aboutsummaryrefslogtreecommitdiff
path: root/web/webqtl/compareCorrelates/trait.py
blob: ff1f811945d803193c11ae29b347e030f6c0e87a (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
#Trait.py
#
#--Individual functions are already annotated, more or less.
#
#Classes:
#RawPoint
#Trait
#ProbeSetTrait
#GenotypeTrait
#PublishTrait
#TempTrait
#-KA

# trait.py: a data structure to represent a trait
import time
import string

CONFIG_pubMedLinkURL = "http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=%s&dopt=Abstract"

# RawPoint: to store information about the relationship between two particular
# traits
# RawPoint represents directly the input file
class RawPoint:
    def __init__(self, i, j):
        self.i = i
        self.j = j
        
    def __eq__(self, other):
        return (self.i == other.i and
                self.j == other.j and
                self.spearman == other.spearman and
                self.pearson == other.pearson)
    
    def __str__(self):
        return "(%s,%s,%s,%s)" % (self.i, self.j, self.spearman, self.pearson)

def tdEscapeList(cols, align="left"):
    """
    A helper function used by tableRow
    in Trait that will convert a list of strings into a set of
    table cells enclosed by <td>%s</td> tags
    """
    html = ""
    for col in cols:
        html += '<td style="text-align: %s">%s</td>' % (align, col)
    return html

def thEscapeList(cols):
    """
    A helper function used by tableRowHeader
    in Trait that will convert a list of strings into a set of
    table cells enclosed by <td>%s</td> tags
    """
    html = ""
    for col in cols:
        html += "<th>%s</th>" % col
    return html

def commaEscapeList(cols):
    """
    A helper function used by csvHeader and csvRow.
    Really it's just a wrapper for string.join
    """
    return '"' + string.join(cols, '","') + '"'


class Trait:
    """
    A trait represents an attribute of an object. In the WebQTL database, traits are stored
    as ProbeSets; that is, the average values of a set of probes are stored.
    """
    def __init__(self, id="", name="", description="", symbol="", href=""):
        self.id  = id
        self.name = name
	self.dbName = ""
        self.symbol = symbol
        self.href = href
        self.strainData = {}

    def populateDataId(self, cursor, freezeId):
        """
        Retrieve the dataId for trait data corresponding to the given database
        The way to do this depends on the particular type of trait, so we leave implementation
        to subclasses.
        """
        raise NotImplementedError
    
    def populateStrainData(self, cursor):
        """
        Load this trait full of train data corresponding to the data id
        The data id can either come from populateDataId
        or can be set manually by the user of this class.
        Xiaodong added: The way to do this depends on the particular type of trait,
        so we leave implementation to subclasses.

        """
        raise NotImplementedError

    def shortName(self):
        """
        To return a short name for this trait; this name should be
        appropriate for a row or column title
        """
        return self.name

    def nameNoDB(self):
        """
        To return only the short name without the database attached
        """
        strArray = self.shortName().split('::')
        
        return strArray[1]
    
    def datasetName(self):
        """
        To return only the name of the dataset
        """
        strArray = self.shortName().split('::')
        
        return strArray[0].strip()

    def longName(self):
        """
        To return a long name for this trait; this name should be
        appropriate for a key to a table
        """
        return self.shortName()

    def __str__(self):
        return self.shortName()

    def tableRowHelper(self, beforeCols, afterCols, color, thisRow):
        """
        tableRowHelper: (arrayof String) -. String
        To generate a table row to represent this object, appending
        the additional information in beforeCols and afterCols
        to the beginning and the end
        """
        thisRow[0] = '<a href="%s">%s</a>' % (self.traitInfoLink(),
                                              self.name)
        html = '<tr bgcolor="%s">' % color
        html += tdEscapeList(beforeCols + thisRow)
        html += tdEscapeList(afterCols, align="right")
        html += "</tr>"
        
