File: rdp_classifier.py

package info (click to toggle)
python-cogent 1.9-14
  • links: PTS, VCS
  • area: main
  • in suites: buster
  • size: 19,752 kB
  • sloc: python: 137,485; makefile: 149; sh: 64
file content (584 lines) | stat: -rw-r--r-- 22,891 bytes parent folder | download | duplicates (2)
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
#!/usr/bin/env python
"""Application controller for rdp_classifier-2.0
"""

__author__ = "Kyle Bittinger"
__copyright__ = "Copyright 2007-2016, The Cogent Project"
__credits__ = ["Kyle Bittinger","Greg Caporaso"]
__license__ = "GPL"
__version__ = "1.9"
__maintainer__ = "Kyle Bittinger"
__email__ = "kylebittinger@gmail.com"
__status__ = "Prototype"


import os.path
import re
from os import remove, environ, getenv, path
from optparse import OptionParser
from shutil import rmtree
import tempfile
import warnings
from cogent.app.parameters import Parameter, ValuedParameter, Parameters
from cogent.parse.fasta import MinimalFastaParser
from cogent.app.util import CommandLineApplication, CommandLineAppResult, \
    FilePath, ResultPath, guess_input_handler, system,\
    ApplicationNotFoundError, ApplicationError
from cogent.util.misc import app_path

class RdpClassifier(CommandLineApplication):
    """RDP Classifier application controller

    The RDP Classifier program is distributed as a java archive (.jar)
    file.  If the file 'rdp_classifier-2.2.jar' is not found in the
    current directory, the app controller uses the JAR file specified
    by the environment variable RDP_JAR_PATH.  If this variable is not
    set, and 'rdp_classifier-2.2.jar' is not found in the current
    directory, the application controller raises an
    ApplicationNotFoundError.

    The RDP Classifier often requires memory in excess of Java's
    default 64M. To correct this situation, the authors recommend
    increasing the maximum heap size for the java virtual machine.  An
    option '-Xmx' (default 1000M) is provided for this purpose.
    Details on this option may be found at
    http://java.sun.com/j2se/1.5.0/docs/tooldocs/solaris/java.html

    The classifier may optionally use a custom training set.  The full
    path to the training set may be provided in the option
    '-training-data'.
    """
    _input_handler = '_input_as_lines'
    _command = "/usr/share/java/rdp_classifier.jar"
    _options = {
        # output file name for classification assignment
        '-o': ValuedParameter('-', Name='o', Delimiter=' ', IsPath=True),
        # a property file contains the mapping of the training
        # files. Note: the training files and the property file should
        # be in the same directory. The default property file is set
        # to data/classifier/rRNAClassifier.properties.
        '-t': ValuedParameter('-', Name='t', Delimiter=' ', IsPath=True),
        # all tab delimited output format: [allrank|fixrank|db].
        # Default is allrank.
        #
        #   allrank: outputs the results for all ranks applied for
        #   each sequence: seqname, orientation, taxon name, rank,
        #   conf, ...
        #
        #   fixrank: only outputs the results for fixed ranks in
        #   order: no rank, domain, phylum, class, order, family,
        #   genus
        #
        #   db: outputs the seqname, trainset_no, tax_id, conf. This
        #   is good for storing in a database
        '-f': ValuedParameter('-', Name='f', Delimiter=' '),
        }

    # The following are available in the attributes JvmParameters,
    # JarParameters, and PositionalParameters

    _jvm_synonyms = {}
    _jvm_parameters = {
        # Maximum heap size for JVM.
        '-Xmx': ValuedParameter('-', Name='Xmx', Delimiter='', Value='1000m'),
        }

    _parameters = {}
    _parameters.update(_options)
    _parameters.update(_jvm_parameters)

    def getHelp(self):
        """Returns documentation string"""
        # Summary paragraph copied from rdp_classifier-2.0, which is
        # licensed under the GPL 2.0 and Copyright 2008 Michigan State
        # University Board of Trustees
        help_str = """\
        usage: ClassifierCmd [-f <arg>] [-o <arg>] [-q <arg>] [-t <arg>]

        -f,--format <arg> all tab delimited output format:
        [allrank|fixrank|db]. Default is allrank.

            allrank: outputs the results for all ranks applied for each
            sequence: seqname, orientation, taxon name, rank, conf, ...

            fixrank: only outputs the results for fixed ranks in order:
            no rank, domain, phylum, class, order, family, genus

            db: outputs the seqname, trainset_no, tax_id, conf. This is
            good for storing in a database

        -o,--outputFile <arg> output file name for classification
        assignment

        -q,--queryFile <arg> query file contains sequences in one of
        the following formats: Fasta, Genbank and EMBL

        -t,--train_propfile <arg> a property file contains the mapping
        of the training files.
        
