mbox2csv 0.1.1 → 0.2.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 1f4a5c7ffcfb3c8deffd7c05068268b129e5a37ffc2ba01f3024e3e9790c3d2d
-  data.tar.gz: c20e300f2ddd8b2a002bb782109af205532e7be93b539f197937480c4ec0400d
+  metadata.gz: fff3b927d3dc15547ce87943b6bfe1b747423222391f82cb336dbb03b710a056
+  data.tar.gz: 143acbaaf7f95ae5f2b61df3e242bf976d328dfa780d4e3439d764a294ddffd7
 SHA512:
-  metadata.gz: b81692ac0aa4648ee89b5f2d3cd3e704bd9a889b140a82f078009c4075067a1a6d09b58398898a4b215352bc4431d83ec3fbbe8e70213388eb4680442468b4ba
-  data.tar.gz: 15442b56f31f99c8cf17db145057b11d6345f912146e4569de628098b6b0574e5bcc3aed9c1791d58cbc5cfc83a29ba5c4685798dbbf579b5225ad933bc900cc
+  metadata.gz: f68e2ca1627bcb9cf8d1325000b6b00d96b7f856227b1ea2a768962312def6bbd29dd4cce478c286ad1a610915fd5b0bbe90310fd0c370a28ca916681f2e4b35
+  data.tar.gz: 0cf68cb50e9886e174d55844f1b68533795c23cb3a38824bf5c9c3d8c90e2b5454317db10b681ca1328eb483a0a7fca1222d18d1b20106e1a81ecfab336452fc

data/lib/mbox2csv.rb

@@ -1,12 +1,14 @@
 require 'base64'
 require 'csv'
 require 'mail'
+require 'fileutils'
+require 'ruby-progressbar'
 
 module Mbox2CSV
-    # Main class
+    # Main class for parsing MBOX files and saving email data and statistics to CSV files.
     class MboxParser
         # Initializes the MboxParser with file paths for the MBOX file, output CSV file,
-        # and statistics CSV files for sender statistics.
+        # and statistics CSV files for sender and recipient statistics.
         #
         # @param [String] mbox_file Path to the MBOX file to be parsed.
         # @param [String] csv_file Path to the output CSV file where parsed email data will be saved.
@@ -18,30 +20,35 @@ module Mbox2CSV
             @statistics = EmailStatistics.new
             @stats_csv_file = stats_csv_file
             @recipient_stats_csv_file = recipient_stats_csv_file
+            @senders_folder = 'senders/'
+            FileUtils.mkdir_p(@senders_folder) # Create the senders folder if it doesn't exist
         end
 
         # Parses the MBOX file and writes the email data to the specified CSV file.
         # It also saves sender and recipient statistics to separate CSV files.
+        # A progress bar is displayed during the parsing process.
         def parse
+            total_lines = File.foreach(@mbox_file).inject(0) { |c, _line| c + 1 }
+            progressbar = ProgressBar.create(title: "Parsing Emails", total: total_lines, format: "%t: |%B| %p%%")
+
             CSV.open(@csv_file, 'w') do |csv|
-                # Write CSV header
                 csv << ['From', 'To', 'Subject', 'Date', 'Body']
 
                 File.open(@mbox_file, 'r') do |mbox|
                     buffer = ""
                     mbox.each_line do |line|
+                        progressbar.increment
                         if line.start_with?("From ")
                             process_email_block(buffer, csv) unless buffer.empty?
-                            buffer = "" # Reset buffer
+                            buffer = ""
                         end
-                        buffer << line # Append line to buffer
+                        buffer << line
                     end
-                    process_email_block(buffer, csv) unless buffer.empty? # Process last email block
+                    process_email_block(buffer, csv) unless buffer.empty?
                 end
             end
             puts "Parsing completed. Data saved to #{@csv_file}"
 
-            # Save and print statistics after parsing
             @statistics.save_sender_statistics(@stats_csv_file)
             @statistics.save_recipient_statistics(@recipient_stats_csv_file)
         rescue => e
@@ -51,7 +58,8 @@ module Mbox2CSV
         private
 
         # Processes an individual email block from the MBOX file, extracts the email fields,
-        # and writes them to the CSV. Also records email statistics for analysis.
+        # and writes them to the CSV. Also records email statistics for analysis and creates
+        # sender-specific CSV files.
         #
         # @param [String] buffer The email block from the MBOX file.
         # @param [CSV] csv The CSV object where email data is written.
@@ -62,14 +70,13 @@ module Mbox2CSV
             to = ensure_utf8(mail.to ? mail.to.join(", ") : '', 'UTF-8')
             subject = ensure_utf8(mail.subject ? mail.subject : '', 'UTF-8')
             date = ensure_utf8(mail.date ? mail.date.to_s : '', 'UTF-8')
-
             body = decode_body(mail)
 