        return html


    def header(self):
        """
        header: (listof String)
        To generate a list of strings describing each piece of data
        returned by row
        """
        raise NotImplementedError

    def row(self):
        """
        row: (listof String)
        To generate a list of strings describing this object. The
        elements of this list should be described by header()
        """
        raise NotImplementedError
    
    def tableRowHeader(self, beforeCols, afterCols, color):
        """
        tableRowHeader: (arrayof String) -> (arrayof String) -> String
        To generate a table row header to represent this object,
        appending the additional information in beforeCols and
        afterCols to the beginning and end
        """
        html = '<tr bgcolor="%s">' % color
        html += thEscapeList(beforeCols + self.header() +
                             afterCols)
        html += "</tr>"
        return html

    def csvHeader(self, beforeCols, afterCols):
        return commaEscapeList(beforeCols + self.header() + afterCols)
    
    def csvRow(self, beforeCols, afterCols):
        return commaEscapeList(beforeCols + self.row() + afterCols)
    
        
    def traitInfoLink(self):
        """
        To build a trait info link to show information about this
        trait. We assume that the database attribute is properly set
        on the hidden form on the page where this link will go.
        """
        return "javascript:showDatabase2('%s','%s','')" % (self.dbName, self.name)

# ProbeSetTrait: a trait with data from a probeset
class ProbeSetTrait(Trait):
    def __init__(self, id="", name="", description="", symbol="", href="",
                 chromosome="", MB="", GeneId=""):
        Trait.__init__(self, id=id, name=name, href=href)
        self.description = description
        self.symbol = symbol
        self.chromosome = chromosome
        self.MB = MB
        self.GeneId = GeneId
        
    def populateDataId(self, cursor, freezeId):
        """
        Look up the data id for this trait given which
        freeze it came from.
        """
        cursor.execute('''
        SELECT
          ProbeSetXRef.DataId
        FROM
          ProbeSetXRef
        WHERE
          ProbeSetId = %s AND
          ProbeSetFreezeId = %s
        ''' % (self.id, freezeId))

        # we hope that there's only one record here
        row = cursor.fetchone()
        self.dataId = row[0]

    #XZ, 03/03/2009: Xiaodong implemented this fuction
    def populateStrainData(self, cursor):
        cursor.execute('''
        SELECT
          ProbeSetData.StrainId,
          ProbeSetData.value
        FROM
          ProbeSetData
        WHERE
          ProbeSetData.Id = %s''' % self.dataId)
        for row in cursor.fetchall():
            self.strainData[int(row[0])] = float(row[1])


    def shortName(self):
        """
        An improved string method that uses the gene symbol where
        we have it
        """
        if self.symbol != "":
            return self.symbol
        else:
            return Trait.shortName(self)

    def longName(self):
        """
        We use several bits of genetic information to give
        useful information about this trait and where it is
        """
        if self.chromosome != "":
            chrPart = " (%s on Chr %s @ %s Mb)" % (self.symbol,
                                                     self.chromosome,
                                                     self.MB)
        else:
            chrPart = ""

        return "%s%s: %s" % (self.name, chrPart, self.description)

    def header(self):
        return ["Name", "Symbol", "Description",
                "Chr", "Position (Mb)"]

    def row(self):
        if type(self.MB) is float:
            MB = "%.2f" % self.MB
        else:
            MB = ""
            
        return [self.name, self.symbol, self.description,
                self.chromosome, MB]
    
    def tableRow(self, beforeCols, afterCols, color):
        """
        tableRow: (arrayof String) -> (arrayof String) -> String
        To generate a table row to represent this object, appending
        the additional information in beforeCols and afterCols to the
        beginning and end
        """
        thisRow = self.row()