        Note: the training files and the property file should be in
        the same directory. The default property file is set to
        data/classifier/rRNAClassifier.properties."""
        return help_str

    def _accept_exit_status(self, status):
        """Returns false if an error occurred in execution
        """
        return (status == 0)

    def _error_on_missing_application(self,params):
        """Raise an ApplicationNotFoundError if the app is not accessible

        In this case, checks for the java runtime and the RDP jar file.
        """
        if not (os.path.exists('java') or app_path('java')):
            raise ApplicationNotFoundError(
                "Cannot find java runtime. Is it installed? Is it in your "
                "path?")
        jar_fp = self._get_jar_fp()
        if jar_fp is None:
            raise ApplicationNotFoundError(
                "JAR file not found in current directory and the RDP_JAR_PATH "
                "environment variable is not set.  Please set RDP_JAR_PATH to "
                "the full pathname of the JAR file.")
        if not os.path.exists(jar_fp):
            raise ApplicationNotFoundError(
                "JAR file %s does not exist." % jar_fp)

    def _get_jar_fp(self):
        """Returns the full path to the JAR file.

        If the JAR file cannot be found in the current directory and
        the environment variable RDP_JAR_PATH is not set, returns
        None.
        """
        # handles case where the jar file is in the current working directory
        if os.path.exists(self._command):
            return self._command
        # handles the case where the user has specified the location via
        # an environment variable
        elif 'RDP_JAR_PATH' in environ:
            return getenv('RDP_JAR_PATH')
        else:
            return None

    # Overridden to pull out JVM-specific command-line arguments.
    def _get_base_command(self):
        """Returns the base command plus command-line options.

        Does not include input file, output file, and training set.
        """
        cd_command = ''.join(['cd ', str(self.WorkingDir), ';'])
        jvm_command = "java"
        jvm_arguments = self._commandline_join(
            [self.Parameters[k] for k in self._jvm_parameters])
        jar_arguments = '-jar "%s"' % self._get_jar_fp()
        rdp_arguments = self._commandline_join(
            [self.Parameters[k] for k in self._options])

        command_parts = [
            cd_command, jvm_command, jvm_arguments, jar_arguments,
            rdp_arguments, '-q']
        return self._commandline_join(command_parts).strip()
    
    BaseCommand = property(_get_base_command)

    def _commandline_join(self, tokens):
        """Formats a list of tokens as a shell command
 
        This seems to be a repeated pattern; may be useful in
        superclass.
        """
        commands = filter(None, map(str, tokens))
        return self._command_delimiter.join(commands).strip()

    def _get_result_paths(self,data):
        """ Return a dict of ResultPath objects representing all possible output
        """
        assignment_fp = str(self.Parameters['-o'].Value).strip('"')
        if not os.path.isabs(assignment_fp):
            assignment_fp = os.path.relpath(assignment_fp, self.WorkingDir)
        return {'Assignments': ResultPath(assignment_fp, IsWritten=True)}


class RdpTrainer(RdpClassifier):
    _input_handler = '_input_as_lines'
    TrainingClass = 'edu.msu.cme.rdp.classifier.train.ClassifierTraineeMaker'
    PropertiesFile = 'RdpClassifier.properties'

    _parameters = {
        'taxonomy_file': ValuedParameter(None, None, IsPath=True),
        'model_output_dir': ValuedParameter(None, None, IsPath=True),
        'training_set_id': ValuedParameter(None, None, Value='1'),
        'taxonomy_version': ValuedParameter(None, None, Value='version1'),
        'modification_info': ValuedParameter(None, None, Value='cogent'),
        }
    _jvm_parameters = {
        # Maximum heap size for JVM.
        '-Xmx': ValuedParameter('-', Name='Xmx', Delimiter='', Value='1000m'),
        }
    _parameters.update(_jvm_parameters)

    def _get_base_command(self):
        """Returns the base command plus command-line options.

