reckon 0.4.4 → 0.5.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
-SHA1:
-  metadata.gz: cf1d814fe911f299cf58401c3b36734e29a172f1
-  data.tar.gz: 9130495690de1f73cfa3570a6e3386f38df86830
+SHA256:
+  metadata.gz: b58a745c730c0dfaf022a98c19dfaf8568c0b3091548d70c9202f8f7c0d78ada
+  data.tar.gz: 030d2036524a19c5eff981608133e8aac15f185b57a891e91a6f1fdabd3796c7
 SHA512:
-  metadata.gz: b7c3e23f96983cd79844ba37524f1b843a1dbba230198cf813b83c7c65a076506fb2c66ab663dee0970db2d42e1d3e0b03adf41ee163d08f3b38dbb163c373a6
-  data.tar.gz: d74ab72db15d54bc0bfe548369ec9a8b3903760bc09fbc16340a6f64c929a69d595c4840900e8a425996bc580e5924321ece84b10d3e033b193b066f2520f7c2
+  metadata.gz: 1daf76a19078f1707e847a7ac54aad5836ee3d49ce9c2f612bfefeb86f8c1bf9c9d1f97b7bbecb309f9f44f8097f8f2dbe0d6b6062bc376a62fc88f45bab1dfc
+  data.tar.gz: 4784ac871dbdb6160dc53ef8a7f5d69f31ccc7f7f7ea876b97bf60941abb2d13a968f5229747dd74d0b574ecc12326485e1eaba7ea368a91162d36d3ef4f9f86

data/.ruby-version

@@ -1 +1 @@
-ruby-2.2.0
+2.5

data/.travis.yml

@@ -1,5 +1,13 @@
 language: ruby
 rvm:
-  - 1.9.3
-  - 2.0.0
+  # Mac High Sierra
+  - 2.0.0-p648
+  # Mac Mojave
+  - 2.3.7
+  # Ubuntu 19.10
+  - 2.5
+  # Mac Catalina
+  - 2.6
 script: "bundle exec rake"
+before_install:
+  - sudo apt-get -y install ledger

