whatsapp-chat-parser 0.1.0 → 0.1.2

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: c63551e684919c385110d12bcbbe8e15bae0eef6e4ffe31a0617970f45f24bd6
-  data.tar.gz: a510d3a34c51154faae21a8f69d7629e6f0f0ca103a4cfddfbb7571013cc8d3c
+  metadata.gz: 4e92031c7088862cb6488753801997b0cb67453b61275068a4b2e430d85f83a0
+  data.tar.gz: 0df5654a6a3590b01596d0fa00e7081699a498db115ab4b8f337b4ababc0c6f2
 SHA512:
-  metadata.gz: c35b864edcaef50faed6aff4c32d80c8d12e7a624819165766a3e935052b7148aff429dfb674c714dae32b938b2f0e3a654e0fbf8628f2147ead5c46e46dfaa6
-  data.tar.gz: 16c197ee4c6ffc1f1e5ecfc94aadabde22a2b5329ab0b7059ffc2897b18afeb3f6b879a799649b98ac162506162d9a965a0ecfcd6e9423c0f11192a2f9880d48
+  metadata.gz: 6167122bde0e631130963719035eb70bce2764c59664c03f451f0ea4c7108511f1bff5a26ee1a3b88706763413b2f20d6c1eff96b91bd13f2292a9689810ef49
+  data.tar.gz: fed7b89133841f13cbef30a270303caea6895db23e8d4b64df6d4350b2d4dfcce1baf83d796a67519c8dae317273673d4421bfb9a79ceeb5d688a08004bbcdc9

data/CHANGELOG.md

@@ -5,6 +5,18 @@ All notable changes to this project will be documented in this file.
 The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
 and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
 
+## [0.1.2] - 2026-02-26
+
+### Fixed
+- Fixed `Encoding::CompatibilityError` caused by stripping an unscrubbed message line
+- Fixed `FrozenError` caused by mutating a frozen `message` string during message line accumulation
+
+## [0.1.1] - 2026-02-22
+
+### Changed
+- Refactored platform-specific parsers (`Android`, `Ios`) into a shared `Base` module to improve code maintainability (DRY).
+- Optimized timestamp extraction logic in `Base` module to handle optional seconds across all platforms.
+
 ## [0.1.0] - 2026-02-18
 
 ### Added

data/README.md

@@ -1,5 +1,8 @@
 # WhatsApp Chat Parser
 
+[![Gem Version](https://badge.fury.io/rb/whatsapp-chat-parser.svg?icon=si%3Arubygems)](https://badge.fury.io/rb/whatsapp-chat-parser)
+[![CI](https://github.com/emmaakachukwu/whatsapp-chat-parser-rb/actions/workflows/ci.yml/badge.svg)](https://github.com/emmaakachukwu/whatsapp-chat-parser-rb/actions/workflows/ci.yml)
+
 A Ruby library that parses exported WhatsApp chat `.txt` files and converts them into structured, machine-readable data. Designed for downstream processing such as analytics, ETL pipelines, storage, and transformation - not for rendering UI or interacting with the WhatsApp API.
 
 ## Features

data/lib/whatsapp-chat-parser/encoding.rb

@@ -22,7 +22,7 @@ module WhatsappChatParser
           )
         end
 
-        str
+        str.scrub(' ').squeeze(' ')
       end
 
       private

data/lib/whatsapp-chat-parser/file_processor.rb

@@ -30,7 +30,7 @@ module WhatsappChatParser
       end
 
       def accumulate_messages(io, &block)
-        message = ''
+        message = +''
 
         io.each_line do |line|
           if message_starts_here?(line)

data/lib/whatsapp-chat-parser/platforms/android.rb

@@ -4,81 +4,29 @@ module WhatsappChatParser
   module Platforms
     # Parser for Android WhatsApp chat exports.
     module Android
+      extend Base
+
       class << self
         # Parses a line from an Android export.
         # @param line [String] The exported line.
         # @return [Models::Message, nil]
         def parse(line)
-          match = line.match(Pattern.regex)
-          return unless match
-
-          timestamp = extract_timestamp(match)
-          author = extract(match, :author)
-          body = extract(match, :body)
-
-          Models::Message.new(timestamp: timestamp, author: author, body: body, platform: :android)
+          super(line, :android, Pattern)
         end
 
