progressor 0.0.1 → 0.1.2

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: fa7eacefe70e6b186f9a72625d6158db19b7f9b45b327c397b7b4402dc56bc47
-  data.tar.gz: 5cb7a42fffe3d6650a48454e41c73d8f74e7cb72fdf521fb2fdd6e59bb2cae8c
+  metadata.gz: 54e39588afa295d9501f32e4a183834d030b536cb902fbc133be2e61d2b97745
+  data.tar.gz: 5732a7f82cd05f0ced057785595d2d4724264ddef570c69fca086823b76318fd
 SHA512:
-  metadata.gz: 2d11d38113127328cb2d0cd6b6d98652f9e6010816aebdbbd0c44cc27b331ba03c9b961308f2ec71e6abb5f0b0694f4952132214a126cdbecb5d062741e7cbe0
-  data.tar.gz: cba7008ee6ca3151ba3c89159dc7835d8d76b5cb981da4017d3dffe0fdb14a9bf0c3ef9e18a6f532ab220f4abececc487a8145e86e34b77b226b9a81f9473f13
+  metadata.gz: e9339fc8632bdd3b0d1c0ab47d5c95d987e6db3f5383f59541511341735c105ebce68a6544eed20a0614f8ee12974d91f9fc65dfe58f4d5e662d96a02746092a
+  data.tar.gz: 5b58b596f4596bccccb352ec2c8ec6949573ad898677b228fcd43073e8b141d561c397b428d5f16dfe53dd7134e2fdb430f2be51c1a0013a1537e95f98d724ae

data/README.md

@@ -1,8 +1,34 @@
-A very basic library to measure loops in a long-running task.
+[![Build Status](https://travis-ci.org/AndrewRadev/progressor.svg?branch=master)](https://travis-ci.org/AndrewRadev/progressor)
+[![Gem Version](https://badge.fury.io/rb/progressor.svg)](https://badge.fury.io/rb/progressor)
 
-*Note: Very incomplete, so mostly for personal usage. Will hopefully flesh it out, write tests, configuration, etc, at some point (PRs welcome). Until then, a similar library can be found here: https://github.com/mkdynamic/ke*
+Full documentation for the latest released version can be found at: https://www.rubydoc.info/gems/progressor
 
-Example usage:
+## Basic example
+
+Here's an example long-running task:
+
+``` ruby
+Product.find_each do |product|
+  next if product.not_something_we_want_to_process?
+  product.calculate_interesting_stats
+end
+```
+
+In order to understand how it's progressing, we might add some print statements:
+
+``` ruby
+Product.find_each do |product|
+  if product.not_something_we_want_to_process?
+    puts "Skipping product: #{product.id}"
+    next
+  end
+
+  puts "Working on product: #{product.id}"
+  product.calculate_interesting_stats
+end
+```
+
+This gives us some indication of progress, but no idea how much time is left. We could take a count and maintain a manual index, and then eyeball it based on how fast the numbers are adding up. Progressor automates that process:
 
 ``` ruby
 progressor = Progressor.new(total_count: Product.count)
@@ -20,12 +46,79 @@ Product.find_each do |product|
 end
 ```
 