data/CHANGELOG.md

@@ -0,0 +1,197 @@
+# Changelog
+
+## [v0.5.0](https://github.com/cantino/reckon/tree/v0.5.0) (2020-02-19)
+
+[Full Changelog](https://github.com/cantino/reckon/compare/v0.4.4...v0.5.0)
+
+**Closed issues:**
+
+- g [\#75](https://github.com/cantino/reckon/issues/75)
+- Learn-from not working [\#74](https://github.com/cantino/reckon/issues/74)
+- Tokens YAML fails to match [\#73](https://github.com/cantino/reckon/issues/73)
+- Missing or stray quote in line error [\#71](https://github.com/cantino/reckon/issues/71)
+- Support ISO 8601 formatting of dates in ledger file [\#70](https://github.com/cantino/reckon/issues/70)
+- Looking for a new maintainer for Reckon [\#68](https://github.com/cantino/reckon/issues/68)
+- Reckon undefined method to\_h when trying to parse csv file  [\#66](https://github.com/cantino/reckon/issues/66)
+- Runtime error [\#65](https://github.com/cantino/reckon/issues/65)
+- Reckon doesn't learn from multiple sources [\#63](https://github.com/cantino/reckon/issues/63)
+- problem of importing file [\#59](https://github.com/cantino/reckon/issues/59)
+- Problem with file in which every column is quoted. [\#58](https://github.com/cantino/reckon/issues/58)
+- Error in reckon for the same format csv file [\#57](https://github.com/cantino/reckon/issues/57)
+- Parsing account names does not work if currency symbol is different from $ [\#56](https://github.com/cantino/reckon/issues/56)
+- Problem reading csv file [\#55](https://github.com/cantino/reckon/issues/55)
+- Problem with mint file [\#53](https://github.com/cantino/reckon/issues/53)
+- --money-column [\#43](https://github.com/cantino/reckon/issues/43)
+
+**Merged pull requests:**
+
+- Fix bugs in ledger file parsing.  Fixes \#56. [\#81](https://github.com/cantino/reckon/pull/81) ([benprew](https://github.com/benprew))
+- Better file encoding suggestions [\#80](https://github.com/cantino/reckon/pull/80) ([benprew](https://github.com/benprew))
+- :bug: fix matching algorithm, add logging and a spec helper.  Fixes \#73 [\#79](https://github.com/cantino/reckon/pull/79) ([benprew](https://github.com/benprew))
+- bug: invalid header lines should be ignored, not parsed. [\#78](https://github.com/cantino/reckon/pull/78) ([benprew](https://github.com/benprew))
+- convert default date format to iso8601 [\#77](https://github.com/cantino/reckon/pull/77) ([benprew](https://github.com/benprew))
+- Fix rspec failure for ruby 2.3 and 2.4 [\#69](https://github.com/cantino/reckon/pull/69) ([BlackEdder](https://github.com/BlackEdder))
+- Allow setting of money and date columns by index [\#67](https://github.com/cantino/reckon/pull/67) ([cantino](https://github.com/cantino))
+
+## [v0.4.4](https://github.com/cantino/reckon/tree/v0.4.4) (2015-12-02)
+
+[Full Changelog](https://github.com/cantino/reckon/compare/v0.4.3...v0.4.4)
+
+**Merged pull requests:**
+
+- Regexp support in the tokens file [\#54](https://github.com/cantino/reckon/pull/54) ([vzctl](https://github.com/vzctl))
+
+## [v0.4.3](https://github.com/cantino/reckon/tree/v0.4.3) (2015-08-16)
+
+[Full Changelog](https://github.com/cantino/reckon/compare/v0.4.2...v0.4.3)
+
+## [v0.4.2](https://github.com/cantino/reckon/tree/v0.4.2) (2015-08-08)
+
+[Full Changelog](https://github.com/cantino/reckon/compare/v0.4.1...v0.4.2)
+
+**Merged pull requests:**
+
+- Ignore empty description columns [\#52](https://github.com/cantino/reckon/pull/52) ([vzctl](https://github.com/vzctl))
+
+## [v0.4.1](https://github.com/cantino/reckon/tree/v0.4.1) (2015-07-08)
+
+[Full Changelog](https://github.com/cantino/reckon/compare/v0.4.0...v0.4.1)
+