        # trim description
        if len(thisRow[2]) > 20:
            thisRow[2] = thisRow[2][:20] + "..."

        # add NCBI info link 
        thisRow[1] = self.ncbiInfoLink()

        return self.tableRowHelper(beforeCols, afterCols, color,
                                   thisRow)


    def ncbiInfoLink(self):
        """
        ncbiInfoLink :: String
        To generate an NCBI info link for this trait. If we have a GeneId,
        then we can go straight to the gene. If not, then we generate a search
        link based on the gene symbol. If we have none of them, then we don't
        generate a link at all.
        """
        if self.GeneId != "":
            cmd = "cmd=Retrieve&dopt=Graphics&list_uids=%s" % self.GeneId
        elif self.symbol != "":
            cmd = "cmd=Search&term=%s" % self.symbol
        else:
            return ""

        return '''
        <a target="_new"
           href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=gene&%s">
           %s</a> ''' % (cmd, self.symbol)


# GenotypeTrait: a trait with data from the genotype
class GenotypeTrait(Trait):
    def __init__(self, id="", name="", href="", chromosome="", MB=""):
        Trait.__init__(self, id=id, name=name, href=href)
        self.chromosome = chromosome
        self.MB = MB

    def populateDataId(self, cursor, freezeId):
        """
        Look up the data id for this trait from the
        genotype.
        """
        cursor.execute('''
        SELECT
          GenoXRef.DataId
        FROM
          GenoXRef
        WHERE
          GenoId = %s AND
          GenoFreezeId = %s
        ''' % (self.id, freezeId))

        # we hope that there's only one record here
        row = cursor.fetchone()
        self.dataId = row[0]

    #XZ, 03/03/2009: Xiaodong implemented this fuction
    def populateStrainData(self, cursor):
        cursor.execute('''
        SELECT
          GenoData.StrainId,
          GenoData.value
        FROM
          GenoData
        WHERE
          GenoData.Id = %s''' % self.dataId)
        for row in cursor.fetchall():
            self.strainData[int(row[0])] = float(row[1])

    def header(self):
        return ["Locus", "Chr", "Position (Mb)"]

    def row(self):
        return [self.name, self.chromosome, "%.3f" % self.MB]

    def tableRow(self, beforeCols, afterCols, color):
        return self.tableRowHelper(beforeCols, afterCols, color, self.row())

# PublishTrait: a trait with data from publications
class PublishTrait(Trait):
    def __init__(self, id="", name="", href="", authors="", title="",
                 phenotype="", year=""):
        Trait.__init__(self, id=id, name=name, href=href)
        self.authors = authors
        self.title = title
        self.phenotype = phenotype
        self.year = year

    def populateDataId(self, cursor, freezeId):
        """
        Look up the data id for this trait from the
        published set. For the moment, we assume that there's
        only one publish freeze.
        """
        cursor.execute('''
        SELECT
          PublishXRef.DataId
        FROM
          PublishXRef, PublishFreeze
        WHERE
          PublishFreeze.Id = %s AND 
          PublishFreeze.InbredSetId = PublishXRef.InbredSetId AND
          PublishXRef.Id = %s 
        ''' % (freezeId, self.id))

        # we hope that there's only one record here
        row = cursor.fetchone()
        self.dataId = row[0]

    #XZ, 03/03/2009: Xiaodong implemented this fuction
    def populateStrainData(self, cursor):
        cursor.execute('''
        SELECT
          PublishData.StrainId,
          PublishData.value
        FROM
          PublishData
        WHERE
          PublishData.Id = %s''' % self.dataId)
        for row in cursor.fetchall():
            self.strainData[int(row[0])] = float(row[1])


    def longName(self):
        """
        A more intelligent string function that uses
        information about the publication from which this trait came
        """
        return "%s: %s by %s" % (self.name, self.title, self.authors)

    def header(self):
        return ["Record", "Phenotype", "Authors", "Year", "URL"]

    def row(self):
        return [self.name,
                self.phenotype,
                self.authors,
                str(self.year),
                ""]
    
    def tableRow(self, beforeCols, afterCols, color):
        """
        tableRow: (arrayof String) -> (arrayof String) ->  String
        To generate a table row to represent this object, appending
        the additional information in beforeCols and afterCols to the
        beginning and end
        """
        thisRow = self.row()

