Class: Cohere::Transcribe::Hub

Inherits:
Object
  • Object
show all
Defined in:
lib/cohere/transcribe/hub.rb

Overview

Minimal Hugging Face Hub client used by the native Ruby runtime. It uses the standard Hub cache layout, so artifacts already fetched by other Hub clients are reused without copying multi-gigabyte model weights.

Defined Under Namespace

Classes: AuthenticationError, ConnectionError, Error, NotFoundError

Constant Summary collapse

COMMIT_PATTERN = 
/\A[0-9a-f]{40}\z/i
DEFAULT_ENDPOINT = 
"https://huggingface.co"

Instance Attribute Summary collapse

Instance Method Summary collapse

Constructor Details

#initialize(cache_dir: nil, endpoint: nil, token: nil, offline: nil) ⇒ Hub

Returns a new instance of Hub.



26
27
28
29
30
31
32
33
 
# File 'lib/cohere/transcribe/hub.rb', line 26

def initialize(cache_dir: nil, endpoint: nil, token: nil, offline: nil)
  hf_home = ENV.fetch("HF_HOME", File.expand_path("~/.cache/huggingface"))
  @cache_dir = Pathname(cache_dir || ENV.fetch("HF_HUB_CACHE", File.join(hf_home, "hub"))).expand_path
  @endpoint = (endpoint || ENV.fetch("HF_ENDPOINT", DEFAULT_ENDPOINT)).sub(%r{/+\z}, "")
  @endpoint_uri = URI(@endpoint)
  @token = token || ENV["HF_TOKEN"] || cached_token(hf_home)
  @offline = offline.nil? ? truthy_environment?(ENV.fetch("HF_HUB_OFFLINE", nil)) : !!offline
end

Instance Attribute Details

#cache_dir ⇒ Object (readonly)

Returns the value of attribute cache_dir.



24
25
26
 
# File 'lib/cohere/transcribe/hub.rb', line 24

def cache_dir
  @cache_dir
end

#endpoint ⇒ Object (readonly)

Returns the value of attribute endpoint.



24
25
26
 
# File 'lib/cohere/transcribe/hub.rb', line 24

def endpoint
  @endpoint
end

Instance Method Details

#cached_file(repo_id, filename, revision: nil) ⇒ Object 



132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
 
# File 'lib/cohere/transcribe/hub.rb', line 132

def cached_file(repo_id, filename, revision: nil)
  validate_repo_id!(repo_id)
  validate_filename!(filename)
  validate_revision!(revision) if revision
  commit = if revision && COMMIT_PATTERN.match?(revision)
             revision.downcase
           else
             cached_revision(repo_id, revision || "main", filename: filename)
           end
  return unless commit

  repository = cache_dir.join(cache_repo_name(repo_id))
  path = repository.join("snapshots", commit, filename)
  safe_cached_file(repository, path)
end

#download(repo_id, filename, revision: nil) ⇒ Object 



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
 
# File 'lib/cohere/transcribe/hub.rb', line 74

def download(repo_id, filename, revision: nil)
  validate_filename!(filename)
  commit = resolve_revision(repo_id, revision, filename: filename)
  repository = cache_dir.join(cache_repo_name(repo_id))
  destination = repository.join("snapshots", commit, filename)
  return destination if safe_cached_file(repository, destination)

  prepare_cache_directory!(repository, destination.dirname)
  encoded_repo = repo_id.split("/").map { |part| URI.encode_www_form_component(part) }.join("/")
  encoded_filename = filename.split("/").map { |part| URI.encode_www_form_component(part) }.join("/")
  uri = URI("#{endpoint}/#{encoded_repo}/resolve/#{commit}/#{encoded_filename}")

  open_download_lock(download_lock_path(destination)) do |lock|
    raise Error, "Cannot acquire Hub download lock for #{destination}" unless lock.flock(File::LOCK_EX)
    return destination if safe_cached_file(repository, destination)

    cleanup_download_temporaries(destination)
    Tempfile.create(
      [download_temporary_prefix(destination), ".download"],
      destination.dirname.to_s,
      binmode: true
    ) do |temporary|
      request(uri, stream: temporary)
      temporary.flush
      temporary.fsync
      temporary.close
      File.rename(temporary.path, destination)
    end
    sync_directory(destination.dirname)
  end
  destination
rescue Errno::EACCES, Errno::ENOSPC, Errno::EROFS => e
  raise Error, "Cannot cache #{repo_id}/#{filename}: #{e.message}"
end

#list_files(repo_id, revision: nil) ⇒ Object 



109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
 
# File 'lib/cohere/transcribe/hub.rb', line 109

def list_files(repo_id, revision: nil)
  commit = resolve_revision(repo_id, revision)
  encoded_repo = repo_id.split("/").map { |part| URI.encode_www_form_component(part) }.join("/")
  response = request(URI("#{endpoint}/api/models/#{encoded_repo}/revision/#{commit}"))
  payload = JSON.parse(response.body)
  siblings = payload["siblings"]
  raise Error, "Hub returned no repository file list for #{repo_id}@#{commit}" unless siblings.is_a?(Array)

  siblings.filter_map do |item|
    name = item.is_a?(Hash) ? item["rfilename"] : nil
    name if name.is_a?(String)
  end.freeze
rescue JSON::ParserError => e
  raise Error, "Invalid Hub response while listing #{repo_id.inspect}: #{e.message}"
end

#offline? ⇒ Boolean

Returns:

  • (Boolean) 


35
36
37
 
# File 'lib/cohere/transcribe/hub.rb', line 35

def offline?
  @offline
end

#resolve_revision(repo_id, revision = nil, filename: "config.json") ⇒ Object 



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
 
# File 'lib/cohere/transcribe/hub.rb', line 39

def resolve_revision(repo_id, revision = nil, filename: "config.json")
  validate_repo_id!(repo_id)
  requested = revision || "main"
  validate_revision!(requested)
  return requested.downcase if COMMIT_PATTERN.match?(requested)

  cached = cached_revision(repo_id, requested, filename: filename)
  if offline?
    return cached if cached

    raise Error,
          "Hub offline mode has no cached #{filename} snapshot for #{repo_id.inspect} at #{requested.inspect}"
  end

  encoded_repo = repo_id.split("/").map { |part| URI.encode_www_form_component(part) }.join("/")
  encoded_revision = URI.encode_www_form_component(requested)
  begin
    response = request(URI("#{endpoint}/api/models/#{encoded_repo}/revision/#{encoded_revision}"))
  rescue ConnectionError
    raise unless cached

    return cached
  end
  payload = JSON.parse(response.body)
  commit = payload["sha"]
  unless commit.is_a?(String) && COMMIT_PATTERN.match?(commit)
    raise Error, "Hub returned no immutable commit for #{repo_id.inspect} at #{requested.inspect}"
  end

  write_ref(repo_id, requested, commit.downcase)
  commit.downcase
rescue JSON::ParserError => e
  raise Error, "Invalid Hub response while resolving #{repo_id.inspect}: #{e.message}"
end

#snapshot_path(repo_id, commit) ⇒ Object

Raises:

  • (ArgumentError) 


125
126
127
128
129
130
 
# File 'lib/cohere/transcribe/hub.rb', line 125

def snapshot_path(repo_id, commit)
  validate_repo_id!(repo_id)
  raise ArgumentError, "Invalid immutable Hub commit: #{commit.inspect}" unless commit.is_a?(String) && COMMIT_PATTERN.match?(commit)

  cache_dir.join(cache_repo_name(repo_id), "snapshots", commit)
end