+**Closed issues:**
+
+- Unattended [\#50](https://github.com/cantino/reckon/issues/50)
+- Debit/Credit Columns from SunTrust [\#42](https://github.com/cantino/reckon/issues/42)
+
+**Merged pull requests:**
+
+- \[RFC\] Fix \#42: Work with suntrust double column csv files [\#48](https://github.com/cantino/reckon/pull/48) ([BlackEdder](https://github.com/BlackEdder))
+
+## [v0.4.0](https://github.com/cantino/reckon/tree/v0.4.0) (2015-06-05)
+
+[Full Changelog](https://github.com/cantino/reckon/compare/v0.3.10...v0.4.0)
+
+**Implemented enhancements:**
+
+- Tab completion for transactions [\#40](https://github.com/cantino/reckon/issues/40)
+- feature: "unattended" mode [\#3](https://github.com/cantino/reckon/issues/3)
+
+**Closed issues:**
+
+- Missing or stray quote error [\#38](https://github.com/cantino/reckon/issues/38)
+
+**Merged pull requests:**
+
+- Better ISO 8601 dates support [\#49](https://github.com/cantino/reckon/pull/49) ([vzctl](https://github.com/vzctl))
+- Unattended mode and custom tokens support [\#47](https://github.com/cantino/reckon/pull/47) ([vzctl](https://github.com/vzctl))
+- \[RFC\] Implement issue \#40: Tab completion [\#46](https://github.com/cantino/reckon/pull/46) ([BlackEdder](https://github.com/BlackEdder))
+- set readline to allow for backspace in ask dialog [\#44](https://github.com/cantino/reckon/pull/44) ([mrtazz](https://github.com/mrtazz))
+
+## [v0.3.10](https://github.com/cantino/reckon/tree/v0.3.10) (2014-08-16)
+
+[Full Changelog](https://github.com/cantino/reckon/compare/v0.3.9...v0.3.10)
+
+**Merged pull requests:**
+
+- Fix --encoding option [\#41](https://github.com/cantino/reckon/pull/41) ([mamciek](https://github.com/mamciek))
+- Bumped version number [\#37](https://github.com/cantino/reckon/pull/37) ([BlackEdder](https://github.com/BlackEdder))
+
+## [v0.3.9](https://github.com/cantino/reckon/tree/v0.3.9) (2014-02-20)
+
+[Full Changelog](https://github.com/cantino/reckon/compare/v0.3.8...v0.3.9)
+
+**Closed issues:**
+
+- Idea/discussion: csv parser [\#25](https://github.com/cantino/reckon/issues/25)
+- Silently misinterprets UK dates [\#18](https://github.com/cantino/reckon/issues/18)
+
+**Merged pull requests:**
+
+- Added spec for csv files from Broker Canada [\#36](https://github.com/cantino/reckon/pull/36) ([BlackEdder](https://github.com/BlackEdder))
+- Date format [\#35](https://github.com/cantino/reckon/pull/35) ([BlackEdder](https://github.com/BlackEdder))
+- Added example from a french bank [\#34](https://github.com/cantino/reckon/pull/34) ([BlackEdder](https://github.com/BlackEdder))
+- Austrian example [\#33](https://github.com/cantino/reckon/pull/33) ([BlackEdder](https://github.com/BlackEdder))
+- Ing csv [\#30](https://github.com/cantino/reckon/pull/30) ([BlackEdder](https://github.com/BlackEdder))
+- Further improvements in nationwide csv handling [\#29](https://github.com/cantino/reckon/pull/29) ([BlackEdder](https://github.com/BlackEdder))
+- Refactor: Add money class [\#28](https://github.com/cantino/reckon/pull/28) ([BlackEdder](https://github.com/BlackEdder))
+- Initial split of CSVparser from class App [\#27](https://github.com/cantino/reckon/pull/27) ([BlackEdder](https://github.com/BlackEdder))
+- Updated version of pull request 24: Allow for other currency symbols while calculating money\_score [\#26](https://github.com/cantino/reckon/pull/26) ([BlackEdder](https://github.com/BlackEdder))
+- Change double column detection [\#23](https://github.com/cantino/reckon/pull/23) ([BlackEdder](https://github.com/BlackEdder))