         # Checks if a line matches the Android format.
         # @param line [String]
         # @return [Boolean]
         def matches?(line)
-          Pattern.regex.match?(line)
+          super(line, Pattern)
         end
 
         private
 
-        def extract(match, key)
-          index = Pattern::PATTERNS.keys.index(key)
-          match[index + 1]
-        end
-
-        def extract_timestamp(match)
-          date_components = extract_date_components(match)
-          time_components = extract_time_components(match)
-
-          format_sql_timestamp(date_components, time_components)
-        end
-
-        def extract_date_components(match)
-          month = extract(match, :month)
-          day = extract(match, :day)
-          year = extract(match, :year).to_i + 2000
-
-          { month: month, day: day, year: year }
-        end
-
-        def extract_time_components(match)
-          hour = extract(match, :hour).to_i
-          minute = extract(match, :minute).to_i
-          meridiem = extract(match, :meridiem)
-          hour = convert_to_24_hour(hour, meridiem)
-
-          { hour: hour, minute: minute }
-        end
-
-        def convert_to_24_hour(hour, meridiem)
-          meridiem = meridiem.upcase
-          if meridiem == 'PM' && hour < 12
-            hour + 12
-          elsif meridiem == 'AM' && hour == 12
-            0
-          else
-            hour
-          end
-        end
-
-        def format_sql_timestamp(date, time)
-          # rubocop:disable Layout/HashAlignment
-          format(
-            '%<year>04d-%<month>02d-%<day>02d %<hour>02d:%<minute>02d:00',
-            year:   date[:year],
-            month:  date[:month],
-            day:    date[:day],
-            hour:   time[:hour],
-            minute: time[:minute]
-          )
-          # rubocop:enable Layout/HashAlignment
+        def extract_date_components(match, pattern_module)
+          components = super
+          components[:year] = components[:year].to_i + 2000
+          components
         end
       end
     end

