists of tuples {"labelgroup":[("SampleName",dDistance)...], "labelgroup":[("SampleName",dDistance)...]}
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
637 :param abundanceTable: An abundance table of the sample data.
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
638 :type: AbundanceTable
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
639 :param lsOriginalSampleNames: Used if the file is being updated as the sample names so that it may be passed in and consistent with other writing.
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
640 Otherwise will use the sample names from the abundance table.
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
641 :type: List of strings
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
642 :param fFromUpdate: Indicates if this is part of an update to the file or not.
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
643 :type: Boolean
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
644 """
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
645
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
646 xInputLabelsFileName = xInputLabelsFile
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
647 if not isinstance(xInputLabelsFile,str):
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
648 xInputLabelsFileName = xInputLabelsFile.name
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
649 f = csv.writer(open(xPredictSupFile,"w") if isinstance(xPredictSupFile, str) else xPredictSupFile,delimiter=ConstantsBreadCrumbs.c_strBreadCrumbsSVMSpace)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
650
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
651 lsAllSampleNames = abundanceTable.funcGetSampleNames()
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
652 lsLabels = SVM.funcReadLabelsFromFile(xSVMFile=xInputLabelsFileName, lsAllSampleNames= lsOriginalSampleNames if fFromUpdate else lsAllSampleNames,
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
653 isPredictFile=False)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
654 dictLabels = dict([(sSample,sLabel) for sLabel in lsLabels.keys() for sSample in lsLabels[sLabel]])
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
655
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
656 #Dictionay keys will be used to order the predict file
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
657 lsMeasurementKeys = dictltpleDistanceMeasurements.keys()
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
658 #Make header
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
659 f.writerow(["labels"]+lsMeasurementKeys)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
660
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
661 #Reformat dictionary to make it easier to use
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
662 for sKey in dictltpleDistanceMeasurements:
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
663 dictltpleDistanceMeasurements[sKey] = dict([ltpl for ltpl in dictltpleDistanceMeasurements[sKey]])
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
664
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
665 for sSample in lsOriginalSampleNames:
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
666 #Make body of file
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
667 f.writerow([dictLabels.get(sSample,ConstantsMicropita.c_sEmptyPredictFileValue)]+
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
668 [str(dictltpleDistanceMeasurements[sKey].get(sSample,ConstantsMicropita.c_sEmptyPredictFileValue))
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
669 for sKey in lsMeasurementKeys])
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
670
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
671 def _funcRunNormalizeSensitiveMethods(self, abndData, iSampleSelectionCount, dictSelectedSamples, lsAlphaMetrics, lsBetaMetrics, lsInverseBetaMetrics,
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
672 fRunDiversity, fRunRepresentative, fRunExtreme, strAlphaMetadata=None,
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
673 istmBetaMatrix=None, istrmTree=None, istrmEnvr=None, fInvertDiversity=False):
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
674 """
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
675 Manages running methods that are sensitive to normalization. This is called twice, once for the set of methods which should not be normalized and the other
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
676 for the set that should be normalized.
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
677
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
678 :param abndData: Abundance table object holding the samples to be measured.
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
679 :type: AbundanceTable
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
680 :param iSampleSelectionCount The number of samples to select per method.
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
681 :type: Integer
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
682 :param dictSelectedSamples Will be added to as samples are selected {"Method:["strSelectedSampleID","strSelectedSampleID"...]}.
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
683 :type: Dictionary
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
684 :param lsAlphaMetrics: List of alpha metrics to use on alpha metric dependent assays (like highest diversity).
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
685 :type: List of strings
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
686 :param lsBetaMetrics: List of beta metrics to use on beta metric dependent assays (like most representative).
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
687 :type: List of strings
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
688 :param lsInverseBetaMetrics: List of inverse beta metrics to use on inverse beta metric dependent assays (like most dissimilar).
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
689 :type: List of strings
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
690 :param fRunDiversity: Run Diversity based methods (true indicates run).
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
691 :type: Boolean
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
692 :param fRunRepresentative: Run Representative based methods (true indicates run).
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
693 :type: Boolean
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
694 :param fRunExtreme: Run Extreme based methods (true indicates run).
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
695 :type: Boolean
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
696 :param istmBetaMatrix: File that has a precalculated beta matrix
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
697 :type: File stream or File path string
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
698 :return Selected Samples: Samples selected by methods.