+- Added optional argument to contains\_header to skip multiple header lines [\#22](https://github.com/cantino/reckon/pull/22) ([BlackEdder](https://github.com/BlackEdder))
+- Add a Bitdeli Badge to README [\#20](https://github.com/cantino/reckon/pull/20) ([bitdeli-chef](https://github.com/bitdeli-chef))
+- Update README to show latest usage info [\#19](https://github.com/cantino/reckon/pull/19) ([purcell](https://github.com/purcell))
+
+## [v0.3.8](https://github.com/cantino/reckon/tree/v0.3.8) (2013-07-03)
+
+[Full Changelog](https://github.com/cantino/reckon/compare/v0.3.7...v0.3.8)
+
+**Implemented enhancements:**
+
+- Support other currencies [\#7](https://github.com/cantino/reckon/issues/7)
+
+**Closed issues:**
+
+- Add support for dates in spanish dd/mm/yyyy [\#13](https://github.com/cantino/reckon/issues/13)
+- Problems with my csv file [\#8](https://github.com/cantino/reckon/issues/8)
+
+**Merged pull requests:**
+
+- add support for spanish dates dd/mm/yyyy closes \#13 [\#14](https://github.com/cantino/reckon/pull/14) ([mauromorales](https://github.com/mauromorales))
+- fix issue showing true when parsing the currency option related to \#7 [\#12](https://github.com/cantino/reckon/pull/12) ([mauromorales](https://github.com/mauromorales))
+
+## [v0.3.7](https://github.com/cantino/reckon/tree/v0.3.7) (2013-06-27)
+
+[Full Changelog](https://github.com/cantino/reckon/compare/v0.3.6...v0.3.7)
+
+**Merged pull requests:**
+
+- Updated the sources to allow for custom curreny [\#11](https://github.com/cantino/reckon/pull/11) ([ghost](https://github.com/ghost))
+- Add --account option on the commandline [\#10](https://github.com/cantino/reckon/pull/10) ([copiousfreetime](https://github.com/copiousfreetime))
+
+## [v0.3.6](https://github.com/cantino/reckon/tree/v0.3.6) (2013-04-30)
+
+[Full Changelog](https://github.com/cantino/reckon/compare/v0.3.5...v0.3.6)
+
+**Closed issues:**
+
+- iso-8859-1 CSV with accented chars =\> invalid byte sequence in UTF-8 \(ArgumentError\) [\#5](https://github.com/cantino/reckon/issues/5)
+- Ruby 2.0 compatibility [\#4](https://github.com/cantino/reckon/issues/4)
+
+**Merged pull requests:**
+
+- Recognize yyyymmdd date in Reckon::App\#date\_for. [\#9](https://github.com/cantino/reckon/pull/9) ([mhoogendoorn](https://github.com/mhoogendoorn))
+
+## [v0.3.5](https://github.com/cantino/reckon/tree/v0.3.5) (2013-03-24)
+
+[Full Changelog](https://github.com/cantino/reckon/compare/v0.3.4...v0.3.5)
+
+**Closed issues:**
+
+- backtrace trying to run reckon -f [\#2](https://github.com/cantino/reckon/issues/2)
+
+**Merged pull requests:**
+
+- Inverse mode [\#6](https://github.com/cantino/reckon/pull/6) ([nathankot](https://github.com/nathankot))
+
+## [v0.3.4](https://github.com/cantino/reckon/tree/v0.3.4) (2013-02-16)
+
+[Full Changelog](https://github.com/cantino/reckon/compare/v0.3.3...v0.3.4)
+
+**Merged pull requests:**
+
+- adds support for Nordea csv files [\#1](https://github.com/cantino/reckon/pull/1) ([x2q](https://github.com/x2q))
+
+## [v0.3.3](https://github.com/cantino/reckon/tree/v0.3.3) (2013-01-13)
+
+[Full Changelog](https://github.com/cantino/reckon/compare/v0.3.1...v0.3.3)
+
+## [v0.3.1](https://github.com/cantino/reckon/tree/v0.3.1) (2012-07-30)
+
+[Full Changelog](https://github.com/cantino/reckon/compare/v0.3.2...v0.3.1)
+
+## [v0.3.2](https://github.com/cantino/reckon/tree/v0.3.2) (2012-07-30)
+
+[Full Changelog](https://github.com/cantino/reckon/compare/5c07bea3fe63f9b909b4b76bd49f22fd8faf7a29...v0.3.2)
+
+
+
+\* *This Changelog was automatically generated by [github_changelog_generator](https://github.com/github-changelog-generator/github-changelog-generator)*