        # for multiple authors, use "et. al" after first two
        authors = thisRow[2].split(",")
        if len(authors) > 2:
            thisRow[2] = string.join(authors[:2], ",") + ", et al"

        # clip phenotype to 20 chars
        if len(thisRow[1]) > 20:
            thisRow[1] = thisRow[1][:20] + "..."

        # add Pub Med URL
        thisRow[4] = '<a href="%s" target="_new">Pub Med</a>' % (CONFIG_pubMedLinkURL % self.href)

        return self.tableRowHelper(beforeCols, afterCols, color,
                                   thisRow)


# TempTrait: a trait with data generate by user and stored in temp table
class TempTrait(Trait):
    def __init__(self, id="", name="", href="", description=""):
        Trait.__init__(self, id=id, name=name, href=href)
        self.description = description

    def populateDataId(self, cursor, freeezeId):
        """
        Look up the data id for this trait from the Temp table, freezeId isn't used, 
        it just for fixing the inherit
        """
        cursor.execute('''
        SELECT
          DataId
        FROM
          Temp
        WHERE
          Id=%s
        ''' % (self.id))

        # we hope that there's only one record here
        row = cursor.fetchone()
        self.dataId = row[0]

    #XZ, 03/03/2009: Xiaodong implemented this fuction
    def populateStrainData(self, cursor):
        cursor.execute('''
        SELECT
          TempData.StrainId,
          TempData.value
        FROM
          TempData
        WHERE
          TempData.Id = %s''' % self.dataId)
        for row in cursor.fetchall():
            self.strainData[int(row[0])] = float(row[1])


    def row(self):
        return [self.id,
                self.name,
                self.description,
                ""]
    

    def longName(self):
        """
        For temp trait, the description always contents whole useful information
        """
        return self.description


# queryGenotypeTraitByName : Cursor -> string -> GenotypeTrait
def queryGenotypeTraitByName(cursor, speciesId, name):
    qry = '''
    SELECT
      Geno.Id,
      Geno.Name,
      Geno.Chr,
      Geno.Mb
    FROM
      Geno
    WHERE
      Geno.SpeciesId = %s and Geno.Name = "%s" ''' % (speciesId, name)

    cursor.execute(qry)
    row = cursor.fetchone()
    return GenotypeTrait(id=row[0], name=row[1],
                         chromosome=row[2], MB=row[3])

# queryPublishTraitByName : Cursor -> string -> PublishTrait
def queryPublishTraitByName(cursor, freezeId, name):
    qry = '''
    SELECT
      PublishXRef.Id,
      Phenotype.Id,
      Publication.Authors,
      Publication.Title,
      Publication.Year,
      Publication.PubMed_ID
    FROM
      Publication, PublishXRef, Phenotype, PublishFreeze
    WHERE
      PublishFreeze.Id = %s AND 
      PublishFreeze.InbredSetId = PublishXRef.InbredSetId AND
      PublishXRef.Id = %s AND 
      PublishXRef.PublicationId = Publication.Id AND
      PublishXRef.PhenotypeId = Phenotype.Id 
      ''' % (freezeId, name)

    cursor.execute(qry)
    if cursor.rowcount == 0:
        return None
    else:
        row = cursor.fetchone()
        
        return PublishTrait(id=row[0], name='%s'%row[0],
                            authors=row[2], title=row[3],
                            year=row[4], href=row[5])


def queryTempTraitByName(cursor, name):
    name=name.strip()
    qry = '''
    SELECT
      Temp.Id,
      Temp.Name,
      Temp.description
    FROM
      Temp
    WHERE
      Temp.Name= "%s"
      ''' % (name)

    cursor.execute(qry)
    if cursor.rowcount == 0:
        return None
    else:
        row = cursor.fetchone()
        return TempTrait(id=row[0], name=row[1], description=row[2], href='')

