BioDSL 1.0.0
Sign up to get free protection for your applications and to get access to all the features.
- checksums.yaml +7 -0
- data/.gitignore +10 -0
- data/BioDSL.gemspec +64 -0
- data/LICENSE +339 -0
- data/README.md +205 -0
- data/Rakefile +94 -0
- data/examples/fastq_to_fasta.rb +8 -0
- data/lib/BioDSL/cary.rb +242 -0
- data/lib/BioDSL/command.rb +133 -0
- data/lib/BioDSL/commands/add_key.rb +110 -0
- data/lib/BioDSL/commands/align_seq_mothur.rb +194 -0
- data/lib/BioDSL/commands/analyze_residue_distribution.rb +222 -0
- data/lib/BioDSL/commands/assemble_pairs.rb +336 -0
- data/lib/BioDSL/commands/assemble_seq_idba.rb +230 -0
- data/lib/BioDSL/commands/assemble_seq_ray.rb +345 -0
- data/lib/BioDSL/commands/assemble_seq_spades.rb +252 -0
- data/lib/BioDSL/commands/classify_seq.rb +217 -0
- data/lib/BioDSL/commands/classify_seq_mothur.rb +226 -0
- data/lib/BioDSL/commands/clip_primer.rb +318 -0
- data/lib/BioDSL/commands/cluster_otus.rb +181 -0
- data/lib/BioDSL/commands/collapse_otus.rb +170 -0
- data/lib/BioDSL/commands/collect_otus.rb +150 -0
- data/lib/BioDSL/commands/complement_seq.rb +117 -0
- data/lib/BioDSL/commands/count.rb +135 -0
- data/lib/BioDSL/commands/count_values.rb +149 -0
- data/lib/BioDSL/commands/degap_seq.rb +253 -0
- data/lib/BioDSL/commands/dereplicate_seq.rb +168 -0
- data/lib/BioDSL/commands/dump.rb +157 -0
- data/lib/BioDSL/commands/filter_rrna.rb +239 -0
- data/lib/BioDSL/commands/genecall.rb +237 -0
- data/lib/BioDSL/commands/grab.rb +535 -0
- data/lib/BioDSL/commands/index_taxonomy.rb +226 -0
- data/lib/BioDSL/commands/mask_seq.rb +175 -0
- data/lib/BioDSL/commands/mean_scores.rb +168 -0
- data/lib/BioDSL/commands/merge_pair_seq.rb +175 -0
- data/lib/BioDSL/commands/merge_table.rb +225 -0
- data/lib/BioDSL/commands/merge_values.rb +113 -0
- data/lib/BioDSL/commands/plot_heatmap.rb +233 -0
- data/lib/BioDSL/commands/plot_histogram.rb +306 -0
- data/lib/BioDSL/commands/plot_matches.rb +282 -0
- data/lib/BioDSL/commands/plot_residue_distribution.rb +278 -0
- data/lib/BioDSL/commands/plot_scores.rb +285 -0
- data/lib/BioDSL/commands/random.rb +153 -0
- data/lib/BioDSL/commands/read_fasta.rb +222 -0
- data/lib/BioDSL/commands/read_fastq.rb +414 -0
- data/lib/BioDSL/commands/read_table.rb +329 -0
- data/lib/BioDSL/commands/reverse_seq.rb +113 -0
- data/lib/BioDSL/commands/slice_align.rb +400 -0
- data/lib/BioDSL/commands/slice_seq.rb +151 -0
- data/lib/BioDSL/commands/sort.rb +223 -0
- data/lib/BioDSL/commands/split_pair_seq.rb +220 -0
- data/lib/BioDSL/commands/split_values.rb +165 -0
- data/lib/BioDSL/commands/trim_primer.rb +314 -0
- data/lib/BioDSL/commands/trim_seq.rb +192 -0
- data/lib/BioDSL/commands/uchime_ref.rb +170 -0
- data/lib/BioDSL/commands/uclust.rb +286 -0
- data/lib/BioDSL/commands/unique_values.rb +145 -0
- data/lib/BioDSL/commands/usearch_global.rb +171 -0
- data/lib/BioDSL/commands/usearch_local.rb +171 -0
- data/lib/BioDSL/commands/write_fasta.rb +207 -0
- data/lib/BioDSL/commands/write_fastq.rb +191 -0
- data/lib/BioDSL/commands/write_table.rb +419 -0
- data/lib/BioDSL/commands/write_tree.rb +167 -0
- data/lib/BioDSL/commands.rb +31 -0
- data/lib/BioDSL/config.rb +55 -0
- data/lib/BioDSL/csv.rb +307 -0
- data/lib/BioDSL/debug.rb +42 -0
- data/lib/BioDSL/fasta.rb +133 -0
- data/lib/BioDSL/fastq.rb +77 -0
- data/lib/BioDSL/filesys.rb +137 -0
- data/lib/BioDSL/fork.rb +145 -0
- data/lib/BioDSL/hamming.rb +128 -0
- data/lib/BioDSL/helpers/aux_helper.rb +44 -0
- data/lib/BioDSL/helpers/email_helper.rb +66 -0
- data/lib/BioDSL/helpers/history_helper.rb +40 -0
- data/lib/BioDSL/helpers/log_helper.rb +55 -0
- data/lib/BioDSL/helpers/options_helper.rb +405 -0
- data/lib/BioDSL/helpers/status_helper.rb +132 -0
- data/lib/BioDSL/helpers.rb +35 -0
- data/lib/BioDSL/html_report.rb +200 -0
- data/lib/BioDSL/math.rb +55 -0
- data/lib/BioDSL/mummer.rb +216 -0
- data/lib/BioDSL/pipeline.rb +354 -0
- data/lib/BioDSL/seq/ambiguity.rb +66 -0
- data/lib/BioDSL/seq/assemble.rb +240 -0