data/Gemfile

@@ -1,5 +1,4 @@
 source "http://rubygems.org"
-
 gemspec
 
 gem 'rake'

data/Gemfile.lock

@@ -1,34 +1,52 @@
 PATH
   remote: .
   specs:
-    reckon (0.4.4)
+    reckon (0.5.0)
       chronic (>= 0.3.0)
-      fastercsv (>= 1.5.1)
       highline (>= 1.5.2)
+      rchardet (>= 1.8.0)
       terminal-table (>= 1.4.2)
 
 GEM
   remote: http://rubygems.org/
   specs:
     chronic (0.10.2)
-    diff-lcs (1.1.3)
-    fastercsv (1.5.5)
-    highline (1.6.21)
-    rake (10.0.4)
-    rspec (2.11.0)
-      rspec-core (~> 2.11.0)
-      rspec-expectations (~> 2.11.0)
-      rspec-mocks (~> 2.11.0)
-    rspec-core (2.11.1)
-    rspec-expectations (2.11.2)
-      diff-lcs (~> 1.1.3)
-    rspec-mocks (2.11.1)
-    terminal-table (1.4.5)
+    coderay (1.1.2)
+    diff-lcs (1.3)
+    highline (2.0.3)
+    method_source (0.9.2)
+    pry (0.12.2)
+      coderay (~> 1.1.0)
+      method_source (~> 0.9.0)
+    rake (12.3.3)
+    rantly (1.2.0)
+    rchardet (1.8.0)
+    rspec (3.9.0)
+      rspec-core (~> 3.9.0)
+      rspec-expectations (~> 3.9.0)
+      rspec-mocks (~> 3.9.0)
+    rspec-core (3.9.1)
+      rspec-support (~> 3.9.1)
+    rspec-expectations (3.9.0)
+      diff-lcs (>= 1.2.0, < 2.0)
+      rspec-support (~> 3.9.0)
+    rspec-mocks (3.9.1)
+      diff-lcs (>= 1.2.0, < 2.0)
+      rspec-support (~> 3.9.0)
+    rspec-support (3.9.2)
+    terminal-table (1.8.0)
+      unicode-display_width (~> 1.1, >= 1.1.1)
+    unicode-display_width (1.6.1)
 
 PLATFORMS
   ruby
 
 DEPENDENCIES
+  pry (>= 0.12.2)
   rake
+  rantly (= 1.2.0)
   reckon!
   rspec (>= 1.2.9)
+
+BUNDLED WITH
+   1.17.3

data/README.md

@@ -1,8 +1,8 @@
 # Reckon
 
-[![Build Status](https://travis-ci.org/cantino/reckon.png)](https://travis-ci.org/cantino/reckon)
+[![Build Status](https://travis-ci.org/cantino/reckon.png?branch=master)](https://travis-ci.org/cantino/reckon)
 
-Reckon automagically converts CSV files for use with the command-line accounting tool [Ledger](https://github.com/jwiegley/ledger/wiki).  It also helps you to select the correct accounts associated with the CSV data using Bayesian machine learning.
+Reckon automagically converts CSV files for use with the command-line accounting tool [Ledger](http://www.ledger-cli.org/).  It also helps you to select the correct accounts associated with the CSV data using Bayesian machine learning.
 
 ## Installation
 
@@ -114,6 +114,3 @@ You can override them with `--default_outof_account` and `--default_into_account
 Copyright (c) 2013 Andrew Cantino. See LICENSE for details.
 
 Thanks to @BlackEdder for many contributions!
-
-[![Bitdeli Badge](https://d2weczhvl823v0.cloudfront.net/cantino/reckon/trend.png)](https://bitdeli.com/free "Bitdeli Badge")
-

data/lib/reckon.rb

@@ -1,19 +1,21 @@
 #!/usr/bin/env ruby
 
 require 'rubygems'
-if RUBY_VERSION =~ /^1\.9/ || RUBY_VERSION =~ /^2/
-  require 'csv'
-else
-  require 'fastercsv'
-end
+require 'rchardet'
+require 'chronic'
+require 'csv'
 require 'highline/import'
 require 'optparse'
-require 'chronic'
-require 'time'
 require 'terminal-table'
+require 'time'
+require 'logger'
+
+LOGGER = Logger.new(STDOUT)
+LOGGER.level = Logger::ERROR
 