        Handles everything up to and including the classpath.  The
        positional training parameters are added by the
        _input_handler_decorator method.
        """
        cd_command = ''.join(['cd ', str(self.WorkingDir), ';'])
        jvm_command = "java"
        jvm_args = self._commandline_join(
            [self.Parameters[k] for k in self._jvm_parameters])
        cp_args = '-cp "%s" %s' % (self._get_jar_fp(), self.TrainingClass)

        command_parts = [cd_command, jvm_command, jvm_args, cp_args]
        return self._commandline_join(command_parts).strip()
    
    BaseCommand = property(_get_base_command)

    def _set_input_handler(self, method_name):
        """Stores the selected input handler in a private attribute.
        """
        self.__InputHandler = method_name

    def _get_input_handler(self):
        """Returns decorator that wraps the requested input handler.
        """
        return '_input_handler_decorator'

    InputHandler = property(_get_input_handler, _set_input_handler)

    @property
    def ModelDir(self):
        """Absolute FilePath to the training output directory.
        """
        model_dir = self.Parameters['model_output_dir'].Value
        absolute_model_dir = os.path.abspath(model_dir)
        return FilePath(absolute_model_dir)

    def _input_handler_decorator(self, data):
        """Adds positional parameters to selected input_handler's results.
        """
        input_handler = getattr(self, self.__InputHandler)
        input_parts = [
            self.Parameters['taxonomy_file'],
            input_handler(data),
            self.Parameters['training_set_id'],
            self.Parameters['taxonomy_version'],
            self.Parameters['modification_info'],
            self.ModelDir,
            ]
        return self._commandline_join(input_parts)

    def _get_result_paths(self, output_dir):
        """Return a dict of output files.
        """
        # Only include the properties file here. Add the other result
        # paths in the __call__ method, so we can catch errors if an
        # output file is not written.
        self._write_properties_file()
        properties_fp = os.path.join(self.ModelDir, self.PropertiesFile)
        result_paths = {
            'properties': ResultPath(properties_fp, IsWritten=True,)
            }
        return result_paths
    
    def _write_properties_file(self):
        """Write an RDP training properties file manually.
        """
        # The properties file specifies the names of the files in the
        # training directory.  We use the example properties file
        # directly from the rdp_classifier distribution, which lists
        # the default set of files created by the application.  We
        # must write this file manually after generating the
        # training data.
        properties_fp = os.path.join(self.ModelDir, self.PropertiesFile)
        properties_file = open(properties_fp, 'w')
        properties_file.write(
            "# Sample ResourceBundle properties file\n"
            "bergeyTree=bergeyTrainingTree.xml\n"
            "probabilityList=genus_wordConditionalProbList.txt\n"
            "probabilityIndex=wordConditionalProbIndexArr.txt\n"
            "wordPrior=logWordPrior.txt\n"
            "classifierVersion=Naive Bayesian rRNA Classifier Version 1.0, "
            "November 2003\n"
            )
        properties_file.close()
      
    def __call__(self, data=None, remove_tmp=True):
        """Run the application with the specified kwargs on data

        data: anything that can be cast into a string or written out
          to a file. Usually either a list of things or a single
          string or number. input_handler will be called on this data
          before it is passed as part of the command-line argument, so
          by creating your own input handlers you can customize what
          kind of data you want your application to accept

        remove_tmp: if True, removes tmp files
        """
        result = super(RdpClassifier, self).__call__(data=data, remove_tmp=remove_tmp)
        training_files = {
            'bergeyTree': 'bergeyTrainingTree.xml',
            'probabilityList': 'genus_wordConditionalProbList.txt',
            'probabilityIndex': 'wordConditionalProbIndexArr.txt',
            'wordPrior': 'logWordPrior.txt',
        }
        for key, training_fn in sorted(training_files.items()):
            training_fp = os.path.join(self.ModelDir, training_fn)
            if not os.path.exists(training_fp):
                exception_msg = (
                    "Training output file %s not found.  This may "
                    "happen if an error occurred during the RDP training "
                    "process.  More details may be available in the "
                    "standard error, printed below.\n\n" % training_fp
                    )
                stderr_msg = result["StdErr"].read()
                result["StdErr"].seek(0)
                raise ApplicationError(exception_msg + stderr_msg)
            # Not in try/except clause because we already know the
            # file exists. Failure would be truly exceptional, and we
            # want to maintain the original exception in that case.
            result[key] = open(training_fp)
        return result
        