- data/lib/BioDSL/seq/backtrack.rb +252 -0
- data/lib/BioDSL/seq/digest.rb +99 -0
- data/lib/BioDSL/seq/dynamic.rb +263 -0
- data/lib/BioDSL/seq/homopolymer.rb +59 -0
- data/lib/BioDSL/seq/kmer.rb +293 -0
- data/lib/BioDSL/seq/levenshtein.rb +113 -0
- data/lib/BioDSL/seq/translate.rb +109 -0
- data/lib/BioDSL/seq/trim.rb +188 -0
- data/lib/BioDSL/seq.rb +742 -0
- data/lib/BioDSL/serializer.rb +98 -0
- data/lib/BioDSL/stream.rb +113 -0
- data/lib/BioDSL/taxonomy.rb +691 -0
- data/lib/BioDSL/test.rb +42 -0
- data/lib/BioDSL/tmp_dir.rb +68 -0
- data/lib/BioDSL/usearch.rb +301 -0
- data/lib/BioDSL/verbose.rb +42 -0
- data/lib/BioDSL/version.rb +31 -0
- data/lib/BioDSL.rb +81 -0
- data/test/BioDSL/commands/test_add_key.rb +105 -0
- data/test/BioDSL/commands/test_align_seq_mothur.rb +99 -0
- data/test/BioDSL/commands/test_analyze_residue_distribution.rb +134 -0
- data/test/BioDSL/commands/test_assemble_pairs.rb +459 -0
- data/test/BioDSL/commands/test_assemble_seq_idba.rb +50 -0
- data/test/BioDSL/commands/test_assemble_seq_ray.rb +51 -0
- data/test/BioDSL/commands/test_assemble_seq_spades.rb +50 -0
- data/test/BioDSL/commands/test_classify_seq.rb +50 -0
- data/test/BioDSL/commands/test_classify_seq_mothur.rb +59 -0
- data/test/BioDSL/commands/test_clip_primer.rb +377 -0
- data/test/BioDSL/commands/test_cluster_otus.rb +128 -0
- data/test/BioDSL/commands/test_collapse_otus.rb +81 -0
- data/test/BioDSL/commands/test_collect_otus.rb +82 -0
- data/test/BioDSL/commands/test_complement_seq.rb +78 -0
- data/test/BioDSL/commands/test_count.rb +103 -0
- data/test/BioDSL/commands/test_count_values.rb +85 -0
- data/test/BioDSL/commands/test_degap_seq.rb +96 -0
- data/test/BioDSL/commands/test_dereplicate_seq.rb +92 -0
- data/test/BioDSL/commands/test_dump.rb +109 -0
- data/test/BioDSL/commands/test_filter_rrna.rb +128 -0
- data/test/BioDSL/commands/test_genecall.rb +50 -0
- data/test/BioDSL/commands/test_grab.rb +398 -0
- data/test/BioDSL/commands/test_index_taxonomy.rb +62 -0
- data/test/BioDSL/commands/test_mask_seq.rb +98 -0
- data/test/BioDSL/commands/test_mean_scores.rb +111 -0
- data/test/BioDSL/commands/test_merge_pair_seq.rb +115 -0
- data/test/BioDSL/commands/test_merge_table.rb +131 -0
- data/test/BioDSL/commands/test_merge_values.rb +83 -0
- data/test/BioDSL/commands/test_plot_heatmap.rb +185 -0
- data/test/BioDSL/commands/test_plot_histogram.rb +194 -0
- data/test/BioDSL/commands/test_plot_matches.rb +157 -0
- data/test/BioDSL/commands/test_plot_residue_distribution.rb +309 -0
- data/test/BioDSL/commands/test_plot_scores.rb +308 -0
- data/test/BioDSL/commands/test_random.rb +88 -0
- data/test/BioDSL/commands/test_read_fasta.rb +229 -0
- data/test/BioDSL/commands/test_read_fastq.rb +552 -0
- data/test/BioDSL/commands/test_read_table.rb +327 -0
- data/test/BioDSL/commands/test_reverse_seq.rb +79 -0
- data/test/BioDSL/commands/test_slice_align.rb +218 -0
- data/test/BioDSL/commands/test_slice_seq.rb +131 -0
- data/test/BioDSL/commands/test_sort.rb +128 -0
- data/test/BioDSL/commands/test_split_pair_seq.rb +164 -0
- data/test/BioDSL/commands/test_split_values.rb +95 -0
- data/test/BioDSL/commands/test_trim_primer.rb +329 -0
- data/test/BioDSL/commands/test_trim_seq.rb +150 -0
- data/test/BioDSL/commands/test_uchime_ref.rb +113 -0
- data/test/BioDSL/commands/test_uclust.rb +139 -0
- data/test/BioDSL/commands/test_unique_values.rb +98 -0
- data/test/BioDSL/commands/test_usearch_global.rb +123 -0
- data/test/BioDSL/commands/test_usearch_local.rb +125 -0
- data/test/BioDSL/commands/test_write_fasta.rb +159 -0
- data/test/BioDSL/commands/test_write_fastq.rb +166 -0
- data/test/BioDSL/commands/test_write_table.rb +411 -0
- data/test/BioDSL/commands/test_write_tree.rb +122 -0
- data/test/BioDSL/helpers/test_options_helper.rb +272 -0
- data/test/BioDSL/seq/test_assemble.rb +98 -0
- data/test/BioDSL/seq/test_backtrack.rb +176 -0
- data/test/BioDSL/seq/test_digest.rb +71 -0
- data/test/BioDSL/seq/test_dynamic.rb +133 -0