-Example output:
+Each invocation of `run` measures how long its block took and records it. The yielded `progress` parameter is an object that can be `to_s`-ed to provide progress information.
+
+The output might look like this:
+
+```
+...
+[0038/1000, (004%), t/i: 0.5s, ETA: 8m:00s] Product 38
+[0039/1000, (004%), t/i: 0.5s, ETA: 7m:58s] Product 39
+[0040/1000, (004%), t/i: 0.5s, ETA: 7m:57s] Product 40
+...
+```
+
+You can check the documentation for the [Progressor](https://www.rubydoc.info/gems/progressor/Progressor) class for details on the methods you can call to get the individual pieces of data shown in the report.
+
+## Limited and unlimited sequences
+
+Initializing a `Progressor` with a provided `total_count:` parameter gives you a limited sequence, which can give you not only a progress report, but an estimation of when it'll be done:
+
+```
+[<current loop>/<total count>, (<progress>%), t/i: <time per iteration>, ETA: <time until it's done>]
+```
+
+The calculation is done by maintaining a list of measurements with a limited size, and a list of averages of those measurements. The average of averages is the "time per iteration" and it's multiplied by the remaining count to produce the estimation.
+
+I can't really say how reliable this is, but it seems to provide smoothly changing estimations that seem more or less correct to me, for similarly-sized chunks of work per iteration.
+
+**Not** providing a `total_count:` parameter leads to less available information:
+
+``` ruby
+progressor = Progressor.new
+
+(1..100).each do |i|
+  progressor.run do |progress|
+    sleep rand
+    puts progress
+  end
+end
+```
+
+A sample of output might look like this:
 
 ```
 ...
-[0038/1000, (004%), t/i: 0.5s, ETA: 8m:0.27s] Product 38
-[0039/1000, (004%), t/i: 0.5s, ETA: 7m:58.47s] Product 39
-[0040/1000, (004%), t/i: 0.5s, ETA: 7m:57.08s] Product 40
+11, t: 5.32s, t/i: 442.39ms
+12, t: 5.58s, t/i: 446.11ms
 ...
 ```
+
+The format is:
+
+```
+<current>, t: <time from start>, t/i: <time per iteration>
+```
+
+## Configuration
+
+Apart from `total_count`, which is optional and affects the kind of sequence that will be stored, you can provide `min_samples` and `max_samples`. You can also provide a custom formatter:
+
+``` ruby
+progressor = Progressor.new({
+  total_count: 1000,
+  min_samples: 5,
+  max_samples: 10,
+  formatter: -> (p) { p.eta }
+})
+```
+
+The option `min_samples` determines how many loops the tool will wait until trying to produce an estimation. A higher number means no information in the beginning, but no wild fluctuations, either. It needs to be at least 1 and the default is 1.
+
+The option `max_samples` is how many measurements will be retained. Those measurements will be averaged, and then those averages averaged to get a time-per-iteration estimate. A smaller number means giving more weight to later events, while a larger one would average over a larger amount of samples. The default is 100.
+
+The `formatter` is a callback that gets a progress object as an argument and you can return your own string to output on every loop. Check `LimitedSequence` and `UnlimitedSequence` for the available methods and accessors you can use.
+
+## Related work
+
+A very similar tool is the gem [ke](https://github.com/mkdynamic/ke). It provides its estimation by maintaining the median quartile range of the stored measurements, removing outliers. It also automates the output of the progress report, only printing it every N loops. Depending on your needs and preferences, it might be better for your use case.

data/lib/progressor/error.rb

@@ -0,0 +1,7 @@
+class Progressor
+  # A custom error class for targeted catching. All Progressor errors will be
+  # wrapped in a Progressor::Error.
+  #
+  class Error < RuntimeError
+  end
+end

data/lib/progressor/formatting.rb

@@ -0,0 +1,39 @@
+class Progressor
+  module Formatting
+    # Formats the given time in seconds to something human readable. Examples:
+    #
+    # - 1 second:      1.00s
+    # - 0.123 seconds: 123.00ms
+    # - 100 seconds:   01m:40s
+    # - 101.5 seconds: 01m:41s
+    # - 3661 seconds:  01h:01m:01s
+    def format_time(time)
+      return "?s" if time.nil?
+
+      if time < 1
+        "#{format_float((time * 1000).round(2))}ms"
+      elsif time < 60
+        "#{format_float(time.round(2))}s"
+      elsif time < 3600
+        minutes = time.to_i / 60
+        seconds = (time - minutes * 60).round(2)
+        "#{format_int(minutes)}m:#{format_int(seconds)}s"
+      else
+        hours = time.to_i / 3600
+        minutes = (time.to_i % 3600) / 60
+        seconds = (time - (hours * 3600 + minutes * 60)).round(2)
+        "#{format_int(hours)}h:#{format_int(minutes)}m:#{format_int(seconds)}s"
+      end
+    end
+
+    # :nodoc:
+    def format_int(value)
+      sprintf("%02d", value)
+    end
+
+    # :nodoc:
+    def format_float(value)
+      sprintf("%0.2f", value)
+    end
+  end
+end