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
699 Dictionary {"Selection Method":["SampleID","SampleID","SampleID",...]}
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
700 """
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
701
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
702 #Sample ids/names
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
703 lsSampleNames = abndData.funcGetSampleNames()
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
704
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
705 #Generate alpha metrics and get most diverse
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
706 if fRunDiversity:
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
707
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
708 #Get Alpha metrics matrix
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
709 internalAlphaMatrix = None
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
710 #Name of technique
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
711 strMethod = [strAlphaMetadata] if strAlphaMetadata else lsAlphaMetrics
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
712
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
713 #If given an alpha-diversity metadata
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
714 if strAlphaMetadata:
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
715 internalAlphaMatrix = [[float(strNum) for strNum in abndData.funcGetMetadata(strAlphaMetadata)]]
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
716 else:
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
717 #Expects Observations (Taxa (row) x sample (column))
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
718 #Returns [[metric1-sample1, metric1-sample2, metric1-sample3],[metric1-sample1, metric1-sample2, metric1-sample3]]
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
719 internalAlphaMatrix = Metric.funcBuildAlphaMetricsMatrix(npaSampleAbundance = abndData.funcGetAbundanceCopy()
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
720 if not abndData.funcIsSummed()
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
721 else abndData.funcGetFeatureAbundanceTable(abndData.funcGetTerminalNodes()).funcGetAbundanceCopy(),
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
722 lsSampleNames = lsSampleNames, lsDiversityMetricAlpha = lsAlphaMetrics)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
723
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
724 if internalAlphaMatrix:
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
725 #Invert measurments
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
726 if fInvertDiversity:
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
727 lldNewDiversity = []
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
728 for lsLine in internalAlphaMatrix:
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
729 lldNewDiversity.append([1/max(dValue,ConstantsMicropita.c_smallNumber) for dValue in lsLine])
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
730 internalAlphaMatrix = lldNewDiversity
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
731 #Get top ranked alpha diversity by most diverse
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
732 #Expects [[sample1,sample2,sample3...],[sample1,sample2,sample3..],...]
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
733 #Returns [[sampleName1, sampleName2, sampleNameN],[sampleName1, sampleName2, sampleNameN]]
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
734 mostDiverseAlphaSamplesIndexes = self.funcGetTopRankedSamples(lldMatrix=internalAlphaMatrix, lsSampleNames=lsSampleNames, iTopAmount=iSampleSelectionCount)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
735
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
736 #Add to results
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
737 for index in xrange(0,len(strMethod)):
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
738 strSelectionMethod = self.dictConvertAMetricDiversity.get(strMethod[index],ConstantsMicropita.c_strDiversity+"="+strMethod[index])
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
739 dictSelectedSamples.setdefault(strSelectionMethod,[]).extend(mostDiverseAlphaSamplesIndexes[index])
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
740
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
741 logging.info("MicroPITA.funcRunNormalizeSensitiveMethods:: Selected Samples 1b")
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
742 logging.info(dictSelectedSamples)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
743
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
744 #Generate beta metrics and
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
745 if fRunRepresentative or fRunExtreme:
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
746
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
747 #Abundance matrix transposed
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
748 npaTransposedAbundance = UtilityMath.funcTransposeDataMatrix(abndData.funcGetAbundanceCopy(), fRemoveAdornments=True)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
749
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
750 #Get center selection using clusters/tiling
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
751 #This will be for beta metrics in normalized space
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
752 if fRunRepresentative:
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
753
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
754 if istmBetaMatrix:
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
755 #Get representative dissimilarity samples
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
756 medoidSamples=self.funcGetCentralSamplesByKMedoids(npaMatrix=npaTransposedAbundance, sMetric=ConstantsMicropita.c_custom, lsSampleNames=lsSampleNames, iNumberSamplesReturned=iSampleSelectionCount, istmBetaMatrix=istmBetaMatrix, istrmTree=istrmTree, istrmEnvr=istrmEnvr)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
757
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
758 if medoidSamples:
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
759 dictSelectedSamples.setdefault(ConstantsMicropita.c_strRepresentative+"="+ConstantsMicropita.c_custom,[]).extend(medoidSamples)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
760 else:
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
761 logging.info("MicroPITA.funcRunNormalizeSensitiveMethods:: Performing representative selection on normalized data.")
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
762 for bMetric in lsBetaMetrics:
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
763
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
764 #Get representative dissimilarity samples
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
765 medoidSamples=self.funcGetCentralSamplesByKMedoids(npaMatrix=npaTransposedAbundance, sMetric=bMetric, lsSampleNames=lsSampleNames, iNumberSamplesReturned=iSampleSelectionCount, istmBetaMatrix=istmBetaMatrix, istrmTree=istrmTree, istrmEnvr=istrmEnvr)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
766
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
767 if medoidSamples:
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
768 dictSelectedSamples.setdefault(self.dictConvertBMetricToMethod.get(bMetric,ConstantsMicropita.c_strRepresentative+"="+bMetric),[]).extend(medoidSamples)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
769
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
770 #Get extreme selection using clusters, tiling
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
771 if fRunExtreme:
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
772 logging.info("MicroPITA.funcRunNormalizeSensitiveMethods:: Performing extreme selection on normalized data.")