- data/test/BioDSL/seq/test_homopolymer.rb +58 -0
- data/test/BioDSL/seq/test_kmer.rb +134 -0
- data/test/BioDSL/seq/test_translate.rb +75 -0
- data/test/BioDSL/seq/test_trim.rb +101 -0
- data/test/BioDSL/test_cary.rb +176 -0
- data/test/BioDSL/test_command.rb +45 -0
- data/test/BioDSL/test_csv.rb +514 -0
- data/test/BioDSL/test_debug.rb +42 -0
- data/test/BioDSL/test_fasta.rb +154 -0
- data/test/BioDSL/test_fastq.rb +46 -0
- data/test/BioDSL/test_filesys.rb +145 -0
- data/test/BioDSL/test_fork.rb +85 -0
- data/test/BioDSL/test_math.rb +41 -0
- data/test/BioDSL/test_mummer.rb +79 -0
- data/test/BioDSL/test_pipeline.rb +187 -0
- data/test/BioDSL/test_seq.rb +790 -0
- data/test/BioDSL/test_serializer.rb +72 -0
- data/test/BioDSL/test_stream.rb +55 -0
- data/test/BioDSL/test_taxonomy.rb +336 -0
- data/test/BioDSL/test_test.rb +42 -0
- data/test/BioDSL/test_tmp_dir.rb +58 -0
- data/test/BioDSL/test_usearch.rb +33 -0
- data/test/BioDSL/test_verbose.rb +42 -0
- data/test/helper.rb +82 -0
- data/www/command.html.haml +14 -0
- data/www/css.html.haml +55 -0
- data/www/input_files.html.haml +3 -0
- data/www/layout.html.haml +12 -0
- data/www/output_files.html.haml +3 -0
- data/www/overview.html.haml +15 -0
- data/www/pipeline.html.haml +4 -0
- data/www/png.html.haml +2 -0
- data/www/status.html.haml +9 -0
- data/www/time.html.haml +11 -0
- metadata +503 -0
@@ -0,0 +1,354 @@
|
|
1
|
+
# >>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>><<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<< #
|
2
|
+
# Copyright (C) 2007-2015 Martin Asser Hansen (mail@maasha.dk). #
|
3
|
+
# #
|
4
|
+
# This program is free software; you can redistribute it and/or #
|
5
|
+
# modify it under the terms of the GNU General Public License #
|
6
|
+
# as published by the Free Software Foundation; either version 2 #
|
7
|
+
# of the License, or (at your option) any later version. #
|
8
|
+
# #
|
9
|
+
# This program is distributed in the hope that it will be useful, #
|
10
|
+
# but WITHOUT ANY WARRANTY; without even the implied warranty of #
|
11
|
+
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the #
|
12
|
+
# GNU General Public License for more details. #
|
13
|
+
# #
|
14
|
+
# You should have received a copy of the GNU General Public License #
|
15
|
+
# along with this program; if not, write to the Free Software #
|
16
|
+
# Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, #
|
17
|
+
# USA. #
|
18
|
+
# #
|
19
|
+
# http://www.gnu.org/copyleft/gpl.html #
|
20
|
+
# #
|
21
|
+
# >>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>><<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<< #
|
22
|
+
# #
|
23
|
+
# This software is part of BioDSL (www.github.com/maasha/BioDSL). #
|
24
|
+
# #
|
25
|
+
# >>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>><<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<< #
|
26
|
+
module BioDSL
|
27
|
+
trap('INT') { fail 'Interrupted: ctrl-c pressed' }
|
28
|
+
|
29
|
+
# Error class for Pipeline errors.
|
30
|
+
PipelineError = Class.new(StandardError)
|
31
|
+
|
32
|
+
# rubocop: disable ClassLength
|
33
|
+
|
34
|
+
# Pipeline class
|
35
|
+
class Pipeline
|
36
|
+
require 'BioDSL/command'
|
37
|
+
require 'BioDSL/helpers/email_helper'
|
38
|
+
require 'BioDSL/helpers/history_helper'
|
39
|
+
require 'BioDSL/helpers/log_helper'
|
40
|
+
require 'BioDSL/helpers/options_helper'
|
41
|
+
require 'BioDSL/helpers/status_helper'
|
42
|
+
require 'mail'
|
43
|
+
require 'yaml'
|
44
|
+
|
45
|
+
include EmailHelper
|
46
|
+
include LogHelper
|
47
|
+
include HistoryHelper
|
48
|
+
include OptionsHelper
|
49
|
+
include StatusHelper
|
50
|
+
|
51
|
+
attr_accessor :commands, :complete
|
52
|
+
|
53
|
+
# Pipeline class constructor.
|
54
|
+
def initialize
|
55
|
+
@commands = [] # Array of Commands in the Pipeline.
|
56
|
+
@options = {} # Options hash.
|
57
|
+
@enums = [[]] # Array of Enumerators.
|
58
|
+
@complete = false # Flag denoting if run was completed.