def parse_command_line_parameters(argv=None):
    """ Parses command line arguments """
    usage =\
     'usage: %prog [options] input_sequences_filepath'
    version = 'Version: %prog ' +  __version__
    parser = OptionParser(usage=usage, version=version)
          
    parser.add_option('-o','--output_fp',action='store',\
          type='string',dest='output_fp',help='Path to store '+\
          'output file [default: generated from input_sequences_filepath]')
          
    parser.add_option('-c','--min_confidence',action='store',\
          type='float',dest='min_confidence',help='minimum confidence '+\
          'level to return a classification [default: %default]')

    parser.set_defaults(verbose=False, min_confidence=0.80)

    opts, args = parser.parse_args(argv)
    if len(args) != 1:
       parser.error('Exactly one argument is required.')

    return opts, args


def assign_taxonomy(
    data, min_confidence=0.80, output_fp=None, training_data_fp=None,
    fixrank=True, max_memory=None, tmp_dir=None):
    """Assign taxonomy to each sequence in data with the RDP classifier
    
        data: open fasta file object or list of fasta lines
        confidence: minimum support threshold to assign taxonomy to a sequence
        output_fp: path to write output; if not provided, result will be 
         returned in a dict of {seq_id:(taxonomy_assignment,confidence)}
    """
    # Going to iterate through this twice in succession, best to force
    # evaluation now
    data = list(data)

    # RDP classifier doesn't preserve identifiers with spaces
    # Use lookup table
    seq_id_lookup = {}
    for seq_id, seq in MinimalFastaParser(data):
        seq_id_lookup[seq_id.split()[0]] = seq_id

    app_kwargs = {}
    if tmp_dir is not None:
        app_kwargs['TmpDir'] = tmp_dir
    app = RdpClassifier(**app_kwargs)

    if max_memory is not None:
        app.Parameters['-Xmx'].on(max_memory)
    
    temp_output_file = tempfile.NamedTemporaryFile(
        prefix='RdpAssignments_', suffix='.txt', dir=tmp_dir)
    app.Parameters['-o'].on(temp_output_file.name)
    if training_data_fp is not None:
        app.Parameters['-t'].on(training_data_fp)

    if fixrank:
        app.Parameters['-f'].on('fixrank')
    else:
        app.Parameters['-f'].on('allrank')

    app_result = app(data)

    assignments = {}

    # ShortSequenceException messages are written to stdout
    # Tag these ID's as unassignable
    for line in app_result['StdOut']:
        excep = parse_rdp_exception(line)
        if excep is not None:
            _, rdp_id = excep
            orig_id = seq_id_lookup[rdp_id]
            assignments[orig_id] = ('Unassignable', 1.0)
    
    for line in app_result['Assignments']:
        rdp_id, direction, taxa = parse_rdp_assignment(line)
        if taxa[0][0] == "Root":
            taxa = taxa[1:]
        orig_id = seq_id_lookup[rdp_id]
        lineage, confidence = get_rdp_lineage(taxa, min_confidence)
        if lineage:
            assignments[orig_id] = (';'.join(lineage), confidence)
        else:
            assignments[orig_id] = ('Unclassified', 1.0)

    if output_fp:
        try:
            output_file = open(output_fp, 'w')
        except OSError:
            raise OSError("Can't open output file for writing: %s" % output_fp)
        for seq_id, assignment in assignments.items():
            lineage, confidence = assignment
            output_file.write(
                '%s\t%s\t%1.3f\n' % (seq_id, lineage, confidence))
        output_file.close()
        return None
    else:
        return assignments


def train_rdp_classifier(
    training_seqs_file, taxonomy_file, model_output_dir, max_memory=None,
    tmp_dir=None):
    """ Train RDP Classifier, saving to model_output_dir

        training_seqs_file, taxonomy_file: file-like objects used to
            train the RDP Classifier (see RdpTrainer documentation for
            format of training data)

        model_output_dir: directory in which to save the files
            necessary to classify sequences according to the training
            data