data/lib/progressor/limited_sequence.rb

@@ -0,0 +1,134 @@
+class Progressor
+  class LimitedSequence
+    include Formatting
+
+    attr_reader :total_count, :min_samples, :max_samples
+
+    # The current loop index, starts at 1
+    attr_reader :current
+
+    # The time the object was created
+    attr_reader :start_time
+
+    # Creates a new LimitedSequence with the given parameters:
+    #
+    # - total_count: The expected number of loops.
+    #
+    # - min_samples: The number of samples to collect before attempting to
+    #   calculate a time per iteration. Default: 1
+    #
+    # - max_samples: The maximum number of measurements to collect and average.
+    #   Default: 100.
+    #
+    # - formatter: A callable that accepts the sequence object and returns a
+    #   custom formatted string.
+    #
+    def initialize(total_count:, min_samples: 1, max_samples: 100, formatter: nil)
+      @total_count = total_count
+      @min_samples = min_samples
+      @max_samples = [max_samples, total_count].min
+      @formatter   = formatter
+
+      raise Error.new("min_samples needs to be a positive number") if min_samples <= 0
+      raise Error.new("max_samples needs to be larger than min_samples") if max_samples <= min_samples
+
+      @start_time         = Time.now
+      @total_count_digits = total_count.to_s.length
+      @current            = 0
+      @measurements       = []
+      @averages           = []
+    end
+
+    # Adds a duration in seconds to the internal storage of samples. Updates
+    # averages accordingly.
+    #
+    def push(duration)
+      @current += 1
+      @measurements << duration
+      # only keep last `max_samples`
+      @measurements.shift if @measurements.count > max_samples
+
+      @averages << average(@measurements)
+      @averages = @averages.compact
+      # only keep last `max_samples`
+      @averages.shift if @averages.count > max_samples
+    end
+
+    # Skips an iteration, updating the total count and ETA
+    #
+    def skip(n)
+      @total_count -= n
+    end
+
+    # Outputs a textual representation of the current state of the
+    # LimitedSequence. Shows:
+    #
+    # - the current number of iterations and the total count
+    # - completion level in percentage
+    # - how long a single iteration takes
+    # - estimated time of arrival (ETA) -- time until it's done
+    #
+    # A custom `formatter` provided at construction time overrides this default
+    # output.
+    #
+    # If the "current" number of iterations goes over the total count, an ETA
+    # can't be shown anymore, so it'll just be the current number over the
+    # expected one, and the time per iteration.
+    #
+    def to_s
+      return @formatter.call(self).to_s if @formatter
+
+      if @current > @total_count
+        return [
+          "#{@current} (expected #{@total_count})",
+          "t/i: #{format_time(per_iteration)}",
+          "ETA: ???",
+        ].join(', ')
+      end
+
+      [
+        "#{@current.to_s.rjust(@total_count_digits, '0')}/#{@total_count}",
+        "#{((@current / @total_count.to_f) * 100).round.to_s.rjust(3, '0')}%",
+        "t/i: #{format_time(per_iteration)}",
+        "ETA: #{format_time(eta)}",
+      ].join(', ')
+    end
+
+    # Returns an estimation for the time per single iteration. Implemented as
+    # an average of averages to provide a smoother gradient from loop to loop.
+    #
+    # Returns nil if not enough samples have been collected yet.
+    #
+    def per_iteration
+      return nil if @measurements.count < min_samples
+      average(@averages)
+    end
+
+    # Returns an estimation for the Estimated Time of Arrival (time until
+    # done).
+    #
+    # Calculated by multiplying the average time per iteration with the
+    # remaining number of loops.
+    #
+    def eta
+      return nil if @measurements.count < min_samples
+
+      remaining_time = per_iteration * (@total_count - @current)
+      remaining_time.round(2)
+    end
+
+    # Returns the time since the object was instantiated, formatted like all
+    # the other durations. Useful for a final message to compare initial
+    # estimation to actual elapsed time.
+    #
+    def elapsed_time
+      format_time(Time.now - @start_time)
+    end
+
+    private
+
+    def average(collection)
+      collection.inject(&:+) / collection.count.to_f
+    end
+  end
+end