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
773 if istmBetaMatrix:
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
774
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
775 #Samples for representative dissimilarity
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
776 #This involves inverting the distance metric,
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
777 #Taking the dendrogram level of where the number cluster == the number of samples to select
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
778 #Returning a repersentative sample from each cluster
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
779 extremeSamples = self.funcSelectExtremeSamplesFromHClust(strBetaMetric=ConstantsMicropita.c_custom, npaAbundanceMatrix=npaTransposedAbundance, lsSampleNames=lsSampleNames, iSelectSampleCount=iSampleSelectionCount, istmBetaMatrix=istmBetaMatrix, istrmTree=istrmTree, istrmEnvr=istrmEnvr)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
780
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
781 #Add selected samples
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
782 if extremeSamples:
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
783 dictSelectedSamples.setdefault(ConstantsMicropita.c_strExtreme+"="+ConstantsMicropita.c_custom,[]).extend(extremeSamples)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
784
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
785 else:
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
786 #Run KMedoids with inverse custom distance metric in normalized space
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
787 for bMetric in lsInverseBetaMetrics:
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
788
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
789 #Samples for representative dissimilarity
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
790 #This involves inverting the distance metric,
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
791 #Taking the dendrogram level of where the number cluster == the number of samples to select
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
792 #Returning a repersentative sample from each cluster
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
793 extremeSamples = self.funcSelectExtremeSamplesFromHClust(strBetaMetric=bMetric, npaAbundanceMatrix=npaTransposedAbundance, lsSampleNames=lsSampleNames, iSelectSampleCount=iSampleSelectionCount, istmBetaMatrix=istmBetaMatrix, istrmTree=istrmTree, istrmEnvr=istrmEnvr)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
794
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
795 #Add selected samples
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
796 if extremeSamples:
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
797 dictSelectedSamples.setdefault(self.dictConvertInvBMetricToMethod.get(bMetric,ConstantsMicropita.c_strExtreme+"="+bMetric),[]).extend(extremeSamples)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
798
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
799 logging.info("MicroPITA.funcRunNormalizeSensitiveMethods:: Selected Samples 2,3b")
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
800 logging.info(dictSelectedSamples)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
801 return dictSelectedSamples
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
802
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
803 def funcRun(self, strIDName, strLastMetadataName, istmInput,
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
804 ostmInputPredictFile, ostmPredictFile, ostmCheckedFile, ostmOutput,
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
805 cDelimiter, cFeatureNameDelimiter, strFeatureSelection,
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
806 istmFeatures, iCount, lstrMethods, strLastRowMetadata = None, strLabel = None, strStratify = None,
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
807 strCustomAlpha = None, strCustomBeta = None, strAlphaMetadata = None, istmBetaMatrix = None, istrmTree = None, istrmEnvr = None,
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
808 iMinSeqs = ConstantsMicropita.c_liOccurenceFilter[0], iMinSamples = ConstantsMicropita.c_liOccurenceFilter[1], fInvertDiversity = False):
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
809 """
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
810 Manages the selection of samples given different metrics.
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
811
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
812 :param strIDName: Sample Id metadata row
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
813 :type: String
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
814 :param strLastMetadataName: The id of the metadata positioned last in the abundance table.
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
815 :type: String String metadata id.
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
816 :param istmInput: File to store input data to supervised methods.
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
817 :type: FileStream of String file path
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
818 :param ostmInputPredictFile: File to store distances from supervised methods.
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
819 :type: FileStream or String file path
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
820 :param ostmCheckedFile: File to store the AbundanceTable data after it is being checked.
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
821 :type: FileStream or String file path
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
822 :param ostmOutPut: File to store sample selection by methods of interest.
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
823 :type: FileStream or String file path
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
824 :param cDelimiter: Delimiter of abundance table.
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
825 :type: Character Char (default TAB).
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
826 :param cFeatureNameDelimiter: Delimiter of the name of features (for instance if they contain consensus lineages indicating clades).
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
827 :type: Character (default |).
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
828 :param stFeatureSelectionMethod: Which method to use to select features in a targeted manner (Using average ranked abundance or average abundance).
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
829 :type: String (specific values indicated in ConstantsMicropita.lsTargetedFeatureMethodValues).
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
830 :param istmFeatures: File which holds the features of interest if using targeted feature methodology.
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
831 :type: FileStream or String file path
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
832 :param iCount: Number of samples to select in each methods, supervised methods select this amount per label if possible.
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
833 :type: Integer integer.
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
834 :param lstrMethods: List of strings indicating selection techniques.
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
835 :type: List of string method names
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
836 :param strLabel: The metadata used for supervised labels.
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
837 :type: String
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
838 :param strStratify: The metadata used to stratify unsupervised data.