|
59
|
+
end
|
60
|
+
|
61
|
+
# @return [Integer] The size or number of commands in a pipeline.
|
62
|
+
def size
|
63
|
+
@commands.size
|
64
|
+
end
|
65
|
+
|
66
|
+
# Method for merging one pipeline onto another.
|
67
|
+
#
|
68
|
+
# @param other [Pipeline] Pipeline to merge.
|
69
|
+
#
|
70
|
+
# @return [self].
|
71
|
+
def <<(other)
|
72
|
+
other.commands.map { |command| commands << command }
|
73
|
+
other.status.map { |status| self.status << status }
|
74
|
+
|
75
|
+
self
|
76
|
+
end
|
77
|
+
|
78
|
+
# Method that adds two Pipelines and return a new Pipeline.
|
79
|
+
def +(other)
|
80
|
+
unless other.is_a?(BioDSL::Pipeline)
|
81
|
+
fail PipelineError, "Not a pipeline: #{other.inspect}"
|
82
|
+
end
|
83
|
+
|
84
|
+
p = self.class.new
|
85
|
+
p << self
|
86
|
+
p << other
|
87
|
+
end
|
88
|
+
|
89
|
+
# Removes last command from a Pipeline and returns a new Pipeline with this
|
90
|
+
# command.
|
91
|
+
def pop
|
92
|
+
p = BioDSL::Pipeline.new
|
93
|
+
p.commands = [@commands.pop]
|
94
|
+
p
|
95
|
+
end
|
96
|
+
|
97
|
+
# Run all the commands in the Pipeline.
|
98
|
+
#
|
99
|
+
# @param options [Hash]
|
100
|
+
# @option options [Boolean] :verbose (false) Enable verbose output.
|
101
|
+
#
|
102
|
+
# @raise [PipelineError] If no commands are added to the pipeline.
|
103
|
+
#
|
104
|
+
# @return [self]
|
105
|
+
def run(options = {})
|
106
|
+
prime_variables(options)
|
107
|
+
|
108
|
+
fail BioDSL::PipelineError, 'Empty pipeline' if @commands.empty?
|
109
|
+
|
110
|
+
@options = options
|
111
|
+
|
112
|
+
check_options
|
113
|
+
command_runner
|
114
|
+
print_status
|
115
|
+
send_email(self)
|
116
|
+
save_report
|
117
|
+
log_ok
|
118
|
+
|
119
|
+
self
|
120
|
+
rescue => exception
|
121
|
+
exit_gracefully(exception)
|
122
|
+
ensure
|
123
|
+
save_history
|
124
|
+
end
|
125
|
+
|
126
|
+
# Return a list of all status hashes from the commands.
|
127
|
+
#
|
128
|
+
# @return [Array] List of status hashes.
|
129
|
+
def status
|
130
|
+
@commands.each_with_object([]) do |e, a|
|
131
|
+
if @complete
|
132
|
+
e.calc_time_elapsed
|
133
|
+
e.calc_delta
|
134
|
+
end
|
135
|
+
|
136
|
+
a << e.status
|
137
|
+
end
|
138
|
+
end
|
139
|
+
|
140
|
+
# Format a Pipeline to a pretty string which is returned.
|
141
|
+
def to_s
|
142
|
+
command_strings = %w(BP new)
|
143
|
+
|
144
|
+
@commands.each { |command| command_strings << command.to_s }
|
145
|
+
|
146
|
+
if @complete
|
147
|
+
if @options.empty?
|
148
|
+
command_strings << 'run'
|
149
|
+
else
|
150
|
+
command_strings << "run(#{options_string})"
|
151
|
+
end
|
152
|
+
end
|
153
|
+
|
154
|
+
command_strings.join('.')
|
155
|
+
end
|
156
|
+
|
157
|
+
private
|
158
|
+
|
159
|
+
# Add a command to the pipeline. This is done by first requiring the
|
160
|
+
# relevant Class/Module and then calling the relevant command.
|
161
|
+
#
|
162
|
+
# @param method [Symbol] Method name.
|
163
|
+
# @param args [Array] Method arguments.
|
164
|
+
# @param block [Proc] Method block.
|
165
|
+
#
|
166
|
+
# @example Here we add the command `dump` to the pipeline.
|
167
|
+
# Pipeline.new.dump
|
168
|
+
# # => self
|
169
|
+
#
|
170
|
+
# @return [self]
|
171
|
+
def method_missing(method, *args, &block)
|
172
|
+
require_file(method)
|
173
|
+
|
174
|
+
const = method.to_s.split('_').map(&:capitalize).join('')
|
175
|
+
|
176
|
+
if BioDSL.const_defined? const
|
177
|
+
options = args.first || {}
|
178
|
+
options_load_rc(options, method)
|
179
|
+
|
180
|
+
klass = BioDSL.const_get(const)
|
181
|
+
klass.send(:include, OptionsHelper)
|
182
|
+
klass.send(:include, StatusHelper)
|
183
|
+
lmb = klass.send(:new, options).lmb
|
184
|
+
|
185
|
+
@commands << Command.new(method, lmb, options)