data/lib/whatsapp-chat-parser/platforms/base.rb

@@ -0,0 +1,118 @@
+# frozen_string_literal: true
+
+module WhatsappChatParser
+  module Platforms
+    # Base module for shared platform parser logic
+    module Base
+      # Parses a line from an export
+      # @param line [String] The exported line
+      # @param platform [Symbol] The platform identifier
+      # @param pattern_module [Module] The pattern module for the platform
+      # @return [Models::Message, nil]
+      def parse(line, platform, pattern_module)
+        match = line.match(pattern_module.regex)
+        return unless match
+
+        timestamp = extract_timestamp(match, pattern_module)
+        author = extract(match, pattern_module, :author)
+        body = extract(match, pattern_module, :body)
+
+        Models::Message.new(timestamp: timestamp, author: author, body: body, platform: platform)
+      end
+
+      # Checks if a line matches the platform format
+      # @param line [String]
+      # @param pattern_module [Module]
+      # @return [Boolean]
+      def matches?(line, pattern_module)
+        pattern_module.regex.match?(line)
+      end
+
+      private
+
+      # Extracts a value from a match based on a pattern key
+      # @param match [MatchData]
+      # @param pattern_module [Module]
+      # @param key [Symbol]
+      # @return [String, nil]
+      def extract(match, pattern_module, key)
+        index = pattern_module::PATTERNS.keys.index(key)
+        return unless index
+
+        match[index + 1]
+      end
+
+      # Extracts and formats a timestamp from a match
+      # @param match [MatchData]
+      # @param pattern_module [Module]
+      # @return [String]
+      def extract_timestamp(match, pattern_module)
+        date_components = extract_date_components(match, pattern_module)
+        time_components = extract_time_components(match, pattern_module)
+
+        format_sql_timestamp(date_components, time_components)
+      end
+
+      # Extracts date components from a match
+      # @param match [MatchData]
+      # @param pattern_module [Module]
+      # @return [Hash]
+      def extract_date_components(match, pattern_module)
+        {
+          month: extract(match, pattern_module, :month),
+          day: extract(match, pattern_module, :day),
+          year: extract(match, pattern_module, :year)
+        }
+      end
+
+      # Extracts time components from a match
+      # @param match [MatchData]
+      # @param pattern_module [Module]
+      # @return [Hash]
+      def extract_time_components(match, pattern_module)
+        hour = extract(match, pattern_module, :hour).to_i
+        minute = extract(match, pattern_module, :minute).to_i
+        second = extract(match, pattern_module, :second).to_i
+        meridiem = extract(match, pattern_module, :meridiem)
+        hour = convert_to_24_hour(hour, meridiem)
+
+        { hour: hour, minute: minute, second: second }
+      end
+
+      # Converts an hour to 24-hour format
+      # @param hour [Integer]
+      # @param meridiem [String, nil]
+      # @return [Integer]
+      def convert_to_24_hour(hour, meridiem)
+        return hour unless meridiem
+
+        meridiem = meridiem.upcase
+        if meridiem == 'PM' && hour < 12
+          hour + 12
+        elsif meridiem == 'AM' && hour == 12
+          0
+        else
+          hour
+        end
+      end
+
+      # Formats date and time components into an SQL timestamp
+      # @param date [Hash]
+      # @param time [Hash]
+      # @return [String]
+      def format_sql_timestamp(date, time)
+        # rubocop:disable Layout/HashAlignment
+        format(
+          '%<year>04d-%<month>02d-%<day>02d %<hour>02d:%<minute>02d:%<second>02d',
+          year:   date[:year],
+          month:  date[:month],
+          day:    date[:day],
+          hour:   time[:hour],
+          minute: time[:minute],
+          second: time[:second] || 0
+        )
+        # rubocop:enable Layout/HashAlignment
+      end
+    end
+  end
+end

data/lib/whatsapp-chat-parser/platforms/ios.rb

@@ -2,84 +2,23 @@
 
 module WhatsappChatParser
   module Platforms
-    # Parser for iOS WhatsApp chat exports.
+    # Parser for iOS WhatsApp chat exports
     module Ios
+      extend Base
+
       class << self
-        # Parses a line from an iOS export.
-        # @param line [String] The exported line.
+        # Parses a line from an iOS export
+        # @param line [String] The exported line
         # @return [Models::Message, nil]
         def parse(line)
-          match = line.match(Pattern.regex)
-          return unless match
-
-          timestamp = extract_timestamp(match)
-          author = extract(match, :author)
-          body = extract(match, :body)
-
-          Models::Message.new(timestamp: timestamp, author: author, body: body, platform: :ios)
+          super(line, :ios, Pattern)
         end
 