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
839 :type: String
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
840 :param strCustomAlpha: Custom alpha diversity metric
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
841 :type: String
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
842 :param strCustomBeta: Custom beta diversity metric
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
843 :type: String
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
844 :param strAlphaMetadata: Metadata id which is a diveristy metric to use in highest diversity sampling
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
845 :type: String
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
846 :param istmBetaMatrix: File containing precalculated beta-diversity matrix for representative sampling
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
847 :type: FileStream or String file path
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
848 :param istrmTree: File containing tree for phylogentic beta-diversity analysis
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
849 :type: FileStream or String file path
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
850 :param istrmEnvr: File containing environment for phylogentic beta-diversity analysis
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
851 :type: FileStream or String file path
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
852 :param iMinSeqs: Minimum sequence in the occurence filter which filters all features not with a minimum number of sequences in each of a minimum number of samples.
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
853 :type: Integer
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
854 :param iMinSamples: Minimum sample count for the occurence filter.
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
855 :type: Integer
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
856 :param fInvertDiversity: When true will invert diversity measurements before using.
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
857 :type: boolean
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
858 :return Selected Samples: Samples selected by methods.
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
859 Dictionary {"Selection Method":["SampleID","SampleID","SampleID",...]}
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
860 """
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
861
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
862 #Holds the top ranked samples from different metrics
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
863 #dict[metric name] = [samplename,samplename...]
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
864 selectedSamples = dict()
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
865
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
866 #If a target feature file is given make sure that targeted feature is in the selection methods, if not add
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
867 if ConstantsMicropita.c_strFeature in lstrMethods:
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
868 if not istmFeatures:
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
869 logging.error("MicroPITA.funcRun:: Did not receive both the Targeted feature file and the feature selection method. MicroPITA did not run.")
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
870 return False
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
871
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
872 #Diversity metrics to run
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
873 #Use custom metrics if specified
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
874 #Custom beta metrics set to normalized only, custom alpha metrics set to count only
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
875 diversityMetricsAlpha = [] if strCustomAlpha or strAlphaMetadata else [MicroPITA.c_strInverseSimpsonDiversity]
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
876 diversityMetricsBeta = [] if istmBetaMatrix else [strCustomBeta] if strCustomBeta else [MicroPITA.c_strBrayCurtisDissimilarity]
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
877 # inverseDiversityMetricsBeta = [MicroPITA.c_strInvBrayCurtisDissimilarity]
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
878 diversityMetricsAlphaNoNormalize = [strAlphaMetadata] if strAlphaMetadata else [strCustomAlpha] if strCustomAlpha else []
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
879 diversityMetricsBetaNoNormalize = []
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
880 # inverseDiversityMetricsBetaNoNormalize = []
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
881
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
882 #Targeted taxa
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
883 userDefinedTaxa = []
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
884
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
885 #Perform different flows flags
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
886 c_RUN_MAX_DIVERSITY_1 = ConstantsMicropita.c_strDiversity in lstrMethods
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
887 c_RUN_REPRESENTIVE_DISSIMILARITY_2 = ConstantsMicropita.c_strRepresentative in lstrMethods
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
888 c_RUN_MAX_DISSIMILARITY_3 = ConstantsMicropita.c_strExtreme in lstrMethods
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
889 c_RUN_RANK_AVERAGE_USER_4 = False
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
890 if ConstantsMicropita.c_strFeature in lstrMethods:
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
891 c_RUN_RANK_AVERAGE_USER_4 = True
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
892 if not istmFeatures:
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
893 logging.error("MicroPITA.funcRun:: No taxa file was given for taxa selection.")
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
894 return False
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
895 #Read in taxa list, break down to lines and filter out empty strings
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
896 userDefinedTaxa = filter(None,(s.strip( ) for s in istmFeatures.readlines()))
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
897 c_RUN_RANDOM_5 = ConstantsMicropita.c_strRandom in lstrMethods
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
898 c_RUN_DISTINCT = ConstantsMicropita.c_strDistinct in lstrMethods
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
899 c_RUN_DISCRIMINANT = ConstantsMicropita.c_strDiscriminant in lstrMethods
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
900
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
901 #Read in abundance data
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
902 #Abundance is a structured array. Samples (column) by Taxa (rows) with the taxa id row included as the column index=0
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
903 #Abundance table object to read in and manage data
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
904 totalAbundanceTable = AbundanceTable.funcMakeFromFile(xInputFile=istmInput, lOccurenceFilter = [iMinSeqs, iMinSamples],
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
905 cDelimiter=cDelimiter, sMetadataID=strIDName, sLastMetadataRow=strLastRowMetadata,
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
906 sLastMetadata=strLastMetadataName, cFeatureNameDelimiter=cFeatureNameDelimiter, xOutputFile=ostmCheckedFile)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
907 if not totalAbundanceTable:
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
908 logging.error("MicroPITA.funcRun:: Could not read in the abundance table. Analysis was not performed."+
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
909 " This often occurs when the Last Metadata is not specified correctly."+
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
910 " Please check to make sure the Last Metadata selection is the row of the last metadata,"+
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
911 " all values after this selection should be microbial measurements and should be numeric.")