    Once the model data has been generated, the RDP Classifier may 
    """
    app_kwargs = {}
    if tmp_dir is not None:
        app_kwargs['TmpDir'] = tmp_dir
    app = RdpTrainer(**app_kwargs)

    if max_memory is not None:
        app.Parameters['-Xmx'].on(max_memory)
    
    temp_taxonomy_file = tempfile.NamedTemporaryFile(
        prefix='RdpTaxonomy_', suffix='.txt', dir=tmp_dir)
    temp_taxonomy_file.write(taxonomy_file.read())
    temp_taxonomy_file.seek(0)

    app.Parameters['taxonomy_file'].on(temp_taxonomy_file.name)
    app.Parameters['model_output_dir'].on(model_output_dir)
    return app(training_seqs_file)
    

def train_rdp_classifier_and_assign_taxonomy(
    training_seqs_file, taxonomy_file, seqs_to_classify, min_confidence=0.80, 
    model_output_dir=None, classification_output_fp=None, max_memory=None,
    tmp_dir=None):
    """ Train RDP Classifier and assign taxonomy in one fell swoop

    The file objects training_seqs_file and taxonomy_file are used to
    train the RDP Classifier (see RdpTrainer documentation for
    details).  Model data is stored in model_output_dir.  If
    model_output_dir is not provided, a temporary directory is created
    and removed after classification.

    The sequences in seqs_to_classify are classified according to the
    model and filtered at the desired confidence level (default:
    0.80).

    The results are saved to classification_output_fp if provided,
    otherwise a dict of {seq_id:(taxonomy_assignment,confidence)} is
    returned.
    """
    if model_output_dir is None:
        training_dir = tempfile.mkdtemp(prefix='RdpTrainer_', dir=tmp_dir)
    else:
        training_dir = model_output_dir

    training_results = train_rdp_classifier(
        training_seqs_file, taxonomy_file, training_dir, max_memory=max_memory,
        tmp_dir=tmp_dir)
    training_data_fp = training_results['properties'].name

    assignment_results = assign_taxonomy(
        seqs_to_classify, min_confidence=min_confidence, 
        output_fp=classification_output_fp, training_data_fp=training_data_fp,
        max_memory=max_memory, fixrank=False, tmp_dir=tmp_dir)

    if model_output_dir is None:
        # Forum user reported an error on the call to os.rmtree:
        # https://groups.google.com/d/topic/qiime-forum/MkNe7-JtSBw/discussion
        # We were not able to replicate the problem and fix it
        # properly.  However, even if an error occurs, we would like
        # to return results, along with a warning.
        try:
            rmtree(training_dir)
        except OSError:
            msg = (
                "Temporary training directory %s not removed" % training_dir)
            if os.path.isdir(training_dir):
                training_dir_files = os.listdir(training_dir)
                msg += "\nDetected files %s" % training_dir_files
            warnings.warn(msg, RuntimeWarning)

    return assignment_results


def get_rdp_lineage(rdp_taxa, min_confidence):
    lineage = []
    obs_confidence = 1.0
    for taxon, rank, confidence in rdp_taxa:
        if confidence >= min_confidence:
            obs_confidence = confidence
            lineage.append(taxon)
        else:
            break
    return lineage, obs_confidence


def parse_rdp_exception(line):
    if line.startswith('ShortSequenceException'):
        matchobj = re.search('recordID=(\S+)', line)
        if matchobj:
            rdp_id = matchobj.group(1)
            return ('ShortSequenceException', rdp_id)
    return None


def parse_rdp_assignment(line):
    """Returns a list of assigned taxa from an RDP classification line
    """
    toks = line.strip().split('\t')
    seq_id = toks.pop(0)
    direction = toks.pop(0)
    if ((len(toks) % 3) != 0):
        raise ValueError(
            "Expected assignments in a repeating series of (rank, name, "
            "confidence), received %s" % toks)
    assignments = []
    # Fancy way to create list of triples using consecutive items from
    # input.  See grouper function in documentation for itertools for
    # more general example.
    itoks = iter(toks)
    for taxon, rank, confidence_str in zip(itoks, itoks, itoks):
        if not taxon:
            continue
        assignments.append((taxon.strip('"'), rank, float(confidence_str)))
    return seq_id, direction, assignments


if __name__ == "__main__":    
    opts, args = parse_command_line_parameters()
    assign_taxonomy(
        open(args[0]), min_confidence=opts.min_confidence,
        output_fp=opts.output_fp)