data/lib/progressor/unlimited_sequence.rb

@@ -0,0 +1,111 @@
+class Progressor
+  class UnlimitedSequence
+    include Formatting
+
+    attr_reader :min_samples, :max_samples
+
+    # The current loop index, starts at 1
+    attr_reader :current
+
+    # The time the object was created
+    attr_reader :start_time
+
+    # Creates a new UnlimitedSequence with the given parameters:
+    #
+    # - min_samples: The number of samples to collect before attempting to
+    #   calculate a time per iteration. Default: 1
+    #
+    # - max_samples: The maximum number of measurements to collect and average.
+    #   Default: 100.
+    #
+    # - formatter: A callable that accepts the sequence object and returns a
+    #   custom formatted string.
+    #
+    def initialize(min_samples: 1, max_samples: 100, formatter: nil)
+      @min_samples = min_samples
+      @max_samples = max_samples
+      @formatter   = formatter
+
+      raise Error.new("min_samples needs to be a positive number") if min_samples <= 0
+      raise Error.new("max_samples needs to be larger than min_samples") if max_samples <= min_samples
+
+      @start_time   = Time.now
+      @current      = 0
+      @measurements = []
+      @averages     = []
+    end
+
+    # Adds a duration in seconds to the internal storage of samples. Updates
+    # averages accordingly.
+    #
+    def push(duration)
+      @current += 1
+      @measurements << duration
+      # only keep last `max_samples`
+      @measurements.shift if @measurements.count > max_samples
+
+      @averages << average(@measurements)
+      @averages = @averages.compact
+      # only keep last `max_samples`
+      @averages.shift if @averages.count > max_samples
+    end
+
+    # "Skips" an iteration, which, in the context of an UnlimitedSequence is a no-op.
+    #
+    def skip(_n)
+      # Nothing to do
+    end
+
+    # Outputs a textual representation of the current state of the
+    # UnlimitedSequence. Shows:
+    #
+    # - the current (1-indexed) number of iterations
+    # - how long since the start time
+    # - how long a single iteration takes
+    #
+    # A custom `formatter` provided at construction time overrides this default
+    # output.
+    #
+    def to_s
+      return @formatter.call(self).to_s if @formatter
+
+      [
+        "#{@current + 1}",
+        "t: #{elapsed_time}",
+        "t/i: #{format_time(per_iteration)}",
+      ].join(', ')
+    end
+
+    # Returns an estimation for the time per single iteration. Implemented as
+    # an average of averages to provide a smoother gradient from loop to loop.
+    #
+    # Returns nil if not enough samples have been collected yet.
+    #
+    def per_iteration
+      return nil if @measurements.count < min_samples
+      average(@averages)
+    end
+
+    # Is supposed to return an estimation for the Estimated Time of Arrival
+    # (time until done).
+    #
+    # For an UnlimitedSequence, this always returns nil.
+    #
+    def eta
+      # No estimation possible
+    end
+
+    # Returns the time since the object was instantiated, formatted like all
+    # the other durations.
+    #
+    def elapsed_time
+      format_time(Time.now - @start_time)
+    end
+
+    private
+
+    def average(collection)
+      collection.inject(&:+) / collection.count.to_f
+    end
+  end
+end

data/lib/progressor/version.rb

@@ -0,0 +1,3 @@
+class Progressor
+  VERSION = '0.1.2'
+end

data/lib/progressor.rb

@@ -1,7 +1,13 @@
+require 'progressor/version'
+require 'progressor/error'
+require 'progressor/formatting'
+require 'progressor/limited_sequence'
+require 'progressor/unlimited_sequence'
+
 require 'benchmark'
 
 # Used to measure the running time of parts of a long-running task and output
-# an estimation based on the average of the last 10-100 measurements.
+# an estimation based on the average of the last 1-100 measurements.
 #
 # Example usage:
 #
@@ -22,98 +28,75 @@ require 'benchmark'
 # Example output:
 #
 #   ...
-#   [0038/1000, (004%), t/i: 0.5s, ETA: 8m:0.27s] Product 38
-#   [0039/1000, (004%), t/i: 0.5s, ETA: 7m:58.47s] Product 39
-#   [0040/1000, (004%), t/i: 0.5s, ETA: 7m:57.08s] Product 40
+#   [0038/1000, 004%, t/i: 0.5s, ETA: 8m:00s] Product 38
+#   [0039/1000, 004%, t/i: 0.5s, ETA: 7m:58s] Product 39
+#   [0040/1000, 004%, t/i: 0.5s, ETA: 7m:57s] Product 40
 #   ...
 #
 class Progressor
-  VERSION = '0.0.1'
+  extend Formatting
 
   # Utility method to print a message with the time it took to run the contents
   # of the block.
   #
-  # > Progressor.puts("Working on a thing") { thing_work }
+  #   Progressor.puts("Working on a thing") { thing_work }
+  #
+  # Output:
   #
-  # Working on a thing...
-  # Working on a thing DONE: 2.1s
+  #   Working on a thing...
+  #   Working on a thing DONE: 2.1s
   #
   def self.puts(message, &block)
     Kernel.puts "#{message}..."
-    measurement = Benchmark.measure { block.call }
+    result = nil
+    measurement = Benchmark.measure { result = block.call }
     Kernel.puts "#{message} DONE: #{format_time(measurement.real)}"
+    result
   end
 
-  def initialize(total_count:)
-    @total_count = total_count
-    @total_count_digits = total_count.to_s.length
-    @current = 0
-    @measurements = []
-    @averages = []
+  # Set up a new Progressor instance. Optional parameters:
+  #
+  # - total_count: If given, the tool will be able to provide an ETA.
+  #
+  # - min_samples: The number of samples to collect before attempting to
+  #   calculate a time per iteration. Default: 1
+  #
+  # - max_samples: The maximum number of measurements to collect and average.
+  #   Default: 100.
+  #
+  # - formatter: A callable that accepts a progress object and returns a
+  #   custom formatted string.
+  #
+  def initialize(total_count: nil, min_samples: 1, max_samples: 100, formatter: nil)
+    params = {
+      min_samples: min_samples,
+      max_samples: max_samples,
+      formatter:   formatter,
+    }
+
+    if total_count
+      @sequence = LimitedSequence.new(total_count: total_count, **params)
+    else
+      @sequence = UnlimitedSequence.new(**params)
+    end
   end
 
