arquivo 0.2.0 → 0.2.1

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 5327acf76aa8ca135c8e6faac8aeb2b5aee4a66a950cf07f3a7498f98da503f5
-  data.tar.gz: d37de979686bfef7182714d1b654da47dae151b3b68b52ebc0b7cf79e25cfa4b
+  metadata.gz: db2e214ab2f4424ed534529d6601643e05256b201c75ac4d6ce13ce63167f62e
+  data.tar.gz: 657d0d3b58f42150e1930d669444ab71a019df9e2b006abc2ac75caee8e4a1ee
 SHA512:
-  metadata.gz: 2d82bbfe746b9dfb2337abcb2c8ef54cb0adfe9466c2b51cde53dfb54eaf1f1cb25828216531350f5c1c5f7e2cf7664f0b80b223b00b17852681826ef714c3fb
-  data.tar.gz: 57dd9546e5d54eb0ad94da2b2fca720e0ac2511e671c1d39f169ff8b03773452f705e25e460ac884d7ac6fb5c4ebd1c558dac341da9d56cfff1a0233ba8560e8
+  metadata.gz: 7aa8742696b7e53e4af0ff69c8c85b71fa48969751d96f162c8546d276373796c2c0dba1a075d5d00fec89ed24bda847083e1061d7698d4919e1b0815211c91c
+  data.tar.gz: 074e49e12a77c243ff62b32bde04b6d72e84576da5f152db4e3c85b643b245cca8f59058a010ea6318c92aed10f2776d339fb5a047e52ab8614f560aa19d00d4

data/Gemfile.lock

@@ -1,7 +1,7 @@
 PATH
   remote: .
   specs:
-    arquivo (0.2.0)
+    arquivo (0.2.1)
       fastimage (~> 2.1)
       google-api-client (~> 0.34)
       pdf-reader (~> 2.3)

data/lib/arquivo.rb

@@ -54,7 +54,7 @@ module Arquivo
                    desc: 'ruido de fundo - sim ou nao'
     option :sound, type: :numeric, default: 1.0,
                    desc: 'minimo som que determina fim do silencio (segundos)'
-    option :amount, type: :numeric, default: 0.0001,
+    option :amount, type: :numeric, default: 0.00001,
                     desc: 'qtd ruido a ser removido'
     option :rate, type: :numeric, default: 16,
                   desc: 'sample rate - radio-16k, CD-44.1k, PC-48k, pro-96k'

data/lib/arquivo/dir.rb

@@ -9,10 +9,11 @@ require 'arquivo/noise'
 module Arquivo
   O1 = '2>/dev/null'
   O2 = '1>/dev/null 2>&1'
+  FT = ['.mp3', '.m4a', '.wav', '.sox'].freeze
 
   # analisar/processar pasta
   class C118dir < Enumerator
-    # @return [Enumerator] lista items dentro duma pasta
+    # @return [Enumerator] items dentro duma pasta
     attr_reader :items
     # @return [String] documento c118
     attr_reader :item
@@ -66,10 +67,10 @@ module Arquivo
       case ext
       when '.jpg' then C118jpg.new(item).processa_jpg(options, dados)
       when '.pdf' then C118pdf.new(item).processa_pdf(options, dados)
-      when '.mp3', '.m4a', '.wav'
+      when *FT
         C118mp3.new(item).processa_mp3(options, noiseprof)
       else
-        puts "erro: #{item} so posso processar mp3, jpg, pdf"
+        puts "erro: #{item} so posso processar"
       end
     end
 

data/lib/arquivo/extrato.rb

@@ -19,14 +19,15 @@ module Arquivo
     end
 
     def processa_extrato?
-      return true if !File.exist?(base) &&
-                     File.exist?(file) && ext == '.pdf' &&
-                     first_extrato
+      return true if ext == '.pdf' &&
+                     size.positive? &&
+                     !File.exist?(base) &&
+                     first_extrato?
 
       if File.exist?(base)
         puts "erro: #{base} pasta ja existe"
       else
-        puts "erro: #{file} nao consigo obter primeira pagina do PDF"
+        puts "erro: #{file} nao consigo obter primeira pagina do EXTRATO"
       end
       false
     end
@@ -58,6 +59,10 @@ module Arquivo
       pagina.include?('45463760224')
     end
 
+    def first_extrato?
+      leitor && proxima_pagina && proximo_extrato
+    end
+
     # @return [Enumerator::Lazy] leitor pdf
     def leitor
       @leitor ||= PDF::Reader.new(file).pages.lazy
@@ -82,10 +87,6 @@ module Arquivo
       @nome = nil
     end
 