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
912 return False
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
913
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
914 lsOriginalLabels = SVM.funcMakeLabels(totalAbundanceTable.funcGetMetadata(strLabel)) if strLabel else strLabel
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
915
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
916 dictTotalMetadata = totalAbundanceTable.funcGetMetadataCopy()
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
917 logging.debug("MicroPITA.funcRun:: Received metadata=" + str(dictTotalMetadata))
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
918 #If there is only 1 unique value for the labels, do not run the Supervised methods
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
919 if strLabel and ( len(set(dictTotalMetadata.get(strLabel,[]))) < 2 ):
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
920 logging.error("The label " + strLabel + " did not have 2 or more values. Labels found=" + str(dictTotalMetadata.get(strLabel,[])))
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
921 return False
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
922
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
923 #Run unsupervised methods###
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
924 #Stratify the data if need be and drop the old data
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
925 lStratifiedAbundanceTables = totalAbundanceTable.funcStratifyByMetadata(strStratify) if strStratify else [totalAbundanceTable]
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
926
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
927 #For each stratified abundance block or for the unstratfified abundance
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
928 #Run the unsupervised blocks
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
929 fAppendSupFiles = False
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
930 for stratAbundanceTable in lStratifiedAbundanceTables:
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
931 logging.info("MicroPITA.funcRun:: Running abundance block:"+stratAbundanceTable.funcGetName())
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
932
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
933 ###NOT SUMMED, NOT NORMALIZED
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
934 #Only perform if the data is not yet normalized
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
935 if not stratAbundanceTable.funcIsNormalized( ):
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
936 #Need to first work with unnormalized data
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
937 if c_RUN_MAX_DIVERSITY_1 or c_RUN_REPRESENTIVE_DISSIMILARITY_2 or c_RUN_MAX_DISSIMILARITY_3:
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
938
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
939 self._funcRunNormalizeSensitiveMethods(abndData=stratAbundanceTable, iSampleSelectionCount=iCount,
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
940 dictSelectedSamples=selectedSamples, lsAlphaMetrics=diversityMetricsAlphaNoNormalize,
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
941 lsBetaMetrics=diversityMetricsBetaNoNormalize,
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
942 lsInverseBetaMetrics=diversityMetricsBetaNoNormalize,
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
943 fRunDiversity=c_RUN_MAX_DIVERSITY_1,fRunRepresentative=c_RUN_REPRESENTIVE_DISSIMILARITY_2,
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
944 fRunExtreme=c_RUN_MAX_DISSIMILARITY_3, strAlphaMetadata=strAlphaMetadata,
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
945 istrmTree=istrmTree, istrmEnvr=istrmEnvr, fInvertDiversity=fInvertDiversity)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
946
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
947
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
948 #Generate selection by the rank average of user defined taxa
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
949 #Expects (Taxa (row) by Samples (column))
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
950 #Expects a column 0 of taxa id that is skipped
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
951 #Returns [(sample name,average,rank)]
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
952 #SUMMED AND NORMALIZED
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
953 stratAbundanceTable.funcSumClades()
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
954 #Normalize data at this point
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
955 stratAbundanceTable.funcNormalize()
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
956 if c_RUN_RANK_AVERAGE_USER_4:
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
957 selectedSamples[ConstantsMicropita.c_strFeature] = self.funcSelectTargetedTaxaSamples(abndMatrix=stratAbundanceTable,
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
958 lsTargetedTaxa=userDefinedTaxa, iSampleSelectionCount=iCount, sMethod=strFeatureSelection)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
959 logging.info("MicroPITA.funcRun:: Selected Samples Rank")
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
960 logging.info(selectedSamples)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
961
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
962 ###SUMMED AND NORMALIZED analysis block
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
963 #Diversity based metric will move reduce to terminal taxa as needed
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
964 if c_RUN_MAX_DIVERSITY_1 or c_RUN_REPRESENTIVE_DISSIMILARITY_2 or c_RUN_MAX_DISSIMILARITY_3:
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
965
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
966 self._funcRunNormalizeSensitiveMethods(abndData=stratAbundanceTable, iSampleSelectionCount=iCount,
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
967 dictSelectedSamples=selectedSamples, lsAlphaMetrics=diversityMetricsAlpha,
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
968 lsBetaMetrics=diversityMetricsBeta,
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
969 lsInverseBetaMetrics=diversityMetricsBeta,
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
970 fRunDiversity=c_RUN_MAX_DIVERSITY_1,fRunRepresentative=c_RUN_REPRESENTIVE_DISSIMILARITY_2,
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
971 fRunExtreme=c_RUN_MAX_DISSIMILARITY_3,
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
972 istmBetaMatrix=istmBetaMatrix, istrmTree=istrmTree, istrmEnvr=istrmEnvr, fInvertDiversity=fInvertDiversity)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
973
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
974 #5::Select randomly
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
975 #Expects sampleNames = List of sample names [name, name, name...]