+  # Run the given block of code, yielding a sequence object that holds progress
+  # information.
+  #
+  # Example usage:
+  #
+  #   progressor.run { |progress| puts progress; long_running_task() }
+  #
   def run
-    @current += 1
-
-    measurement = Benchmark.measure { yield self }
-
-    @measurements << measurement.real
-    # only keep last 1000
-    @measurements.shift if @measurements.count > 1000
-
-    @averages << average(@measurements)
-    @averages = @averages.compact
-    # only keep last 100
-    @averages.shift if @averages.count > 100
+    measurement = Benchmark.measure { yield @sequence }
+    @sequence.push(measurement.real)
   end
 
+  # Skips the given number of loops (will likely be 1), updating the
+  # estimations appropriately.
+  #
   def skip(n)
-    @total_count -= n
-  end
-
-  def to_s
-    [
-      "#{@current.to_s.rjust(@total_count_digits, '0')}/#{@total_count}",
-      "(#{((@current / @total_count.to_f) * 100).round.to_s.rjust(3, '0')}%)",
-      "t/i: #{self.class.format_time(per_iteration)}",
-      "ETA: #{self.class.format_time(eta)}",
-    ].join(', ')
-  end
-
-  def per_iteration
-    return nil if @measurements.count < 10
-    average(@averages)
-  end
-
-  def eta
-    return nil if @measurements.count < 10
-
-    remaining_time = per_iteration * (@total_count - @current)
-    remaining_time.round(2)
-  end
-
-  private
-
-  def self.format_time(time)
-    return "?s" if time.nil?
-
-    if time < 0.1
-      "#{(time * 1000).round(2)}ms"
-    elsif time < 60
-      "#{time.round(2)}s"
-    elsif time < 3600
-      minutes = time.to_i / 60
-      seconds = (time - minutes * 60).round(2)
-      "#{minutes}m:#{seconds}s"
-    else
-      hours = time.to_i / 3600
-      minutes = (time.to_i % 3600) / 60
-      seconds = (time - (hours * 3600 + minutes * 60)).round(2)
-      "#{hours}h:#{minutes}m:#{seconds}s"
-    end
-  end
-
-  def average(collection)
-    collection.inject(&:+) / collection.count.to_f
+    @sequence.skip(n)
   end
 end

metadata

@@ -1,57 +1,57 @@
 --- !ruby/object:Gem::Specification
 name: progressor
 version: !ruby/object:Gem::Version
-  version: 0.0.1
+  version: 0.1.2
 platform: ruby
 authors:
 - Andrew Radev
-autorequire: 
+autorequire:
 bindir: bin
 cert_chain: []
-date: 2019-02-20 00:00:00.000000000 Z
+date: 2022-04-03 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
-  name: bundler
+  name: rake
   requirement: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '1.17'
+        version: 12.3.3
   type: :development
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '1.17'
+        version: 12.3.3
 - !ruby/object:Gem::Dependency
-  name: rake
+  name: rspec
   requirement: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '10.0'
+        version: '3.0'
   type: :development
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '10.0'
+        version: '3.0'
 - !ruby/object:Gem::Dependency
-  name: rspec
+  name: timecop
   requirement: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '3.0'
+        version: '0.9'
   type: :development
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '3.0'
+        version: '0.9'
 description: |
   Provides a way to measure how long each loop in a task took, outputting a
   report with an estimated time till the task is done.
@@ -65,11 +65,16 @@ files:
 - LICENSE
 - README.md
 - lib/progressor.rb
+- lib/progressor/error.rb
+- lib/progressor/formatting.rb
+- lib/progressor/limited_sequence.rb
+- lib/progressor/unlimited_sequence.rb
+- lib/progressor/version.rb
 homepage: https://github.com/AndrewRadev/progressor
 licenses:
 - MIT
 metadata: {}
-post_install_message: 
+post_install_message:
 rdoc_options: []
 require_paths:
 - lib
@@ -84,8 +89,8 @@ required_rubygems_version: !ruby/object:Gem::Requirement
     - !ruby/object:Gem::Version
       version: '0'
 requirements: []
-rubygems_version: 3.0.0
-signing_key: 
+rubygems_version: 3.1.6
+signing_key:
 specification_version: 4
 summary: Measure iterations in a long-running task
 test_files: []