|
186
|
+
else
|
187
|
+
super
|
188
|
+
end
|
189
|
+
|
190
|
+
self
|
191
|
+
end
|
192
|
+
|
193
|
+
# Require a file form the lib/commands directory given a method name that
|
194
|
+
# must match the file name. E.g. `require_file(:dump)` requires the file
|
195
|
+
# `lib/commands/dump.rb`.
|
196
|
+
#
|
197
|
+
# @param method [Symbol]
|
198
|
+
# The name of the method.
|
199
|
+
#
|
200
|
+
# @raise [Errno::ENOENT] If no such file was found.
|
201
|
+
def require_file(method)
|
202
|
+
return if BioDSL.const_defined? method.to_s.capitalize
|
203
|
+
|
204
|
+
# FIXME
|
205
|
+
# file = File.join('lib', 'BioDSL', 'commands', "#{method}.rb")
|
206
|
+
# fail Errno::ENOENT, "No such file: #{file}" unless File.exist? file
|
207
|
+
|
208
|
+
require File.join('BioDSL', 'commands', method.to_s)
|
209
|
+
end
|
210
|
+
|
211
|
+
# Print status.
|
212
|
+
def print_status
|
213
|
+
return unless @options[:verbose]
|
214
|
+
|
215
|
+
@commands.each do |command|
|
216
|
+
hash = {}
|
217
|
+
hash[:command] = command.name
|
218
|
+
hash[:options] = command.options
|
219
|
+
hash[:status] = command.status
|
220
|
+
puts hash.to_yaml
|
221
|
+
end
|
222
|
+
end
|
223
|
+
|
224
|
+
# Check all run options.
|
225
|
+
def check_options
|
226
|
+
options_allowed(@options, :debug, :verbose, :email, :progress, :subject,
|
227
|
+
:input, :output, :output_dir, :report, :force)
|
228
|
+
options_allowed_values(@options, debug: [true, false, nil])
|
229
|
+
options_allowed_values(@options, verbose: [true, false, nil])
|
230
|
+
options_conflict(@options, progress: :verbose)
|
231
|
+
options_tie(@options, subject: :email)
|
232
|
+
options_files_exist_force(@options, :report)
|
233
|
+
end
|
234
|
+
|
235
|
+
# Run all commands in the Pipeline.
|
236
|
+
def run_commands
|
237
|
+
prefix_output_dir
|
238
|
+
run_time_start
|
239
|
+
run_add_enumerators
|
240
|
+
run_enumerate
|
241
|
+
end
|
242
|
+
|
243
|
+
# Add start time to the status of all commands.
|
244
|
+
def run_time_start
|
245
|
+
time = Time.now
|
246
|
+
|
247
|
+
@commands.each do |command|
|
248
|
+
command.status[:time_start] = time
|
249
|
+
end
|
250
|
+
end
|
251
|
+
|
252
|
+
# Add enumerators to instance array.
|
253
|
+
def run_add_enumerators
|
254
|
+
@commands.each do |command|
|
255
|
+
input = @options[:input] || @enums.last
|
256
|
+
@enums << Enumerator.new { |output| command.call(input, output) }
|
257
|
+
end
|
258
|
+
end
|
259
|
+
|
260
|
+
# Iterate through all enumerators.
|
261
|
+
def run_enumerate
|
262
|
+
if @options[:output]
|
263
|
+
@enums.last.each { |record| @options[:output].write record }
|
264
|
+
@options[:output].close # TODO: this close is ugly here
|
265
|
+
else
|
266
|
+
@enums.last.each {}
|
267
|
+
end
|
268
|
+
end
|
269
|
+
|
270
|
+
# Create an output directory and prefix all output files in the commands
|
271
|
+
# with this directory.
|
272
|
+
def prefix_output_dir
|
273
|
+
return unless @options[:output_dir]
|
274
|
+
|
275
|
+
unless File.exist?(@options[:output_dir])
|
276
|
+
FileUtils.mkdir_p(@options[:output_dir])
|
277
|
+
end
|
278
|
+
|
279
|
+
@commands.each do |command|
|
280
|
+
if (value = command.options[:output])
|
281
|
+
command.options[:output] = File.join(@options[:output_dir], value)
|
282
|
+
end
|
283
|
+
end
|
284
|
+
end
|
285
|
+
|
286
|
+
# Save a HTML status report to file.
|
287
|
+
def save_report
|
288
|
+
return unless @options[:report]
|
289
|
+
|
290
|
+
file = if @options[:output_dir]
|
291
|
+
File.join(@options[:output_dir], @options[:report])
|
292
|
+
else
|
293
|
+
@options[:report]
|
294
|
+
end
|
295
|
+
|
296
|
+
File.open(file, 'w') do |ios|
|
297
|
+
ios.puts BioDSL::HtmlReport.new(self).to_html
|
298
|
+
end
|
299
|
+
end
|
300
|
+
|
301
|
+
# Run all commands.
|
302
|
+
def command_runner
|
303
|
+
return if @complete
|
304
|
+
|
305
|
+
if @options[:progress]
|
306
|
+
status_progress(@commands) { run_commands }
|
307
|
+
else
|
308
|
+
run_commands
|
309
|
+
end
|
310
|
+
|
311
|
+
@complete = true
|
312
|
+
end
|
313
|
+
|
314
|
+
# Set some global variables.
|
315
|
+
#
|
316
|
+
# @param options [Hash] Options hash.
|
317
|
+
# @option options [Booleon] :debug Debug flag.
|
318
|
+
# @option options [Booleon] :verbose Verbose flag.
|
319
|
+
def prime_variables(options)
|
320
|
+
BioDSL.test = ENV['BP_TEST']
|
321
|
+
BioDSL.debug = options[:debug]
|
322
|
+
BioDSL.verbose = options[:verbose]