+require_relative 'reckon/version'
+require_relative 'reckon/cosine_similarity'
 require File.expand_path(File.join(File.dirname(__FILE__), "reckon", "app"))
 require File.expand_path(File.join(File.dirname(__FILE__), "reckon", "ledger_parser"))
 require File.expand_path(File.join(File.dirname(__FILE__), "reckon", "csv_parser"))
 require File.expand_path(File.join(File.dirname(__FILE__), "reckon", "money"))
-

data/lib/reckon/app.rb

@@ -1,21 +1,20 @@
-#coding: utf-8
+# coding: utf-8
 require 'pp'
 require 'yaml'
 
 module Reckon
   class App
-    VERSION = "Reckon 0.4.4"
-    attr_accessor :options, :accounts, :tokens, :seen, :csv_parser, :regexps
+    attr_accessor :options, :seen, :csv_parser, :regexps, :matcher
 
     def initialize(options = {})
+      LOGGER.level = Logger::INFO if options[:verbose]
       self.options = options
-      self.tokens = {}
       self.regexps = {}
-      self.accounts = {}
       self.seen = {}
       self.options[:currency] ||= '$'
       options[:string] = File.read(options[:file]) unless options[:string]
       @csv_parser = CSVParser.new( options )
+      @matcher = CosineSimilarity.new(options)
       learn!
     end
 
@@ -24,21 +23,44 @@ module Reckon
       puts str
     end
 
+    def learn!
+      learn_from_account_tokens(options[:account_tokens_file])
+
+      ledger_file = options[:existing_ledger_file]
+      return unless ledger_file
+      fail "#{ledger_file} doesn't exist!" unless File.exists?(ledger_file)
+      learn_from(File.read(ledger_file))
+    end
+
+    def learn_from_account_tokens(filename)
+      return unless filename
+
+      fail "#{filename} doesn't exist!" unless File.exists?(filename)
+
+      extract_account_tokens(YAML.load_file(filename)).each do |account, tokens|
+        tokens.each do |t|
+          if t.start_with?('/')
+            add_regexp(account, t)
+          else
+            @matcher.add_document(account, t)
+          end
+        end
+      end
+    end
+
     def learn_from(ledger)
       LedgerParser.new(ledger).entries.each do |entry|
         entry[:accounts].each do |account|
-          learn_about_account( account[:name],
-                              [entry[:desc], account[:amount]].join(" ") ) unless account[:name] == options[:bank_account]
-          seen[entry[:date]] ||= {}
-          seen[entry[:date]][@csv_parser.pretty_money(account[:amount])] = true
+          str = [entry[:desc], account[:amount]].join(" ")
+          @matcher.add_document(account[:name], str) unless account[:name] == options[:bank_account]
+          pretty_date = entry[:date].iso8601
+          seen[pretty_date] ||= {}
+          seen[pretty_date][@csv_parser.pretty_money(account[:amount])] = true
         end
       end
     end
 
-    def already_seen?(row)
-      seen[row[:pretty_date]] && seen[row[:pretty_date]][row[:pretty_money]]
-    end
-
+    # Add tokens from account_tokens_file to accounts
     def extract_account_tokens(subtree, account = nil)
       if subtree.nil?
         puts "Warning: empty #{account} tree"
@@ -46,50 +68,26 @@ module Reckon
       elsif subtree.is_a?(Array)
         { account => subtree }
       else
-        at = subtree.map { |k, v| extract_account_tokens(v, [account, k].compact.join(':')) }
-        at.inject({}) { |k, v| k = k.merge(v)}
-      end
-    end
-
-    def learn!
-      if options[:account_tokens_file]
-        fail "#{options[:account_tokens_file]} doesn't exist!" unless File.exists?(options[:account_tokens_file])
-        extract_account_tokens(YAML.load_file(options[:account_tokens_file])).each do |account, tokens|
-          tokens.each { |t| learn_about_account(account, t, true) }
+        at = subtree.map do |k, v|
+          merged_acct = [account, k].compact.join(':')
+          extract_account_tokens(v, merged_acct)
         end
+        at.inject({}) { |memo, e| memo.merge!(e)}
       end
-      return unless options[:existing_ledger_file]
-      fail "#{options[:existing_ledger_file]} doesn't exist!" unless File.exists?(options[:existing_ledger_file])
-      ledger_data = File.read(options[:existing_ledger_file])
-      learn_from(ledger_data)
     end
 