-            # Write to CSV
             csv << [from, to, subject, date, body]
 
-            # Record email for statistics
             @statistics.record_email(from, to, body.length)
+
+            save_email_to_sender_csv(from, to, subject, date, body)
         rescue => e
             puts "Error processing email block: #{e.message}"
         end
@@ -108,15 +115,45 @@ module Mbox2CSV
         text = text.force_encoding(charset) if charset
         text.encode('UTF-8', invalid: :replace, undef: :replace, replace: '?')
     end
+
+    # Saves an email to a sender-specific CSV file.
+    #
+    # @param [String] from The sender of the email.
+    # @param [String] to The recipient(s) of the email.
+    # @param [String] subject The subject of the email.
+    # @param [String] date The date of the email.
+    # @param [String] body The body of the email.
+    def save_email_to_sender_csv(from, to, subject, date, body)
+        return if from.empty?
+
+        sender_file = File.join(@senders_folder, "#{sanitize_filename(from)}.csv")
+
+        CSV.open(sender_file, 'a') do |csv|
+            if File.size(sender_file).zero?
+                csv << ['From', 'To', 'Subject', 'Date', 'Body'] # Add header if file is new
+            end
+            csv << [from, to, subject, date, body]
+        end
+    rescue => e
+        puts "Error writing to sender CSV file for #{from}: #{e.message}"
+    end
+
+    # Sanitizes filenames by replacing invalid characters with underscores.
+    #
+    # @param [String] filename The input filename.
+    # @return [String] A sanitized version of the filename.
+    def sanitize_filename(filename)
+        filename.gsub(/[^0-9A-Za-z\-]/, '_')
+    end
 end
 
 # The EmailStatistics class is responsible for gathering and writing statistics related to emails.
 # It tracks sender frequency, recipient frequency, and calculates the average email body length per sender.
 class EmailStatistics
     def initialize
-        @sender_counts = Hash.new(0) # Keeps count of emails per sender
-        @recipient_counts = Hash.new(0) # Keeps count of emails per recipient
-        @body_lengths = Hash.new { |hash, key| hash[key] = [] } # Stores body lengths per sender
+        @sender_counts = Hash.new(0)
+        @recipient_counts = Hash.new(0)
+        @body_lengths = Hash.new { |hash, key| hash[key] = [] }
     end
 
     # Records an email's sender, recipients, and body length for statistical purposes.

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: mbox2csv
 version: !ruby/object:Gem::Version
-  version: 0.1.1
+  version: 0.2.0
 platform: ruby
 authors:
 - firefly-cpp
 autorequire:
 bindir: bin
 cert_chain: []
-date: 2024-09-21 00:00:00.000000000 Z
+date: 2025-01-25 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: base64
@@ -52,6 +52,20 @@ dependencies:
     - - "~>"
       - !ruby/object:Gem::Version
         version: 2.8.1
+- !ruby/object:Gem::Dependency
+  name: ruby-progressbar
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: '1.11'
+  type: :runtime
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: '1.11'
 description:
 email:
 - iztok@iztok-jr-fister.eu
@@ -62,13 +76,13 @@ files:
 - LICENSE
 - README.md
 - lib/mbox2csv.rb
-homepage: https://codeberg.org/firefly-cpp/mbox2csv
+homepage: https://github.com/firefly-cpp/mbox2csv
 licenses:
 - MIT
 metadata:
-  homepage_uri: https://codeberg.org/firefly-cpp/mbox2csv
-  source_code_uri: https://codeberg.org/firefly-cpp/mbox2csv
-  changelog_uri: https://codeberg.org/firefly-cpp/mbox2csv
+  homepage_uri: https://github.com/firefly-cpp/mbox2csv
+  source_code_uri: https://github.com/firefly-cpp/mbox2csv
+  changelog_uri: https://github.com/firefly-cpp/mbox2csv
 post_install_message:
 rdoc_options: []
 require_paths:
@@ -84,7 +98,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
     - !ruby/object:Gem::Version
       version: '0'
 requirements: []
-rubygems_version: 3.5.16
+rubygems_version: 3.5.22
 signing_key:
 specification_version: 4
 summary: Parse MBOX files and export email data into CSV format