# queryPopulatedProbeSetTraits: Cursor -> Integer -> dictof Trait
# to retrieve an entire probeset fully populated with data
# this query can take 15+ sec the old way (22,000 traits * 35 strains = half
# a million records)
# so we ask for the data in bulk
#
# cursor should be SSCursor for MySQL so rows are stored on the server side
# and tuples are used
# we explicitly close the cursor here as well
#XZ, 03/04/2009: It seems to me that this function is never be executed.
#XZ: Although it can be called from multitrait.loadDatabase,
#XZ: but the loadDatabase function will not be called
#XZ: if the targetDatabaseType is probeset.
#XZ: The probeset traits of target database are retrieved by execute
#XZ: queryPopulatedProbeSetTraits2 from correlation.calcProbeSetPearsonMatrix
def queryPopulatedProbeSetTraits(cursor, freezeId):
    step1 = time.time()
    traits = queryProbeSetTraits(cursor, freezeId)
    traitDict = {}
    for trait in traits:
        traitDict[trait.id] = trait
        
    step2 = time.time()
    print 
    #XZ, 03/04/2009: Xiaodong changed Data to ProbeSetData
    cursor.execute('''
    SELECT
      ProbeSetXRef.ProbeSetId,
      ProbeSetData.StrainId,
      ProbeSetData.value
    FROM
      ProbeSetXRef
    Left Join ProbeSetData ON
      ProbeSetXRef.DataId = ProbeSetData.Id
    WHERE
      ProbeSetXRef.ProbeSetFreezeId = %s
    ''' % freezeId)

    step3 = time.time()
    totalrows = 0
    somerows = cursor.fetchmany(1000)
    while len(somerows) > 0:
        totalrows += len(somerows)
        for row in somerows:
            # this line of code can execute more than one million times
            traitDict[row[0]].strainData[int(row[1])] = row[2]
        somerows = cursor.fetchmany(1000)

    #cursor.close()
    step4 = time.time()
    
    time1 = step2 - step1
    time2 = step3 - step2
    time3 = step4 - step3
    time4 = step4 - step1
    #print "%f %f %f %f %d rows" % (round(time1, 2),
    #                               round(time2, 2),
    #                               round(time3, 2),
    #                               round(time4, 2),
    #                               totalrows)
    #print "Fetched %d traits" % len(traits)
    return traits


# queryPopulatedProbeSetTraits2: Cursor -> Integer -> dictof Trait
# to retrieve probeset fully populated whose ProbeSetId in a range
# a special ProbeSetId with data
# this query can take 15+ sec the old way (22,000 traits * 35 strains = half
# a million records)
# so we ask for the data in bulk
#
# cursor should be SSCursor for MySQL so rows are stored on the server side
# and tuples are used
# we explicitly close the cursor here as well
def queryPopulatedProbeSetTraits2(cursor, freezeId, ProbeSetId1, ProbeSetId2):
    step1 = time.time()
    traits = queryProbeSetTraits2(cursor, freezeId, ProbeSetId1, ProbeSetId2)
    traitDict = {}
    for trait in traits:
        traitDict[trait.id] = trait

    step2 = time.time()
    print
    #XZ, 03/04/2009: Xiaodong changed Data to ProbeSetData
    cursor.execute('''
    SELECT
      ProbeSetXRef.ProbeSetId,
      ProbeSetData.StrainId,
      ProbeSetData.value
    FROM
      ProbeSetXRef
    Left Join ProbeSetData ON
      ProbeSetXRef.DataId = ProbeSetData.Id
    WHERE
      ProbeSetXRef.ProbeSetFreezeId = %s AND
      ProbeSetXRef.ProbeSetId >= %s AND
      ProbeSetXRef.ProbeSetId <= %s
      ''' % (freezeId, ProbeSetId1, ProbeSetId2))

    step3 = time.time()
    totalrows = 0
    somerows = cursor.fetchmany(1000)
    while len(somerows) > 0:
        totalrows += len(somerows)
        for row in somerows:
            # this line of code can execute more than one million times
            traitDict[row[0]].strainData[int(row[1])] = row[2]
        somerows = cursor.fetchmany(1000)