|
323
|
+
end
|
324
|
+
|
325
|
+
# Output exception message and possibly stack tracre to STDERR,
|
326
|
+
# log error message and exit with non-zero status.
|
327
|
+
def exit_gracefully(exception)
|
328
|
+
fail exception if BioDSL.test
|
329
|
+
|
330
|
+
STDERR.puts "Error in run: #{exception.message}"
|
331
|
+
STDERR.puts exception.backtrace if BioDSL.verbose
|
332
|
+
log_error(exception)
|
333
|
+
exit 2
|
334
|
+
end
|
335
|
+
|
336
|
+
# Generate a comma separated string from the options ensuring that
|
337
|
+
# values are in "" if need be.
|
338
|
+
#
|
339
|
+
# Return [Array] List of options.
|
340
|
+
def options_string
|
341
|
+
options = []
|
342
|
+
|
343
|
+
@options.each_pair do |key, value|
|
344
|
+
if value.is_a? String
|
345
|
+
options << %(#{key}: "#{value}")
|
346
|
+
else
|
347
|
+
options << %(#{key}: #{value})
|
348
|
+
end
|
349
|
+
end
|
350
|
+
|
351
|
+
options.join(', ')
|
352
|
+
end
|
353
|
+
end
|
354
|
+
end
|
@@ -0,0 +1,66 @@
|
|
1
|
+
# >>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>><<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<< #
|
2
|
+
# #
|
3
|
+
# Copyright (C) 2007-2015 Martin Asser Hansen (mail@maasha.dk). #
|
4
|
+
# #
|
5
|
+
# This program is free software; you can redistribute it and/or #
|
6
|
+
# modify it under the terms of the GNU General Public License #
|
7
|
+
# as published by the Free Software Foundation; either version 2 #
|
8
|
+
# of the License, or (at your option) any later version. #
|
9
|
+
# #
|
10
|
+
# This program is distributed in the hope that it will be useful, #
|
11
|
+
# but WITHOUT ANY WARRANTY; without even the implied warranty of #
|
12
|
+
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the #
|
13
|
+
# GNU General Public License for more details. #
|
14
|
+
# #
|
15
|
+
# You should have received a copy of the GNU General Public License #
|
16
|
+
# along with this program; if not, write to the Free Software #
|
17
|
+
# Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, #
|
18
|
+
# USA. #
|
19
|
+
# #
|
20
|
+
# http://www.gnu.org/copyleft/gpl.html #
|
21
|
+
# #
|
22
|
+
# >>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>><<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<< #
|
23
|
+
# #
|
24
|
+
# This software is part of BioDSL (www.github.com/maasha/BioDSL). #
|
25
|
+
# #
|
26
|
+
# >>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>><<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<< #
|
27
|
+
|
28
|
+
# Namespace for BioDSL.
|
29
|
+
module BioDSL
|
30
|
+
# Namespace for Ambiguity.
|
31
|
+
module Ambiguity
|
32
|
+
# Add C functions to Inline::C object.
|
33
|
+
#
|
34
|
+
# @param inline_builder [Inline::C] Inline C object.
|
35
|
+
def add_ambiguity_macro(inline_builder)
|
36
|
+
# Macro for matching nucleotides including ambiguity codes.
|
37
|
+
inline_builder.prefix %(
|
38
|
+
#define MATCH(A,B) ((bitmap[(int) A] & bitmap[(int) B]) != 0)
|
39
|
+
)
|
40
|
+
|
41
|
+
# Bitmap for matching nucleotides including ambiguity codes.
|
42
|
+
# For each value bits are set from the left: bit pos 1 for A,
|
43
|
+
# bit pos 2 for T, bit pos 3 for C, and bit pos 4 for G.
|
44
|
+
inline_builder.prefix %(
|
45
|
+
char bitmap[256] = {
|
46
|
+
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
|
47
|
+
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
|
48
|
+
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
|
49
|
+
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
|
50
|
+
0, 1,14, 4,11, 0, 0, 8, 7, 0, 0,10, 0, 5,15, 0,
|
51
|
+
0, 0, 9,12, 2, 2,13, 3, 0, 6, 0, 0, 0, 0, 0, 0,
|
52
|
+
0, 1,14, 4,11, 0, 0, 8, 7, 0, 0,10, 0, 5,15, 0,
|
53
|
+
0, 0, 9,12, 2, 2,13, 3, 0, 6, 0, 0, 0, 0, 0, 0,
|
54
|
+
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
|
55
|
+
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
|
56
|
+
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
|
57
|
+
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
|
58
|
+
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
|
59
|
+
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
|
60
|
+
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
|
61
|
+
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
|
62
|
+
};
|
63
|
+
)
