Class: Ollama::Response

Inherits:
Object
  • Object
show all
Defined in:
lib/ollama/response.rb

Overview

Response wrapper for chat() that provides method access to response data

Example:

response = client.chat(messages: [...])
response.message&.content      # Access content
response.message&.thinking     # Access thinking output
response.message&.tool_calls   # Access tool_calls
response.done?                 # Check if generation finished
response.done_reason           # Why generation stopped
response.total_duration        # Total time in nanoseconds

Direct Known Subclasses

Ollama::Responses::Base

Defined Under Namespace

Classes: Message

Instance Method Summary collapse

Constructor Details

#initialize(data) ⇒ Response

Returns a new instance of Response.



17
18
19
 
# File 'lib/ollama/response.rb', line 17

def initialize(data)
  @data = data || {}
end

Dynamic Method Handling

This class handles dynamic methods through the method_missing method

#method_missing(method) ⇒ Object

Delegate unknown methods to underlying hash for forward compatibility. This allows accessing any raw response fields not covered by explicit accessors.



121
122
123
124
125
 
# File 'lib/ollama/response.rb', line 121

def method_missing(method, ...)
  return super unless @data.respond_to?(method)

  @data.public_send(method, ...)
end

Instance Method Details

#[](key) ⇒ Object

Delegate hash access to underlying data



115
116
117
 
# File 'lib/ollama/response.rb', line 115

def [](key)
  @data[key]
end

#content ⇒ Object

Convenient content accessor (shorthand for message&.content)



110
111
112
 
# File 'lib/ollama/response.rb', line 110

def content
  message&.content
end

#created_at ⇒ Object

ISO 8601 timestamp of response creation



45
46
47
 
# File 'lib/ollama/response.rb', line 45

def created_at
  @data["created_at"] || @data[:created_at]
end

#done? ⇒ Boolean

Whether generation has finished

Returns:

  • (Boolean) 


30
31
32
 
# File 'lib/ollama/response.rb', line 30

def done?
  @data["done"] || @data[:done] || false
end

#done_reason ⇒ Object

Reason the generation stopped



35
36
37
 
# File 'lib/ollama/response.rb', line 35

def done_reason
  @data["done_reason"] || @data[:done_reason]
end

#eval_count ⇒ Object

Number of tokens generated in the response



71
72
73
 
# File 'lib/ollama/response.rb', line 71

def eval_count
  @data["eval_count"] || @data[:eval_count] || @data.dig("usage", "completion_tokens") || @data.dig(:usage, :completion_tokens)
end

#eval_duration ⇒ Object

Time spent generating tokens in nanoseconds



76
77
78
 
# File 'lib/ollama/response.rb', line 76

def eval_duration
  @data["eval_duration"] || @data[:eval_duration]
end

#latency_ms ⇒ Float?

Total generation latency in milliseconds (converted from nanoseconds). Returns nil when timing data is unavailable.

Returns:

  • (Float, nil) 


98
99
100
101
102
 
# File 'lib/ollama/response.rb', line 98

def latency_ms
  return nil unless total_duration

  (total_duration / 1_000_000.0).round(2)
end

#load_duration ⇒ Object

Time spent loading the model in nanoseconds



55
56
57
 
# File 'lib/ollama/response.rb', line 55

def load_duration
  @data["load_duration"] || @data[:load_duration]
end

#logprobs ⇒ Object

Log probability information when logprobs are enabled



81
82
83
 
# File 'lib/ollama/response.rb', line 81

def logprobs
  @data["logprobs"] || @data[:logprobs]
end

#message ⇒ Object

Access the message object



22
23
24
25
26
27
 
# File 'lib/ollama/response.rb', line 22

def message
  msg = @data["message"] || @data[:message]
  return nil unless msg

  Message.new(msg)
end

#model ⇒ Object

Model name used



40
41
42
 
# File 'lib/ollama/response.rb', line 40

def model
  @data["model"] || @data[:model]
end

#prompt_eval_count ⇒ Object

Number of tokens in the prompt



60
61
62
63
 
# File 'lib/ollama/response.rb', line 60

def prompt_eval_count
  @data["prompt_eval_count"] || @data[:prompt_eval_count] || @data.dig("usage",
                                                                       "prompt_tokens") || @data.dig(:usage, :prompt_tokens)
end

#prompt_eval_duration ⇒ Object

Time spent evaluating the prompt in nanoseconds



66
67
68
 
# File 'lib/ollama/response.rb', line 66

def prompt_eval_duration
  @data["prompt_eval_duration"] || @data[:prompt_eval_duration]
end

#respond_to_missing?(method, include_private = false) ⇒ Boolean

Returns:

  • (Boolean) 


127
128
129
 
# File 'lib/ollama/response.rb', line 127

def respond_to_missing?(method, include_private = false)
  @data.respond_to?(method, include_private) || super
end

#to_h ⇒ Object

Access raw data as hash



105
106
107
 
# File 'lib/ollama/response.rb', line 105

def to_h
  @data
end

#total_duration ⇒ Object

Total time spent generating in nanoseconds



50
51
52
 
# File 'lib/ollama/response.rb', line 50

def total_duration
  @data["total_duration"] || @data[:total_duration]
end

#usage ⇒ Hash

Aggregated token usage for observability and cost tracking.

Returns:

  • (Hash) —

    { prompt_tokens:, completion_tokens:, total_tokens: }



87
88
89
90
91
92
93
 
# File 'lib/ollama/response.rb', line 87

def usage
  {
    prompt_tokens: prompt_eval_count,
    completion_tokens: eval_count,
    total_tokens: (prompt_eval_count.to_i + eval_count.to_i).then { |v| v.zero? ? nil : v }
  }.compact
end