Class: Bitfab::Datasets

Inherits:
Object
  • Object
show all
Defined in:
lib/bitfab/datasets.rb

Overview

Dataset operations for the authenticated organization, reached as client.datasets. A dataset is a named bucket of traces scoped to one trace function. Experiments replay against it and its graders score its members. Every method returns the parsed JSON response as a Hash with string keys matching the HTTP API ("traceCount", "addedTraceIds", ...).

Constant Summary collapse

DEFAULT_RERUN_TIMEOUT_SECONDS = 
90
DEFAULT_RERUN_POLL_INTERVAL_SECONDS = 
1
TERMINAL_RERUN_STATUSES = 
["completed", "errored"].freeze

Instance Method Summary collapse

Constructor Details

#initialize(http_client) ⇒ Datasets

Returns a new instance of Datasets.



14
15
16
 
# File 'lib/bitfab/datasets.rb', line 14

def initialize(http_client)
  @http_client = http_client
end

Instance Method Details

#add_graders(dataset_id, grader_ids) ⇒ Object

Assign graders to the dataset (1 to 100 ids per call). Graders outside the organization or under another trace function are reported in "skippedGraderIds" rather than failing the call.



61
62
63
 
# File 'lib/bitfab/datasets.rb', line 61

def add_graders(dataset_id, grader_ids)
  @http_client.request(dataset_path(dataset_id, "/graders"), {"graderIds" => grader_ids})
end

#add_traces(dataset_id, trace_ids) ⇒ Object

Add traces to the dataset (1 to 100 ids per call). Traces outside the organization or under another trace function are reported in "skippedTraceIds" rather than failing the call.



49
50
51
 
# File 'lib/bitfab/datasets.rb', line 49

def add_traces(dataset_id, trace_ids)
  @http_client.request(dataset_path(dataset_id, "/traces"), {"traceIds" => trace_ids})
end

#get(dataset_id) ⇒ Object

Fetch one dataset by id. A dataset outside this organization raises Net::HTTPError with a 404.



36
37
38
 
# File 'lib/bitfab/datasets.rb', line 36

def get(dataset_id)
  @http_client.get(dataset_path(dataset_id))["dataset"]
end

#get_grader_rerun(dataset_id, run_id: nil) ⇒ Object

The dataset's active grader re-run, or the run named by run_id:. Returns nil when nothing is active or the run is not this dataset's.



92
93
94
95
 
# File 'lib/bitfab/datasets.rb', line 92

def get_grader_rerun(dataset_id, run_id: nil)
  query = run_id.nil? ? "" : "?#{URI.encode_www_form("runId" => run_id)}"
  @http_client.get(dataset_path(dataset_id, "/rerunGraders#{query}"))["run"]
end

#list(trace_function_key: nil) ⇒ Object

List datasets, scoped to one trace function when given and organization-wide otherwise.



29
30
31
32
 
# File 'lib/bitfab/datasets.rb', line 29

def list(trace_function_key: nil)
  query = trace_function_key.nil? ? "" : "?#{URI.encode_www_form("traceFunctionKey" => trace_function_key)}"
  @http_client.get("/api/sdk/datasets#{query}")["datasets"]
end

#list_traces(dataset_id) ⇒ Object

The ids of every trace in the dataset, the same membership a replay with dataset_id: selects.



42
43
44
 
# File 'lib/bitfab/datasets.rb', line 42

def list_traces(dataset_id)
  @http_client.get(dataset_path(dataset_id, "/traces"))
end

#remove_graders(dataset_id, grader_ids) ⇒ Object

Unassign graders from the dataset.



66
67
68
 
# File 'lib/bitfab/datasets.rb', line 66

def remove_graders(dataset_id, grader_ids)
  @http_client.request(dataset_path(dataset_id, "/removeGraders"), {"graderIds" => grader_ids})
end

#remove_traces(dataset_id, trace_ids) ⇒ Object

Remove traces from the dataset. The traces themselves are never deleted.



54
55
56
 
# File 'lib/bitfab/datasets.rb', line 54

def remove_traces(dataset_id, trace_ids)
  @http_client.request(dataset_path(dataset_id, "/removeTraces"), {"traceIds" => trace_ids})
end

#rerun_graders(dataset_id, grader_ids: nil, wait: true, timeout: DEFAULT_RERUN_TIMEOUT_SECONDS, poll_interval: DEFAULT_RERUN_POLL_INTERVAL_SECONDS) ⇒ Object

Re-run graders over every trace in the dataset. Defaults to every assigned grader, and an unassigned id is rejected. Waits for the run to finish (up to timeout seconds) unless wait: is false, and returns the last run state seen either way. A request matching an in-flight run joins it.



75
76
77
78
79
80
81
82
83
84
85
86
87
88
 
# File 'lib/bitfab/datasets.rb', line 75

def rerun_graders(dataset_id, grader_ids: nil, wait: true,
  timeout: DEFAULT_RERUN_TIMEOUT_SECONDS, poll_interval: DEFAULT_RERUN_POLL_INTERVAL_SECONDS)
  payload = grader_ids.nil? ? {} : {"graderIds" => grader_ids}
  started = @http_client.request(dataset_path(dataset_id, "/rerunGraders"), payload)
  return started unless wait

  deadline = monotonic_now + timeout
  run = started["run"]
  while !TERMINAL_RERUN_STATUSES.include?(run["status"]) && monotonic_now < deadline
    sleep(poll_interval)
    run = get_grader_rerun(dataset_id, run_id: run["id"]) || run
  end
  {"run" => run, "joinedExisting" => started["joinedExisting"]}
end

#save(trace_function_key:, name:, description: nil) ⇒ Object

Create a dataset, or update the one already named this way under the same trace function. The result's "created" reports which happened. A nil description leaves an existing description untouched.



21
22
23
24
25
 
# File 'lib/bitfab/datasets.rb', line 21

def save(trace_function_key:, name:, description: nil)
  payload = {"traceFunctionKey" => trace_function_key, "name" => name}
  payload["description"] = description unless description.nil?
  @http_client.request("/api/sdk/datasets", payload)
end