-    def first_extrato
-      leitor && proxima_pagina && proximo_extrato
-    end
-
     def split
       system "pdftk #{file} burst output #{base}/pg%04d-#{base}.pdf;" \
              "rm -f #{base}/*.txt"

data/lib/arquivo/mp3.rb

@@ -25,13 +25,12 @@ module Arquivo
 
     def processa_mp3(options, npr)
       cmd = if npr
-              "noisered #{npr} #{format('%<v>.5f', v: options[:amount])} "
+              "noisered #{npr} #{format('%<v>.9f', v: options[:amount])} "
             else
               ''
             end
-      cmd += "rate -v #{options[:rate]}k"
+      cmd += "rate -v #{options[:rate]}k channels 1"
       system "sox -G #{file} tmp/zip/#{base}.mp3 #{cmd} #{O2}"
-      # puts base
     end
 
     def segmenta(tps, pse, cmd)
@@ -57,14 +56,14 @@ module Arquivo
     end
 
     def processa_minuta?
-      return true if ['.mp3', '.m4a', '.wav'].include?(ext) &&
-                     size.positive? &&
+      return true if FT.include?(ext) && size.positive? &&
                      !File.exist?(base)
 
       if File.exist?(base)
         puts "erro: #{base} pasta ja existe"
       else
-        puts "erro: #{file} nao consigo processar minuta"
+        puts 'erro: so consigo processar minutas com som ' \
+             "e tipo #{FT}"
       end
 
       false

data/lib/arquivo/noise.rb

@@ -3,49 +3,52 @@
 module Arquivo
   # analisar/processar pasta
   class C118dir < Enumerator
-    def obtem_noiseprof(dir, options)
-      return unless /minuta/i.match?(dir) && !options[:noise]
+    def obtem_noiseprof(pasta, options)
+      return unless /minuta/i.match?(pasta) && !options[:noise]
 
-      silencio(1, duracao(item), options[:sound]) while next_item
-      items.rewind
-      @noiseprof = processa_noiseprof
-    end
+      d = Dir.glob(File.join(pasta, '*')).map { |e| [e, duracao(e)] }
+      t = 0
+      s = ['', 0, 0]
+      s = maximo(d, t += 1, options[:sound]) while t < 9 && s[2] <= silence
 
-    def silencio(thr, tse, som)
-      o = "tmp/silencio-#{File.basename(item)}"
+      processa_silencio(s)
+    end
 
-      system "sox #{item} #{o} " \
-             "silence 1 #{format('%<valor>.5f', valor: som)}t #{thr}% #{O2}"
+    def processa_silencio(seg)
+      return unless seg[2] > silence
 
-      return if silencio?(o, tse) || thr == 3
+      o = "tmp/silencio-#{File.basename(seg[0])}"
+      system "sox #{seg[0]} #{o} trim 0 #{seg[2]} #{O2}"
+      seg[2] = duracao(o)
+      return unless seg[2].positive?
 
-      silencio(thr + 1, tse, som)
+      processa_noiseprof(seg, o)
     end
 
-    def silencio?(fss, tse)
-      s = duracao(fss)
-      return false unless s.positive? && (tse - s > silence)
+    def processa_noiseprof(seg, trm)
+      o = "tmp/noiseprof-#{File.basename(seg[0], File.extname(seg[0]))}"
+      # obter noiseprof do silencio encontrado
+      system "sox #{trm} -n noiseprof #{o} #{O2}"
 
-      @silence = tse - s
-      @noiseprof = fss
+      # so noiseprof validos sao devolvidos
+      @silence = File.size?(o) ? seg[2] : 0.0
+      @noiseprof = silence.positive? ? o : nil
     end
 
-    def duracao(seg)
-      `soxi -V0 -D #{seg} #{O1}`.to_f
+    def maximo(seg, thr, som)
+      seg.sort.map { |e| add_silencio(e, thr, som) }.max_by { |_, _, s| s }
     end
 
-    def processa_noiseprof
-      return unless silence&.positive?
+    def add_silencio(seg, thr, som)
+      o = "tmp/thr-#{File.basename(seg[0])}"
+      system "sox #{seg[0]} #{o} silence 1 #{som}t #{thr}% #{O2}"
+      s = (seg[1] - duracao(o)).round(2, half: :down)
 
-      e = File.extname(noiseprof)
-      o = "tmp/noiseprof-#{File.basename(noiseprof, e)}"
-      # obter noiseprof do silencio encontrado
-      system "sox #{noiseprof} #{o}#{e} trim 0 #{silence} #{O2};" \
-             "sox #{o}#{e} -n noiseprof #{o} #{O2}"
+      seg + [s > som ? s : 0.0]
+    end
 
-      # so noiseprof validos sao devolvidos
-      @silence = 0.0 unless File.size?(o)
-      silence.positive? ? o : nil
+    def duracao(seg)
+      `soxi -V0 -D #{seg} #{O1}`.to_f
     end
   end
 end

data/lib/arquivo/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module Arquivo
-  VERSION = '0.2.0'
+  VERSION = '0.2.1'
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: arquivo
 version: !ruby/object:Gem::Version
-  version: 0.2.0
+  version: 0.2.1
 platform: ruby
 authors:
 - Hernâni Rodrigues Vaz
 autorequire: 
 bindir: exe
 cert_chain: []
-date: 2019-12-20 00:00:00.000000000 Z
+date: 2019-12-23 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: bundler