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
976 if(c_RUN_RANDOM_5):
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
977 #Select randomly from sample names
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
978 selectedSamples[ConstantsMicropita.c_strRandom] = self.funcGetRandomSamples(lsSamples=stratAbundanceTable.funcGetSampleNames(), iNumberOfSamplesToReturn=iCount)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
979 logging.info("MicroPITA.funcRun:: Selected Samples Random")
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
980 logging.info(selectedSamples)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
981
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
982 #Perform supervised selection
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
983 if c_RUN_DISTINCT or c_RUN_DISCRIMINANT:
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
984 if strLabel:
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
985 dictSelectionRet = self.funcRunSupervisedDistancesFromCentroids(abundanceTable=stratAbundanceTable,
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
986 fRunDistinct=c_RUN_DISTINCT, fRunDiscriminant=c_RUN_DISCRIMINANT,
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
987 xOutputSupFile=ostmInputPredictFile,xPredictSupFile=ostmPredictFile,
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
988 strSupervisedMetadata=strLabel, iSampleSupSelectionCount=iCount,
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
989 lsOriginalSampleNames = totalAbundanceTable.funcGetSampleNames(),
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
990 lsOriginalLabels = lsOriginalLabels,
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
991 fAppendFiles=fAppendSupFiles)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
992
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
993 [selectedSamples.setdefault(sKey,[]).extend(lValue) for sKey,lValue in dictSelectionRet.items()]
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
994
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
995 if not fAppendSupFiles:
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
996 fAppendSupFiles = True
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
997 logging.info("MicroPITA.funcRun:: Selected Samples Unsupervised")
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
998 logging.info(selectedSamples)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
999 return selectedSamples
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1000
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1001 #Testing: Happy path tested
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1002 @staticmethod
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1003 def funcWriteSelectionToFile(dictSelection,xOutputFilePath):
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1004 """
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1005 Writes the selection of samples by method to an output file.
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1006
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1007 :param dictSelection: The dictionary of selections by method to be written to a file.
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1008 :type: Dictionary The dictionary of selections by method {"method":["sample selected","sample selected"...]}
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1009 :param xOutputFilePath: FileStream or String path to file inwhich the dictionary is written.
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1010 :type: String FileStream or String path to file
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1011 """
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1012
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1013 if not dictSelection:
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1014 return
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1015
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1016 #Open file
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1017 f = csv.writer(open(xOutputFilePath,"w") if isinstance(xOutputFilePath, str) else xOutputFilePath, delimiter=ConstantsMicropita.c_outputFileDelim )
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1018
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1019 #Create output content from dictionary
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1020 for sKey in dictSelection:
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1021 f.writerow([sKey]+dictSelection[sKey])
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1022 logging.debug("MicroPITA.funcRun:: Selected samples output to file:"+str(dictSelection[sKey]))
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1023
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1024 #Testing: Happy Path tested
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1025 @staticmethod
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1026 def funcReadSelectionFileToDictionary(xInputFile):
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1027 """
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1028 Reads in an output selection file from micropita and formats it into a dictionary.
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1029
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1030 :param xInputFile: String path to file or file stream to read and translate into a dictionary.
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1031 {"method":["sample selected","sample selected"...]}
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1032 :type: FileStream or String Path to file
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1033 :return Dictionary: Samples selected by methods.