-    def learn_about_account(account, data, parse_regexps = false)
-      accounts[account] ||= 0
-      if parse_regexps && data.start_with?('/')
-        # https://github.com/tenderlove/psych/blob/master/lib/psych/visitors/to_ruby.rb
-        match = data.match(/^\/(.*)\/([ix]*)$/m)
-        fail "failed to parse regexp #{data}" unless match
-        options = 0
-        (match[2] || '').split('').each do |option|
-          case option
-          when 'x' then options |= Regexp::EXTENDED
-          when 'i' then options |= Regexp::IGNORECASE
-          end
-        end
-        regexps[Regexp.new(match[1], options)] = account
-      else
-        tokenize(data).each do |token|
-          tokens[token] ||= {}
-          tokens[token][account] ||= 0
-          tokens[token][account] += 1
-          accounts[account] += 1
+    def add_regexp(account, regex_str)
+      # https://github.com/tenderlove/psych/blob/master/lib/psych/visitors/to_ruby.rb
+      match = regex_str.match(/^\/(.*)\/([ix]*)$/m)
+      fail "failed to parse regexp #{regex_str}" unless match
+      options = 0
+      (match[2] || '').split('').each do |option|
+        case option
+        when 'x' then options |= Regexp::EXTENDED
+        when 'i' then options |= Regexp::IGNORECASE
         end
       end
-    end
-
-    def tokenize(str)
-      str.downcase.split(/[\s\-]/)
+      regexps[Regexp.new(match[1], options)] = account
     end
 
     def walk_backwards
@@ -107,8 +105,7 @@ module Reckon
           seen_anything_new = true
         end
 
-        possible_answers = most_specific_regexp_match(row)
-        possible_answers = weighted_account_match( row ).map! { |a| a[:account] } if possible_answers.empty?
+        possible_answers = suggest(row)
 
         ledger = if row[:money] > 0
           if options[:unattended]
@@ -156,15 +153,19 @@ module Reckon
       end
     end
 
-    def finish
-      options[:output_file].close unless options[:output_file] == STDOUT
-      interactive_output "Exiting."
-      exit
-    end
-
-    def output(ledger_line)
-      options[:output_file].puts ledger_line
-      options[:output_file].flush
+    def each_row_backwards
+      rows = []
+      (0...@csv_parser.columns.first.length).to_a.each do |index|
+        rows << { :date => @csv_parser.date_for(index),
+                  :pretty_date => @csv_parser.pretty_date_for(index),
+                  :pretty_money => @csv_parser.pretty_money_for(index),
+                  :pretty_money_negated => @csv_parser.pretty_money_for(index, :negate),
+                  :money => @csv_parser.money_for(index),
+                  :description => @csv_parser.description_for(index) }
+      end
+      rows.sort { |a, b| a[:date] <=> b[:date] }.each do |row|
+        yield row
+      end
     end
 
     def most_specific_regexp_match( row )
@@ -176,41 +177,9 @@ module Reckon
       matches.sort_by! { |account, matched_text| matched_text.length }.map(&:first)
     end
 
-    # Weigh accounts by how well they match the row
-    def weighted_account_match( row )
-      query_tokens = tokenize(row[:description])
-
-      search_vector = []
-      account_vectors = {}
-
-      query_tokens.each do |token|
-        idf = Math.log((accounts.keys.length + 1) / ((tokens[token] || {}).keys.length.to_f + 1))
-        tf = 1.0 / query_tokens.length.to_f
-        search_vector << tf*idf
-
-        accounts.each do |account, total_terms|
-          tf = (tokens[token] && tokens[token][account]) ? tokens[token][account] / total_terms.to_f : 0
-          account_vectors[account] ||= []
-          account_vectors[account] << tf*idf
-        end
-      end
-
-      # Should I normalize the vectors?  Probably unnecessary due to tf-idf and short documents.
-
-      account_vectors = account_vectors.to_a.map do |account, account_vector|
-        { :cosine => (0...account_vector.length).to_a.inject(0) { |m, i| m + search_vector[i] * account_vector[i] },
-          :account => account }
-      end
-      account_vectors.sort! {|a, b| b[:cosine] <=> a[:cosine] }
-
-      # Return empty set if no accounts matched so that we can fallback to the defaults in the unattended mode
-      if options[:unattended]
-        if account_vectors.first && account_vectors.first[:account]
-          account_vectors = [] if account_vectors.first[:cosine] == 0
-        end
-      end
-
-      return account_vectors
+    def suggest(row)
+      most_specific_regexp_match(row) +
+        @matcher.find_similar(row[:description]).map { |n| n[:account] }
     end
 
