Class: Bioroebe::ShowFastaHeaders

Inherits:
CommandlineApplication show all
Defined in:
lib/bioroebe/fasta_and_fastq/show_fasta_headers.rb

Overview

Biroebe::ShowFastaHeaders

Constant Summary collapse

BASE_DIR =
#

BASE_DIR

This base-directory will only work on my home setup, by default.

#
'/home/x/data/science/bioinformatics/data/fasta/'

Constants inherited from CommandlineApplication

CommandlineApplication::OLD_VERBOSE_VALUE

Constants included from ColoursForBase

ColoursForBase::ARRAY_HTML_COLOURS_IN_USE

Constants inherited from Base

Base::NAMESPACE

Instance Method Summary collapse

Methods inherited from CommandlineApplication

#all_aminoacids?, #append_what_into, #at_home?, #be_silent, #be_verbose?, #cat, #ccliner, #change_directory, #cliner, #codon_table_dataset?, #codon_to_aminoacid, #codons_for?, #colourize_this_dna_sequence, #complement, #cp, #disable_warnings, #download_dir?, #editor?, #enable_warnings, #ensure_that_the_base_directories_exist, #esystem, #extract, #is_this_a_start_codon?, #is_this_a_stop_codon?, #leading_five_prime, #load_bioroebe_yaml_file, #log_directory?, #one_letter_to_long_name, #one_to_three, #only_numbers?, #open_in_browser, #opnerev, #opnn, #pad_with_double_quotes, #pad_with_single_quotes, #partner_nucleotide, #remove_numbers, #remove_trailing_ansii_escape_code, #return_all_possible_start_codons, #return_array_of_one_letter_aminoacids, #return_cheerful_person, #return_chunked_display, #return_ubiquitin_sequence, #runmode?, #set_be_verbose, #set_runmode, #start_codon?, #stop_codons?, #strict_filter_away_invalid_aminoacids, #taxonomy_download_directory?, #three_to_one, #to_rna, #trailing_three_prime, #use_opn?, #verbose_truth, #was_or_were, #without_extname, #write_what_into

Methods included from BaseModule

#absolute_path, #default_file_read, #file_readlines

Methods included from CommandlineArguments

#commandline_arguments?, #commandline_arguments_that_are_files?, #e, #first?, #first_non_hyphen_argument?, #remove_hyphens_from_the_commandline_arguments, #return_commandline_arguments_as_string, #return_commandline_arguments_that_are_not_files, #return_entries_without_two_leading_hyphens, #select_commandline_arguments, #select_entries_starting_with_two_hyphens, #set_commandline_arguments

Methods included from ColoursForBase

#colourize_this_aminoacid_sequence_for_the_commandline, #colourize_this_nucleotide_sequence, #disable_colours, #ecomment, #efancy, #egold, #enable_colours, #eorange, #eparse, #erev, #red, #remove_trailing_escape_part, #return_colour_for_nucleotides, #rev, #sdir, #set_will_we_use_colours, #sfancy, #sfile, #simp, #swarn, #use_colours?, #use_colours_within_the_bioroebe_namespace?

Methods inherited from Base

#append_what_into, #can_base_pair?, #convert_global_env, #delete_file, #directory_to_the_codon_tables?, #is_on_roebe?, #is_palindrome?, #main_encoding?, #mkdir, #move_file, #mv, #no_file_exists_at, #no_newlines, #project_yaml_directory?, #rds, #register_sigint, #return_pwd, #return_the_first_line_of_this_file, #word_wrap, #write_what_into

Methods included from InternalHashModule

#internal_hash?, #reset_the_internal_hash

Methods included from InferTheNamespaceModule

#infer_the_namespace, #namespace?

Constructor Details

#initialize(i = BASE_DIR, run_already = true) ⇒ ShowFastaHeaders

#

initialize

#


33
34
35
36
37
38
39
40
 
# File 'lib/bioroebe/fasta_and_fastq/show_fasta_headers.rb', line 33

def initialize(
    i           = BASE_DIR,
    run_already = true
  )
  reset
  set_input(i)
  run if run_already
end

Instance Method Details

#input? ⇒ Boolean

#

input?

#

Returns:

  • (Boolean) 


77
78
79
 
# File 'lib/bioroebe/fasta_and_fastq/show_fasta_headers.rb', line 77

def input?
  @input
end

#input_without_extension ⇒ Object

#

input_without_extension

#


84
85
86
 
# File 'lib/bioroebe/fasta_and_fastq/show_fasta_headers.rb', line 84

def input_without_extension
  input?.sub(/\*\.ffn$/,'')
end

#report_these_files(files) ⇒ Object

#

report_these_files

#


91
92
93
94
95
96
97
98
 
# File 'lib/bioroebe/fasta_and_fastq/show_fasta_headers.rb', line 91

def report_these_files(files)
  files.each {|this_file|
    dataset = File.readlines(this_file)
    dataset.reject! {|line| line.strip.start_with? '#' }
    header = dataset.first
    report_this_header_from_that_file(header, this_file)
  }
end

#report_this_header_from_that_file(header, file) ⇒ Object

#

report_this_header_from_that_file

#


103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
 
# File 'lib/bioroebe/fasta_and_fastq/show_fasta_headers.rb', line 103

def report_this_header_from_that_file(header, file)
  if be_verbose?
    erev "The file at #{sfile(file)}#{rev} has the following FASTA header:"
    e
  end
  e "  #{simp(header)}"
  if header.include? '|'
    splitted = header.split('|')
    if splitted.size > 2
      assumed_name_of_the_organism = splitted[2].strip
      if assumed_name_of_the_organism.include? 'OS='
        assumed_name_of_the_organism = assumed_name_of_the_organism.scan(/ OS=(.+)/).flatten
        assumed_name_of_the_organism = assumed_name_of_the_organism.first
        if assumed_name_of_the_organism.include? ' ('
          assumed_name_of_the_organism = assumed_name_of_the_organism.split(' (').first
        end
      end
      erev 'The organism may be: '+
           steelblue(assumed_name_of_the_organism)+rev
    end
  end
end

#reset ⇒ Object

#

reset (reset tag)

#


45
46
47
48
49
50
51
52
 
# File 'lib/bioroebe/fasta_and_fastq/show_fasta_headers.rb', line 45

def reset
  super()
  infer_the_namespace
  # ======================================================================= #
  # === :be_verbose
  # ======================================================================= #
  set_be_verbose
end

#run ⇒ Object

#

run (run tag)

#


129
130
131
132
133
134
135
136
 
# File 'lib/bioroebe/fasta_and_fastq/show_fasta_headers.rb', line 129

def run
  files = Dir[input?]
  if files.empty?
    opnn; e "No files could be found for #{input?}."
  else
    report_these_files(files)
  end
end

#set_input(i = '') ⇒ Object

#

set_input

#


57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
 
# File 'lib/bioroebe/fasta_and_fastq/show_fasta_headers.rb', line 57

def set_input(i = '')
  i = i.first if i.is_a? Array
  i = i.to_s.dup
  if i.empty?
    i = BASE_DIR
  end
  i = i.dup if i.frozen?
  if i.end_with?('/') and !File.directory?(i) # Reject invalid directories here.
    i.chop!
  end
  if File.directory? i
    i << '/*.ffn'
  end
  i.squeeze!('/')
  @input = i
end