Module: MiGA::Cli::Action::Download::Base

Included in:
Gtdb, Ncbi, Seqcode
Defined in:
lib/miga/cli/action/download/base.rb

Overview

Helper module including download functions for the *_get actions

Instance Method Summary collapse

Instance Method Details

#cli_base_flags(opt) ⇒ Object 



10
11
12
13
14
15
16
17
18
19
20
 
# File 'lib/miga/cli/action/download/base.rb', line 10

def cli_base_flags(opt)
  opt.on(
    '--max INT', Integer,
    'Maximum number of datasets to download (by default: unlimited)'
  ) { |v| cli[:max_datasets] = v }
  opt.on(
    '-m', '--metadata STRING',
    'Metadata as key-value pairs separated by = and delimited by comma',
    'Values are saved as strings except for booleans (true / false) or nil'
  ) { |v| cli[:metadata] = v }
end

#cli_filters(opt) ⇒ Object 



22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
 
# File 'lib/miga/cli/action/download/base.rb', line 22

def cli_filters(opt)
  opt.on(
    '--exclude PATH',
    'A file with dataset names to exclude'
  ) { |v| cli[:exclude] = v }
  cli.opt_flag(opt, 'dry', 'Do not download or save the datasets')
  opt.on(
    '--ignore-until STRING',
    'Ignores all datasets until a name is found (useful for large reruns)'
  ) { |v| cli[:ignore_until] = v }
  opt.on(
    '--ignore-removed',
    'Ignores entries removed from NCBI (by default fails on removed entries)'
  ) { |v| cli[:ignore_removed] = v }
  cli.opt_flag(
    opt, 'get-metadata',
    'Only download and update metadata for existing datasets', :get_md
  )
end

#cli_save_actions(opt) ⇒ Object 



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
 
# File 'lib/miga/cli/action/download/base.rb', line 42

def cli_save_actions(opt)
  cli.opt_flag(
    opt, 'only-metadata',
    'Create datasets without input data but retrieve all metadata',
    :only_md
  )
  opt.on(
    '--save-every INT', Integer,
    'Save project every this many downloaded datasets',
    'If zero, it saves the project only once upon completion',
    "By default: #{cli[:save_every]}"
  ) { |v| cli[:save_every] = v }
  opt.on(
    '-q', '--query',
    'Register the datasets as queries, not reference datasets'
  ) { |v| cli[:query] = v }
  opt.on(
    '-u', '--unlink',
    'Unlink all datasets in the project missing from the download list'
  ) { |v| cli[:unlink] = v }
  opt.on(
    '-R', '--remote-list PATH',
    'Path to an output file with the list of all datasets listed remotely'
  ) { |v| cli[:remote_list] = v }
end

#discard_excluded(ds) ⇒ Object 



102
103
104
105
106
107
108
109
110
111
 
# File 'lib/miga/cli/action/download/base.rb', line 102

def discard_excluded(ds)
  unless cli[:exclude].nil?
    cli.say "Discarding datasets in #{cli[:exclude]}"
    File.readlines(cli[:exclude])
        .select { |i| i !~ /^#/ }
        .map(&:chomp)
        .each { |i| ds.delete i }
  end
  ds
end

#download_entries(ds, p) ⇒ Object 



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
 
# File 'lib/miga/cli/action/download/base.rb', line 123

def download_entries(ds, p)
  cli.say "Downloading #{ds.size} " + (ds.size == 1 ? 'entry' : 'entries')
  p.do_not_save = true if cli[:save_every] != 1
  ignore = !cli[:ignore_until].nil?
  downloaded = 0
  d = []
  ds.each do |name, body|
    d << name
    cli.puts name
    ignore = false if ignore && name == cli[:ignore_until]
    next if ignore || p.dataset(name).nil? == cli[:get_md]

    downloaded += 1
    unless cli[:dry]
      unless save_entry(name, body, p)
        downloaded -= 1
        d.pop
        next
      end
      p.save! if cli[:save_every] > 1 && (downloaded % cli[:save_every]).zero?
    end
  end
  p.do_not_save = false
  p.save! if cli[:save_every] != 1
  [d, downloaded]
end

#finalize_tasks(d, downloaded) ⇒ Object 



86
87
88
89
90
91
92
93
94
95
 
# File 'lib/miga/cli/action/download/base.rb', line 86

def finalize_tasks(d, downloaded)
  cli.say "Datasets listed: #{d.size}"
  act = cli[:dry] ? 'to download' : 'downloaded'
  cli.say "Datasets #{act}: #{downloaded}"
  unless cli[:remote_list].nil?
    File.open(cli[:remote_list], 'w') do |fh|
      d.each { |i| fh.puts i }
    end
  end
end

#generic_perform ⇒ Object 



68
69
70
71
72
73
74
75
 
# File 'lib/miga/cli/action/download/base.rb', line 68

def generic_perform
  p, ds = load_tasks
  d, downloaded = download_entries(ds, p)

  # Finalize
  finalize_tasks(d, downloaded)
  unlink_entries(p, p.dataset_names - d) if cli[:unlink]
end

#impose_limit(ds) ⇒ Object 



113
114
115
116
117
118
119
120
121
 
# File 'lib/miga/cli/action/download/base.rb', line 113

def impose_limit(ds)
  max = cli[:max_datasets].to_i
  if !max.zero? && max < ds.size
    cli.say "Subsampling list from #{ds.size} to #{max} datasets"
    sample = ds.keys.sample(max)
    ds.select! { |k, _| sample.include? k }
  end
  ds
end

#load_tasks ⇒ Object 



77
78
79
80
81
82
83
84
 
# File 'lib/miga/cli/action/download/base.rb', line 77

def load_tasks
  sanitize_cli
  p = cli.load_project
  ds = remote_list
  ds = discard_excluded(ds)
  ds = impose_limit(ds)
  [p, ds]
end

#save_entry(name, body, p) ⇒ Object

Saves the (generic remote) entry identified by name with body into the project p, and returns true on success and false otherwise



153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
 
# File 'lib/miga/cli/action/download/base.rb', line 153

def save_entry(name, body, p)
  cli.say "  Locating remote dataset: #{name}"
  body[:md][:metadata_only] = true if cli[:only_md]
  rd = MiGA::RemoteDataset.new(body[:ids], body[:db], body[:universe])
  if cli[:get_md]
    cli.say '  Updating dataset'
    rd.update_metadata(p.dataset(name), body[:md])
  else
    cli.say '  Creating dataset'
    rd.save_to(p, name, !cli[:query], body[:md])
    cli.add_metadata(p.add_dataset(name))
  end
  true
rescue MiGA::RemoteDataMissingError => e
  raise(e) unless cli[:ignore_removed]
  cli.say "    Removed dataset ignored: #{name}"
  false
end

#unlink_entries(p, unlink) ⇒ Object 



97
98
99
100
 
# File 'lib/miga/cli/action/download/base.rb', line 97

def unlink_entries(p, unlink)
  unlink.each { |i| p.unlink_dataset(i).remove! }
  cli.say "Datasets unlinked: #{unlink.size}"
end