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1034 Dictionary {"Selection Method":["SampleID","SampleID","SampleID",...]}
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1035 """
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1036
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1037 #Open file
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1038 istmReader = csv.reader(open(xInputFile,'r') if isinstance(xInputFile, str) else xInputFile, delimiter = ConstantsMicropita.c_outputFileDelim)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1039
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1040 #Dictionary to hold selection data
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1041 return dict([(lsLine[0], lsLine[1:]) for lsLine in istmReader])
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1042
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1043 #Set up arguments reader
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1044 argp = argparse.ArgumentParser( prog = "MicroPITA.py",
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1045 description = """Selects samples from abundance tables based on various selection schemes.""" )
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1046
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1047 args = argp.add_argument_group( "Common", "Commonly modified options" )
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1048 args.add_argument(ConstantsMicropita.c_strCountArgument,"--num", dest="iCount", metavar = "samples", default = 10, type = int, help = ConstantsMicropita.c_strCountHelp)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1049 args.add_argument("-m","--method", dest = "lstrMethods", metavar = "method", default = [], help = ConstantsMicropita.c_strSelectionTechniquesHelp,
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1050 choices = ConstantsMicropita.c_lsAllMethods, action = "append")
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1051
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1052 args = argp.add_argument_group( "Custom", "Selecting and inputing custom metrics" )
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1053 args.add_argument("-a","--alpha", dest = "strAlphaDiversity", metavar = "AlphaDiversity", default = None, help = ConstantsMicropita.c_strCustomAlphaDiversityHelp, choices = Metric.setAlphaDiversities)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1054 args.add_argument("-b","--beta", dest = "strBetaDiversity", metavar = "BetaDiversity", default = None, help = ConstantsMicropita.c_strCustomBetaDiversityHelp, choices = list(Metric.setBetaDiversities)+[Metric.c_strUnifracUnweighted,Metric.c_strUnifracWeighted])
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1055 args.add_argument("-q","--alphameta", dest = "strAlphaMetadata", metavar = "AlphaDiversityMetadata", default = None, help = ConstantsMicropita.c_strCustomAlphaDiversityMetadataHelp)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1056 args.add_argument("-x","--betamatrix", dest = "istmBetaMatrix", metavar = "BetaDiversityMatrix", default = None, help = ConstantsMicropita.c_strCustomBetaDiversityMatrixHelp)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1057 args.add_argument("-o","--tree", dest = "istrmTree", metavar = "PhylogeneticTree", default = None, help = ConstantsMicropita.c_strCustomPhylogeneticTreeHelp)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1058 args.add_argument("-i","--envr", dest = "istrmEnvr", metavar = "EnvironmentFile", default = None, help = ConstantsMicropita.c_strCustomEnvironmentFileHelp)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1059 args.add_argument("-f","--invertDiversity", dest = "fInvertDiversity", action="store_true", default = False, help = ConstantsMicropita.c_strInvertDiversityHelp)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1060
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1061 args = argp.add_argument_group( "Miscellaneous", "Row/column identifiers and feature targeting options" )
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1062 args.add_argument("-d",ConstantsMicropita.c_strIDNameArgument, dest="strIDName", metavar="sample_id", help= ConstantsMicropita.c_strIDNameHelp)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1063 args.add_argument("-l",ConstantsMicropita.c_strLastMetadataNameArgument, dest="strLastMetadataName", metavar = "metadata_id", default = None,
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1064 help= ConstantsMicropita.c_strLastMetadataNameHelp)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1065 args.add_argument("-r",ConstantsMicropita.c_strTargetedFeatureMethodArgument, dest="strFeatureSelection", metavar="targeting_method", default=ConstantsMicropita.lsTargetedFeatureMethodValues[0],
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1066 choices=ConstantsMicropita.lsTargetedFeatureMethodValues, help= ConstantsMicropita.c_strTargetedFeatureMethodHelp)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1067 args.add_argument("-t",ConstantsMicropita.c_strTargetedSelectionFileArgument, dest="istmFeatures", metavar="feature_file", type=argparse.FileType("rU"), help=ConstantsMicropita.c_strTargetedSelectionFileHelp)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1068 args.add_argument("-w",ConstantsMicropita.c_strFeatureMetadataArgument, dest="strLastFeatureMetadata", metavar="Last_Feature_Metadata", default=None, help=ConstantsMicropita.c_strFeatureMetadataHelp)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1069
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1070 args = argp.add_argument_group( "Data labeling", "Metadata IDs for strata and supervised label values" )
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1071 args.add_argument("-e",ConstantsMicropita.c_strSupervisedLabelArgument, dest="strLabel", metavar= "supervised_id", help=ConstantsMicropita.c_strSupervisedLabelHelp)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1072 args.add_argument("-s",ConstantsMicropita.c_strUnsupervisedStratifyMetadataArgument, dest="strUnsupervisedStratify", metavar="stratify_id",
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1073 help= ConstantsMicropita.c_strUnsupervisedStratifyMetadataHelp)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1074
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1075 args = argp.add_argument_group( "File formatting", "Rarely modified file formatting options" )
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1076 args.add_argument("-j",ConstantsMicropita.c_strFileDelimiterArgument, dest="cFileDelimiter", metavar="column_delimiter", default="\t", help=ConstantsMicropita.c_strFileDelimiterHelp)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1077 args.add_argument("-k",ConstantsMicropita.c_strFeatureNameDelimiterArgument, dest="cFeatureNameDelimiter", metavar="taxonomy_delimiter", default="|", help=ConstantsMicropita.c_strFeatureNameDelimiterHelp)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1078
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1079 args = argp.add_argument_group( "Debugging", "Debugging options - modify at your own risk!" )