    #cursor.close()
    step4 = time.time()

    time1 = step2 - step1
    time2 = step3 - step2
    time3 = step4 - step3
    time4 = step4 - step1
    #print "%f %f %f %f %d rows" % (round(time1, 2),
    #                               round(time2, 2),
    #                               round(time3, 2),
    #                               round(time4, 2),
    #                               totalrows)
    #print "Fetched %d traits" % len(traits)
    return traits


# def noneFilter : string or none -> string
# to replace a possible None by an empty string
def noneFilter(x):
    if x is None:
        return ""
    else:
        return x

# queryProbeSetTraits: Cursor -> Integer -> dictof Trait
def queryProbeSetTraits(cursor, freezeId):
    """
    To locate all of the traits in a particular probeset
    """
    qry = '''
    SELECT
      ProbeSet.Id,
      ProbeSet.Name,
      ProbeSet.description,
      ProbeSet.symbol,
      ProbeSet.Chr,
      ProbeSet.Mb,
      ProbeSet.GeneId,
      ProbeSetXRef.DataId
    FROM
      ProbeSet,
      ProbeSetXRef
    WHERE
      ProbeSetXRef.ProbeSetId = ProbeSet.Id AND
      ProbeSetXRef.ProbeSetFreezeId = %s
    ORDER BY ProbeSet.Id
    ''' % freezeId

    cursor.execute(qry)
    rows = cursor.fetchall()
    traits = []

    for row in rows:
        row = map(noneFilter, row)
        trait = ProbeSetTrait(id=row[0], name=row[1],
                              description=row[2],
                              chromosome=row[4],
                              MB=row[5],
                              symbol=row[3],
                              GeneId=row[6])
        trait.dataId = row[7]
        traits.append(trait)

    return traits


# queryProbeSetTraits2: Cursor -> Integer -> dictof Trait
def queryProbeSetTraits2(cursor, freezeId, ProbeSetId1, ProbeSetId2):
    """
    To locate all of the traits in a particular probeset
    """
    qry = '''
    SELECT
      ProbeSet.Id,
      ProbeSet.Name,
      ProbeSet.description,
      ProbeSet.symbol,
      ProbeSet.Chr,
      ProbeSet.Mb,
      ProbeSet.GeneId,
      ProbeSetXRef.DataId
    FROM
      ProbeSet,
      ProbeSetXRef
    WHERE
      ProbeSetXRef.ProbeSetId = ProbeSet.Id AND
      ProbeSetXRef.ProbeSetFreezeId = %s AND
      ProbeSet.Id >= %s AND
      ProbeSet.Id <= %s
    ORDER BY ProbeSet.Id
    ''' % (freezeId, ProbeSetId1, ProbeSetId2)

    cursor.execute(qry)
    rows = cursor.fetchall()
    traits = []

    for row in rows:
        row = map(noneFilter, row)
        trait = ProbeSetTrait(id=row[0], name=row[1],
                              description=row[2],
                              chromosome=row[4],
                              MB=row[5],
                              symbol=row[3],
                              GeneId=row[6])
        trait.dataId = row[7]
        traits.append(trait)

    return traits


# queryPublishTraits : Cursor -> arrayof Trait
def queryPublishTraits(cursor, freezeId):
    """
    To locate all published traits
    """
    qry = '''
    SELECT
      Publication.Id,
      Publication.Name,
      Publication.PhenoType,
      Publication.Authors,
      Publication.Title,
      Publication.Year,
      Publication.PubMed_ID,
      PublishXRef.DataId
    FROM
      Publication,
      PublishXRef
    WHERE
      PublishXRef.PublishFreezeId = %s AND
      PublishXRef.PublishId = Publication.Id
    ''' % freezeId
    