|
64
|
+
end
|
65
|
+
end
|
66
|
+
end
|
@@ -0,0 +1,240 @@
|
|
1
|
+
# >>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>><<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<< #
|
2
|
+
# #
|
3
|
+
# Copyright (C) 2007-2015 Martin Asser Hansen (mail@maasha.dk). #
|
4
|
+
# #
|
5
|
+
# This program is free software; you can redistribute it and/or #
|
6
|
+
# modify it under the terms of the GNU General Public License #
|
7
|
+
# as published by the Free Software Foundation; either version 2 #
|
8
|
+
# of the License, or (at your option) any later version. #
|
9
|
+
# #
|
10
|
+
# This program is distributed in the hope that it will be useful, #
|
11
|
+
# but WITHOUT ANY WARRANTY; without even the implied warranty of #
|
12
|
+
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the #
|
13
|
+
# GNU General Public License for more details. #
|
14
|
+
# #
|
15
|
+
# You should have received a copy of the GNU General Public License #
|
16
|
+
# along with this program; if not, write to the Free Software #
|
17
|
+
# Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, #
|
18
|
+
# USA. #
|
19
|
+
# #
|
20
|
+
# http://www.gnu.org/copyleft/gpl.html #
|
21
|
+
# #
|
22
|
+
# >>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>><<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<< #
|
23
|
+
# #
|
24
|
+
# This software is part of the BioDSL framework (www.BioDSL.org). #
|
25
|
+
# #
|
26
|
+
# >>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>><<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<< #
|
27
|
+
|
28
|
+
# Namespace for BioDSL.
|
29
|
+
module BioDSL
|
30
|
+
# Error class for all Assemble errors.
|
31
|
+
AssembleError = Class.new(StandardError)
|
32
|
+
|
33
|
+
# rubocop: disable ClassLength
|
34
|
+
|
35
|
+
# Class with methods for assembling pair-end reads.
|
36
|
+
class Assemble
|
37
|
+
require 'inline'
|
38
|
+
|
39
|
+
extend Ambiguity
|
40
|
+
|
41
|
+
# Class method to assemble two Seq objects.
|
42
|
+
def self.pair(entry1, entry2, options = {})
|
43
|
+
assemble = new(entry1, entry2, options)
|
44
|
+
assemble.match
|
45
|
+
end
|
46
|
+
|
47
|
+
# Method to initialize an Assembly object.
|
48
|
+
def initialize(entry1, entry2, options)
|
49
|
+
@entry1 = entry1
|
50
|
+
@entry2 = entry2
|
51
|
+
@overlap = 0
|
52
|
+
@offset1 = 0
|
53
|
+
@offset2 = 0
|
54
|
+
@options = options
|
55
|
+
@options[:mismatches_max] ||= 0
|
56
|
+
@options[:overlap_min] ||= 1
|
57
|
+
|
58
|
+
check_options
|
59
|
+
end
|
60
|
+
|
61
|
+
# Check option values are sane.
|
62
|
+
#
|
63
|
+
# @raise [AssembleError] on bad values.
|
64
|
+
def check_options
|
65
|
+
if @options[:mismatches_max] < 0
|
66
|
+
fail AssembleError, "mismatches_max must be zero or greater - not: \
|
67
|
+
#{@options[:mismatches_max]}"
|
68
|
+
end
|
69
|
+
|
70
|
+
if @options[:overlap_max] && @options[:overlap_max] <= 0
|
71
|
+
fail AssembleError, "overlap_max must be one or greater - not: \
|
72
|
+
#{@options[:overlap_max]}"
|
73
|
+
end
|
74
|
+
|
75
|
+
if @options[:overlap_min] <= 0
|
76
|
+
fail AssembleError, "overlap_min must be one or greater - not: \
|
77
|
+
#{@options[:overlap_min]}"
|
78
|
+
end
|
79
|
+
end
|
80
|
+
|
81
|
+
# Method to locate overlapping matches between two sequences.
|
82
|
+
def match
|
83
|
+
calc_overlap
|
84
|
+
diff = calc_diff
|
85
|
+
|
86
|
+
@offset1 = @entry1.length - @overlap - diff
|
87
|
+
|
88
|
+
while @overlap >= @options[:overlap_min]
|
89
|
+
mismatches_max = (@overlap * @options[:mismatches_max] * 0.01).round
|
90
|
+
|
91
|
+
if (mismatches = match_C(@entry1.seq, @entry2.seq, @offset1, @offset2,
|
92
|
+
@overlap, mismatches_max)) && mismatches >= 0
|
93
|
+
entry_merged = entry_left + entry_overlap + entry_right
|
94
|
+
entry_merged.seq_name = @entry1.seq_name +
|
95
|
+
":overlap=#{@overlap}:hamming=#{mismatches}" if @entry1.seq_name
|
96
|
+
|
97
|
+
return entry_merged
|
98
|
+
end
|
99
|
+
|
100
|
+
diff > 0 ? diff -= 1 : @overlap -= 1
|
101
|
+
|
102
|
+
@offset1 += 1
|
103
|
+
end
|
104
|
+
end
|
105
|
+
|
106
|
+
# Calculate the overlap to be matched.
|
107
|
+
def calc_overlap
|
108
|
+
@overlap = if @options[:overlap_max]
|
109
|
+
[@options[:overlap_max], @entry1.length, @entry2.length].min
|
110
|
+
else
|
111
|
+
[@entry1.length, @entry2.length].min
|
112
|
+
end
|
113
|
+
end
|
114
|
+
|
115
|
+
# Calculate the diff between sequence lengths and return this.
|
116
|
+
#
|
117
|
+
# @return [Fixnum] Diff.
|
118
|
+
def calc_diff
|
119
|
+
diff = @entry1.length - @entry2.length
|
120
|
+
diff = 0 if diff < 0
|
121
|
+
diff
|
122
|
+
end
|
123
|
+
|
124
|
+
# Method to extract and downcase the left part of an assembled pair.
|
125
|
+
#
|
126
|
+
# @return [BioDSL::Seq] Left part.
|
127
|
+
def entry_left
|
128
|
+
entry = @entry1[0...@offset1]
|
129
|
+
entry.seq.downcase!