     def ledger_format(row, line1, line2)
@@ -220,6 +189,21 @@ module Reckon
       out
     end
 
+    def output(ledger_line)
+      options[:output_file].puts ledger_line
+      options[:output_file].flush
+    end
+
+    def already_seen?(row)
+      seen[row[:pretty_date]] && seen[row[:pretty_date]][row[:pretty_money]]
+    end
+
+    def finish
+      options[:output_file].close unless options[:output_file] == STDOUT
+      interactive_output "Exiting."
+      exit
+    end
+
     def output_table
       output = Terminal::Table.new do |t|
         t.headings = 'Date', 'Amount', 'Description'
@@ -230,21 +214,6 @@ module Reckon
       interactive_output output
     end
 
-    def each_row_backwards
-      rows = []
-      (0...@csv_parser.columns.first.length).to_a.each do |index|
-        rows << { :date => @csv_parser.date_for(index),
-          :pretty_date => @csv_parser.pretty_date_for(index),
-          :pretty_money => @csv_parser.pretty_money_for(index),
-          :pretty_money_negated => @csv_parser.pretty_money_for(index, :negate),
-          :money => @csv_parser.money_for(index),
-          :description => @csv_parser.description_for(index) }
-      end
-      rows.sort { |a, b| a[:date] <=> b[:date] }.each do |row|
-        yield row
-      end
-    end
-
     def self.parse_opts(args = ARGV)
       options = { :output_file => STDOUT }
       parser = OptionParser.new do |opts|
@@ -255,7 +224,7 @@ module Reckon
           options[:file] = file
         end
 
-        opts.on("-a", "--account name", "The Ledger Account this file is for") do |a|
+        opts.on("-a", "--account NAME", "The Ledger Account this file is for") do |a|
           options[:bank_account] = a
         end
 
@@ -283,6 +252,14 @@ module Reckon
           options[:ignore_columns] = ignore.split(",").map { |i| i.to_i }
         end
 
+        opts.on("", "--money-column 2", Integer, "Specify the money column instead of letting Reckon guess - the first column is column 1") do |column_number|
+          options[:money_column] = column_number
+        end
+
+        opts.on("", "--date-column 3", Integer, "Specify the date column instead of letting Reckon guess - the first column is column 1") do |column_number|
+          options[:date_column] = column_number
+        end
+
         opts.on("", "--contains-header [N]", "The first row of the CSV is a header and should be skipped. Optionally add the number of rows to skip.") do |contains_header|
           options[:contains_header] = 1
           options[:contains_header] = contains_header.to_i if contains_header
@@ -316,11 +293,11 @@ module Reckon
           options[:account_tokens_file] = a
         end
 
-        opts.on("", "--default-into-account name", "Default into account") do |a|
+        opts.on("", "--default-into-account NAME", "Default into account") do |a|
           options[:default_into_account] = a
         end
 
-        opts.on("", "--default-outof-account name", "Default 'out of' account") do |a|
+        opts.on("", "--default-outof-account NAME", "Default 'out of' account") do |a|
           options[:default_outof_account] = a
         end
 
@@ -351,7 +328,6 @@ module Reckon
       end
 
       unless options[:bank_account]
-
         fail "Please specify --account for the unattended mode" if options[:unattended]
 
         options[:bank_account] = ask("What is the account name of this bank account in Ledger? ") do |q|