    qry = '''
    SELECT
      Publication.Id,
      PublishXRef.Id,
      Phenotype.Pre_publication_description,
      Phenotype.Post_publication_description,
      Publication.Authors,
      Publication.Title,
      Publication.Year,
      Publication.PubMed_ID,
      PublishXRef.DataId
    FROM
      Publication, PublishXRef, Phenotype, PublishFreeze
    WHERE
      PublishFreeze.Id = %s AND 
      PublishFreeze.InbredSetId = PublishXRef.InbredSetId AND
      PublishXRef.PublicationId = Publication.Id AND
      PublishXRef.PhenotypeId = Phenotype.Id 
      ''' % freezeId
    cursor.execute(qry)
    rows = cursor.fetchall()
    traits = []
    for row in rows:
        PhenotypeString = row[3]
        if not row[7] and row[2]:
            PhenotypeString = row[2]
        trait = PublishTrait(id=row[0], name= '%s' %row[1],
                             phenotype=PhenotypeString,
                             authors=row[4],
                             title=row[5],
                             year=row[6],
                             href=row[7])
        trait.dataId = row[8]
        traits.append(trait)
        
    return traits

# queryGenotypeTraits : Cursor -> arrayof Trait
def queryGenotypeTraits(cursor, freezeId):
    """
    To locate all traits in the genotype
    """
    qry =    '''
    SELECT
      Geno.Id,
      Geno.Name,
      Geno.Chr,
      GenoXRef.DataId,
      Geno.Mb
    FROM
      Geno,
      GenoXRef
    WHERE
      GenoXRef.GenoId = Geno.Id
      AND GenoXRef.GenoFreezeId = %s
    ''' % freezeId
    cursor.execute(qry)
    rows = cursor.fetchall()
    traits = []
    
    for row in rows:
        trait = GenotypeTrait(id=row[0], name=row[1],
                              chromosome=row[2], MB=row[4])
        trait.dataId = row[3]
        traits.append(trait)
        
    return traits

# queryProbeSetTraitByName : Cursor -> string -> Trait
# to find a particular trait given its name 
def queryProbeSetTraitByName(cursor, name):
    qry = '''
    SELECT
      ProbeSet.Id,
      ProbeSet.Name,
      ProbeSet.description,
      ProbeSet.symbol,
      ProbeSet.Chr,
      ProbeSet.Mb,
      ProbeSet.GeneId
    FROM
      ProbeSet
    WHERE
      ProbeSet.Name = "%s"
    ''' % name
    #print qry
    cursor.execute(qry)
    row = cursor.fetchone()

    # convert a MySQL NULL value to an empty string
    # for gene symbol
    if row[3] is None:
        sym = ""
    else:
        sym = row[3]
        
    return ProbeSetTrait(id=row[0], name=row[1], description=row[2],
                         symbol=sym, chromosome=row[4], MB=row[5],
                         GeneId=row[6])
    
                     
# queryTraits : Cursor -> string -> string -> arrayof Traits
# to find all of the traits whose descriptions match a certain string in a
# particular database
def queryTraits(cursor, dbId, queryString):
    # we do this in two steps:
    # first we get the data id for the matching traits
    qry = '''
    SELECT
      ProbeSet.Id,
      ProbeSet.Name,
      ProbeSet.description,
      ProbeSetXRef.DataId
    FROM
      ProbeSet,
      ProbeSetXRef
    WHERE
      ProbeSetXRef.ProbeSetFreezeId = %s AND
      ProbeSet.Id = ProbeSetXRef.ProbeSetId AND
      ProbeSet.description LIKE "%%%s%%"
    ''' % (dbId, queryString)
    #    print qry
    cursor.execute(qry)

    if cursor.rowcount == 0:
        print "No traits found"
        return []
    else:
        print "%s traits found" % (cursor.rowcount)