|
130
|
+
entry
|
131
|
+
end
|
132
|
+
|
133
|
+
# Method to extract and downcase the right part of an assembled pair.
|
134
|
+
#
|
135
|
+
# @return [BioDSL::Seq] Right part.
|
136
|
+
def entry_right
|
137
|
+
entry = if @entry1.length > @offset1 + @overlap
|
138
|
+
@entry1[@offset1 + @overlap..-1]
|
139
|
+
else
|
140
|
+
@entry2[@offset2 + @overlap..-1]
|
141
|
+
end
|
142
|
+
|
143
|
+
entry.seq.downcase!
|
144
|
+
entry
|
145
|
+
end
|
146
|
+
|
147
|
+
# Method to extract and upcase the overlapping part of an assembled pair.
|
148
|
+
#
|
149
|
+
# @return [BioDSL::Seq] Overlapping part.
|
150
|
+
def entry_overlap
|
151
|
+
if @entry1.qual && @entry2.qual
|
152
|
+
entry_overlap1 = @entry1[@offset1...@offset1 + @overlap]
|
153
|
+
entry_overlap2 = @entry2[@offset2...@offset2 + @overlap]
|
154
|
+
|
155
|
+
entry = merge_overlap(entry_overlap1, entry_overlap2)
|
156
|
+
else
|
157
|
+
entry = @entry1[@offset1...@offset1 + @overlap]
|
158
|
+
end
|
159
|
+
|
160
|
+
entry.seq.upcase!
|
161
|
+
entry
|
162
|
+
end
|
163
|
+
|
164
|
+
# Method to merge sequence and quality scores in an overlap.
|
165
|
+
# The residue with the highest score at mismatch positions is selected.
|
166
|
+
# The quality scores of the overlap are the mean of the two sequences.
|
167
|
+
def merge_overlap(entry_overlap1, entry_overlap2)
|
168
|
+
na_seq = NArray.byte(entry_overlap1.length, 2)
|
169
|
+
na_seq[true, 0] = NArray.to_na(entry_overlap1.seq.downcase, 'byte')
|
170
|
+
na_seq[true, 1] = NArray.to_na(entry_overlap2.seq.downcase, 'byte')
|
171
|
+
|
172
|
+
na_qual = NArray.byte(entry_overlap1.length, 2)
|
173
|
+
na_qual[true, 0] = NArray.to_na(entry_overlap1.qual, 'byte')
|
174
|
+
na_qual[true, 1] = NArray.to_na(entry_overlap2.qual, 'byte')
|
175
|
+
|
176
|
+
mask_xor = na_seq[true, 0] ^ na_seq[true, 1] > 0
|
177
|
+
mask_seq = ((na_qual * mask_xor).eq((na_qual * mask_xor).max(1)))
|
178
|
+
|
179
|
+
merged = Seq.new
|
180
|
+
merged.seq = (na_seq * mask_seq).max(1).to_s
|
181
|
+
merged.qual = na_qual.mean(1).round.to_type('byte').to_s
|
182
|
+
|
183
|
+
merged
|
184
|
+
end
|
185
|
+
|
186
|
+
inline do |builder|
|
187
|
+
add_ambiguity_macro(builder)
|
188
|
+
|
189
|
+
# C method for determining if two strings of equal length match
|
190
|
+
# given a maximum allowed mismatches and allowing for IUPAC
|
191
|
+
# ambiguity codes. Returns number of mismatches is true if match, else
|
192
|
+
# false.
|
193
|
+
builder.c %{
|
194
|
+
VALUE match_C(
|
195
|
+
VALUE _string1, // String 1
|
196
|
+
VALUE _string2, // String 2
|
197
|
+
VALUE _offset1, // Offset 1
|
198
|
+
VALUE _offset2, // Offset 2
|
199
|
+
VALUE _length, // String length
|
200
|
+
VALUE _max_mismatch // Maximum mismatches
|
201
|
+
)
|
202
|
+
{
|
203
|
+
char *string1 = StringValuePtr(_string1);
|
204
|
+
char *string2 = StringValuePtr(_string2);
|
205
|
+
unsigned int offset1 = FIX2UINT(_offset1);
|
206
|
+
unsigned int offset2 = FIX2UINT(_offset2);
|
207
|
+
unsigned int length = FIX2UINT(_length);
|
208
|
+
unsigned int max_mismatch = FIX2UINT(_max_mismatch);
|
209
|
+
|
210
|
+
unsigned int max_match = length - max_mismatch;
|
211
|
+
unsigned int match = 0;
|
212
|
+
unsigned int mismatch = 0;
|
213
|
+
unsigned int i = 0;
|
214
|
+
|
215
|
+
for (i = 0; i < length; i++)
|
216
|
+
{
|
217
|
+
if (MATCH(string1[i + offset1], string2[i + offset2]))
|
218
|
+
{
|
219
|
+
match++;
|
220
|
+
|
221
|
+
if (match >= max_match) {
|
222
|
+
return UINT2NUM(mismatch);
|
223
|
+
}
|
224
|
+
}
|
225
|
+
else
|
226
|
+
{
|
227
|
+
mismatch++;
|
228
|
+
|
229
|
+
if (mismatch > max_mismatch) {
|
230
|
+
return INT2NUM(-1);
|
231
|
+
}
|
232
|
+
}
|
233
|
+
}
|
234
|
+
|
235
|
+
return INT2NUM(-1);
|
236
|
+
}
|
237
|
+
}
|
238
|
+
end
|
239
|
+
end
|
240
|
+
end
|