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1080 args.add_argument("-v",ConstantsMicropita.c_strLoggingArgument, dest="strLogLevel", metavar = "log_level", default="WARNING",
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1081 choices=ConstantsMicropita.c_lsLoggingChoices, help= ConstantsMicropita.c_strLoggingHelp)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1082 args.add_argument("-c",ConstantsMicropita.c_strCheckedAbundanceFileArgument, dest="ostmCheckedFile", metavar = "output_qc", type = argparse.FileType("w"), help = ConstantsMicropita.c_strCheckedAbundanceFileHelp)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1083 args.add_argument("-g",ConstantsMicropita.c_strLoggingFileArgument, dest="ostmLoggingFile", metavar = "output_log", type = argparse.FileType("w"), help = ConstantsMicropita.c_strLoggingFileHelp)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1084 args.add_argument("-u",ConstantsMicropita.c_strSupervisedInputFile, dest="ostmInputPredictFile", metavar = "output_scaled", type = argparse.FileType("w"), help = ConstantsMicropita.c_strSupervisedInputFileHelp)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1085 args.add_argument("-p",ConstantsMicropita.c_strSupervisedPredictedFile, dest="ostmPredictFile", metavar = "output_labels", type = argparse.FileType("w"), help = ConstantsMicropita.c_strSupervisedPredictedFileHelp)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1086
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1087 argp.add_argument("istmInput", metavar = "input.pcl/biome", type = argparse.FileType("rU"), help = ConstantsMicropita.c_strAbundanceFileHelp,
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1088 default = sys.stdin)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1089 argp.add_argument("ostmOutput", metavar = "output.txt", type = argparse.FileType("w"), help = ConstantsMicropita.c_strGenericOutputDataFileHelp,
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1090 default = sys.stdout)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1091
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1092 __doc__ = "::\n\n\t" + argp.format_help( ).replace( "\n", "\n\t" ) + __doc__
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1093
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1094 def _main( ):
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1095 args = argp.parse_args( )
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1096
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1097 #Set up logger
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1098 iLogLevel = getattr(logging, args.strLogLevel.upper(), None)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1099 logging.basicConfig(stream = args.ostmLoggingFile if args.ostmLoggingFile else sys.stderr, filemode = 'w', level=iLogLevel)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1100
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1101 #Run micropita
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1102 logging.info("MicroPITA:: Start microPITA")
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1103 microPITA = MicroPITA()
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1104
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1105 #Argparse will append to the default but will not remove the default so I do this here
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1106 if not len(args.lstrMethods):
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1107 args.lstrMethods = [ConstantsMicropita.c_strRepresentative]
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1108
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1109 dictSelectedSamples = microPITA.funcRun(
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1110 strIDName = args.strIDName,
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1111 strLastMetadataName = args.strLastMetadataName,
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1112 istmInput = args.istmInput,
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1113 ostmInputPredictFile = args.ostmInputPredictFile,
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1114 ostmPredictFile = args.ostmPredictFile,
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1115 ostmCheckedFile = args.ostmCheckedFile,
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1116 ostmOutput = args.ostmOutput,
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1117 cDelimiter = args.cFileDelimiter,
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1118 cFeatureNameDelimiter = args.cFeatureNameDelimiter,
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1119 istmFeatures = args.istmFeatures,
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1120 strFeatureSelection = args.strFeatureSelection,
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1121 iCount = args.iCount,
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1122 strLastRowMetadata = args.strLastFeatureMetadata,
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1123 strLabel = args.strLabel,
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1124 strStratify = args.strUnsupervisedStratify,
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1125 strCustomAlpha = args.strAlphaDiversity,
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1126 strCustomBeta = args.strBetaDiversity,
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1127 strAlphaMetadata = args.strAlphaMetadata,
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1128 istmBetaMatrix = args.istmBetaMatrix,
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1129 istrmTree = args.istrmTree,
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1130 istrmEnvr = args.istrmEnvr,
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1131 lstrMethods = args.lstrMethods,
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1132 fInvertDiversity = args.fInvertDiversity
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1133 )
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1134
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1135 if not dictSelectedSamples:
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1136 logging.error("MicroPITA:: Error, did not get a result from analysis.")
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1137 return -1
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1138 logging.info("End microPITA")
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1139
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1140 #Log output for debugging
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1141 logging.debug("MicroPITA:: Returned the following samples:"+str(dictSelectedSamples))
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1142
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1143 #Write selection to file
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1144 microPITA.funcWriteSelectionToFile(dictSelection=dictSelectedSamples, xOutputFilePath=args.ostmOutput)
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1145
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1146 if __name__ == "__main__":
2f4f6f08c8c4 Uploaded
george-weingart
parents:
diff changeset
1147 _main( )