    # maybe fetchall is bad; we will see
    traits = []
    for row in cursor.fetchall():
        myTrait = Trait(row[0], row[1], row[2])
        myTrait.dataId = row[3]
        traits.append(myTrait)

    # second we pull all of the strain data for each trait
    print "Retrieving individual trait data..."
    for trait in traits:
        trait.populateStrainData(cursor, trait.dataId)
        print "%s (%s) -- %s" % (trait.name, trait.id, trait.description)

    print "done"
    return traits

# queryProbeSetFreezes : Cursor -> arrayof String,String tuples
# to return the short and long name for each ProbeSetFreeze
# this function is designed specifically for building
# a database selector
def queryProbeSetFreezes(cursor):
    cursor.execute("""
    SELECT
      ProbeSetFreeze.Name,
      ProbeSetFreeze.FullName
    FROM
      ProbeSetFreeze
    ORDER BY
      ProbeSetFreeze.Name
    """)

    # for now, fetchall returns the data as a list of tuples
    # which is what we want
    return list(cursor.fetchall())

# queryProbeSetFreezeIdName: Cursor -> String -> String, String
# this function returns the
# id and the long name of a probesetfreeze given its name
# once again, it's designed specifically for building
# the database selector
def queryProbeSetFreezeIdName(cursor, name):
    qry = ('''
    SELECT
      ProbeSetFreeze.Id,
      ProbeSetFreeze.FullName
    FROM
      ProbeSetFreeze
    WHERE
      ProbeSetFreeze.Name = "%s" 
    ''' % name)
    #print qry
    cursor.execute(qry)

    row = cursor.fetchone()
    return row

# queryProbeSetFreezeName: Cursor -> String -> String
# to return the name of a probe set freeze given its id
def queryProbeSetFreezeName(cursor, id):
    cursor.execute('''
    SELECT
      ProbeSetFreeze.FullName
    FROM
      ProbeSetFreeze
    WHERE
      ProbeSetFreeze.Id = %s
    ''' % id)

    row = cursor.fetchone()
    return row[0]

# dbNameToTypeId : Cursor -> String -> (String, String)
# to convert a database name to a (type, id) pair
def dbNameToTypeId(cursor, name):
    types = ["ProbeSet", "Geno", "Publish"]
    for type_ in types:
        count = cursor.execute('''
        SELECT
          %sFreeze.Id
        FROM
          %sFreeze
        WHERE
          Name = "%s"
        ''' % (type_, type_,  name))

        if count != 0:
            id = cursor.fetchone()[0]
            return type_, id

    return None, None

# dbTypeIdToName : Cursor -> String -> String -> String
# to convert a database (type,id) pair into a name
def dbTypeIdToName(cursor, dbType, dbId):
    cursor.execute('''
    SELECT
      %sFreeze.Name
    FROM
      %sFreeze
    WHERE
    Id = %s
    ''' % (dbType, dbType, dbId))
    
    row = cursor.fetchone()
    return row[0]

#XZ, July 21, 2010: I add this function
def getSpeciesIdByDbTypeId (cursor, dbType, dbId):
    cursor.execute('''
    SELECT
      SpeciesId
    FROM
      InbredSet, %sFreeze
    WHERE
    %sFreeze.Id = %s
    and InbredSetId = InbredSet.Id
    ''' % (dbType, dbType, dbId))

    row = cursor.fetchone()
    return row[0]


# queryStrainCount : Cursor -> int
# return the number of strains in the database
def queryStrainCount(cursor):
    cursor.execute('''
    SELECT
      Max(Strain.Id)
    FROM
      Strain
    ''')
    return (cursor.fetchone())[0]