Class: StaticEmbeddings::CLI

Inherits:
Object
  • Object
show all
Defined in:
lib/static_embeddings/cli.rb

Constant Summary collapse

COMMANDS = 
{
  "convert" => :convert,
  "verify" => :verify,
  "inspect" => :inspect_model,
  "tokenize" => :tokenize,
  "embed" => :embed,
  "cache-path" => :cache_path,
  "help" => :usage,
  "-h" => :usage,
  "--help" => :usage,
  nil => :usage
}.freeze
HELP = 
<<~TEXT
  static_embeddings <command> [options]

    convert SOURCE_DIR    Convert a HuggingFace/Model2Vec directory to .semb
      --out PATH          Output file (default: <cache>/models/<id>.semb)
      --id ID             Model id recorded in provenance
      --max-tokens N      Truncation limit baked into the file (default 512)

    verify PATH           Recompute the SHA-256 embedded in the header
    inspect PATH          Print header fields and provenance
    tokenize PATH TEXT    Print token ids
    embed PATH TEXT       Print the vector, token count and UNK ratio
    cache-path            Print the model cache directory
TEXT

Class Method Summary collapse

Instance Method Summary collapse

Class Method Details

.run(argv) ⇒ Object 



35
36
37
 
# File 'lib/static_embeddings/cli.rb', line 35

def self.run(argv)
  new.run(argv)
end

Instance Method Details

#run(argv) ⇒ Object 



39
40
41
42
43
44
45
46
47
 
# File 'lib/static_embeddings/cli.rb', line 39

def run(argv)
  method = COMMANDS[argv.shift]
  return unknown unless method

  send(method, argv)
rescue StaticEmbeddings::Error => e
  warn "#{e.class.name.split('::').last}: #{e.message}"
  1
end