-        # Checks if a line matches the iOS format.
+        # Checks if a line matches the iOS format
         # @param line [String]
         # @return [Boolean]
         def matches?(line)
-          Pattern.regex.match?(line)
-        end
-
-        private
-
-        def extract(match, key)
-          index = Pattern::PATTERNS.keys.index(key)
-          match[index + 1]
-        end
-
-        def extract_timestamp(match)
-          date_components = extract_date_components(match)
-          time_components = extract_time_components(match)
-
-          format_sql_timestamp(date_components, time_components)
-        end
-
-        def extract_date_components(match)
-          month = extract(match, :month)
-          day = extract(match, :day)
-          year = extract(match, :year)
-
-          { month: month, day: day, year: year }
-        end
-
-        def extract_time_components(match)
-          hour = extract(match, :hour).to_i
-          minute = extract(match, :minute).to_i
-          second = extract(match, :second)
-          meridiem = extract(match, :meridiem)
-          hour = convert_to_24_hour(hour, meridiem)
-
-          { hour: hour, minute: minute, second: second }
-        end
-
-        def convert_to_24_hour(hour, meridiem)
-          if meridiem == 'PM' && hour < 12
-            hour + 12
-          elsif meridiem == 'AM' && hour == 12
-            0
-          else
-            hour
-          end
-        end
-
-        def format_sql_timestamp(date, time)
-          # rubocop:disable Layout/HashAlignment
-          format(
-            '%<year>04d-%<month>02d-%<day>02d %<hour>02d:%<minute>02d:%<second>02d',
-            year:   date[:year],
-            month:  date[:month],
-            day:    date[:day],
-            hour:   time[:hour],
-            minute: time[:minute],
-            second: time[:second]
-          )
-          # rubocop:enable Layout/HashAlignment
+          super(line, Pattern)
         end
       end
     end

data/lib/whatsapp-chat-parser/platforms.rb

@@ -1,6 +1,7 @@
 # frozen_string_literal: true
 
 require_relative 'encoding'
+require_relative 'platforms/base'
 require_relative 'platforms/android'
 require_relative 'platforms/ios'
 require_relative 'platforms/android/pattern'
@@ -31,7 +32,7 @@ module WhatsappChatParser
       end
 
       def sanitize(line)
-        Encoding.normalize_to_utf8(line).strip.scrub(' ').squeeze(' ')
+        Encoding.normalize_to_utf8(line).strip
       end
     end
   end

data/lib/whatsapp-chat-parser/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module WhatsappChatParser
-  VERSION = '0.1.0'
+  VERSION = '0.1.2'
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: whatsapp-chat-parser
 version: !ruby/object:Gem::Version
-  version: 0.1.0
+  version: 0.1.2
 platform: ruby
 authors:
 - Emmanuel Akachukwu
 autorequire:
 bindir: bin
 cert_chain: []
-date: 2026-02-18 00:00:00.000000000 Z
+date: 2026-02-26 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: rspec
@@ -100,6 +100,7 @@ files:
 - lib/whatsapp-chat-parser/platforms.rb
 - lib/whatsapp-chat-parser/platforms/android.rb
 - lib/whatsapp-chat-parser/platforms/android/pattern.rb
+- lib/whatsapp-chat-parser/platforms/base.rb
 - lib/whatsapp-chat-parser/platforms/ios.rb
 - lib/whatsapp-chat-parser/platforms/ios/pattern.rb
 - lib/whatsapp-chat-parser/platforms/pattern_helpers.rb
@@ -110,9 +111,9 @@ licenses:
 metadata:
   homepage_uri: https://github.com/emmaakachukwu/whatsapp-chat-parser-rb
   bug_tracker_uri: https://github.com/emmaakachukwu/whatsapp-chat-parser-rb/issues
-  changelog_uri: https://github.com/emmaakachukwu/whatsapp-chat-parser-rb/blob/v0.1.0/CHANGELOG.md
-  documentation_uri: https://www.rubydoc.info/gems/whatsapp-chat-parser/0.1.0
-  source_code_uri: https://github.com/emmaakachukwu/whatsapp-chat-parser-rb/tree/v0.1.0
+  changelog_uri: https://github.com/emmaakachukwu/whatsapp-chat-parser-rb/blob/v0.1.2/CHANGELOG.md
+  documentation_uri: https://www.rubydoc.info/gems/whatsapp-chat-parser/0.1.2
+  source_code_uri: https://github.com/emmaakachukwu/whatsapp-chat-parser-rb/tree/v0.1.2
   keywords: whatsapp chat parser whatsapp-chat-parser text export android ios
   rubygems_mfa_required: 'true'
 post_install_message: