universal_excel_parser 1.1.0

checksums.yaml

@@ -0,0 +1,7 @@
+---
+SHA1:
+  metadata.gz: a042c3feaaebe5d532e85ccb7b6cfd9ff6d787dc
+  data.tar.gz: e46cb8657a18bc2e93581e8e2d48565685fa62aa
+SHA512:
+  metadata.gz: 2b99cff666a7d8ba425f3e018d58048b82df2bd533be7c8af70d7585b92cd095c45c2f8d2ce2bf13d938e428500d041643b6412e3a6defadf3516c698f5f6258
+  data.tar.gz: 5c3c75374cec78681f9a9cb0efd0b57d1b2158143952f356d3ee27afd72023ec27cfdd310acd02dc0712c1763ab5ecd6537be980c74a0be74f397572c9fca685

data/lib/universal_excel_parser.rb

@@ -0,0 +1,221 @@
+require 'simple_xlsx_reader'
+require 'logger'
+class UniversalExcelParser
+  attr_accessor :xls_file, :file_path, :data_hash, :validator_hashes, :logger
+
+  def initialize(p_file_path, p_data_hash, p_validator_hashes, p_logger_level = Logger::ERROR)
+    self.file_path = p_file_path
+    self.data_hash = p_data_hash
+    self.validator_hashes = p_validator_hashes
+    self.logger = Logger.new(STDOUT)
+    logger.level = p_logger_level
+  end
+
+  def self.parse_hash_with_sheet_identification(file_path, data_hash, validator_hashes={})
+    uep = UniversalExcelParser.new(file_path, data_hash, validator_hashes)
+    uep.logger.info "READING FILE #{file_path})"
+    return false unless uep.validate_file_type(uep.file_path)
+    return false unless uep.read_file(uep.file_path)
+    return false unless uep.validate_readability_of_all_sheets_with_identification(data_hash)
+    data = uep.read_hash_with_sheet_identification(data_hash)
+    uep.logger.info "READING FILE #{file_path}: FINISHED"
+    data
+  end
+
+  def read_hash_with_sheet_identification(input)
+    logger.debug "[+] READING THE FILES WITH SHEET IDENTIFICATION"
+    # HASH SYNTAX:  {:CUSTOM_NAME => {:sheet_identification =>  {:row => 123, :column => 123, :content => "TEXT"}, :data => [{:KEY_IS_KEY => [ROW, COLUMN]}}
+    unless input.class == Hash
+      logger.error "  [ERROR] INPUT IS NOT AN HASH"
+      return false
+    end
+    output_data = {}
+    sheet_ids = (0..(xls_file.sheets.count-1)).to_a
+    input.each do |key,value|
+      sheet_ids.delete_if do |sid|
+        next unless field_exists?(sid, value[:sheet_identification][:row], value[:sheet_identification][:column])
+        next unless self.xls_file.sheets[sid].rows[value[:sheet_identification][:row]][value[:sheet_identification][:column]] == value[:sheet_identification][:content]
+        hash = {}
+        value[:data].each do |data_key, data_value|
+          unless field_exists?(sid, data_value[0], data_value[1])
+            logger.error "FIELD DOES NOT EXISTS [#{sid}][#{data_value[0]}][#{data_value[1]}]"
+          end
+          hash[data_key] = self.xls_file.sheets[sid].rows[data_value[0]][data_value[1]]
+        end
+        output_data[key] = [] if output_data[key].class != Array
+        output_data[key] << hash
+        true
+      end   
+    end
+    logger.debug " [ERROR] SHEETS THAT NOT READ: #{sheet_ids}"
+    output_data
+  end
+
+  def validate_readability_of_all_sheets_with_identification(input)
+    logger.debug "[+] TRYING TO READ ALL SHEETS"
+    # HASH SYNTAX:  {:CUSTOM_NAME => {:sheet_identification =>  {:row => 123, :column => 123, :content => "TEXT"}, :data => [{:KEY_IS_KEY => [ROW, COLUMN]}}
+    unless input.class == Hash
+      logger.error "  [ERROR] INPUT IS NOT AN HASH"
+      return false
+    end
+    output_data = {}
+    sheet_ids = (0..(xls_file.sheets.count-1)).to_a
+    input.each do |key,value|
+      sheet_ids.delete_if do |sid|
+        next unless field_exists?(sid, value[:sheet_identification][:row], value[:sheet_identification][:column])
+        next unless self.xls_file.sheets[sid].rows[value[:sheet_identification][:row]][value[:sheet_identification][:column]] == value[:sheet_identification][:content]
+        hash = {}
+        value[:data].each do |data_key, data_value|
+          return false unless field_exists?(sid, data_value[0], data_value[1])
+          hash[data_key] = self.xls_file.sheets[sid].rows[data_value[0]][data_value[1]]
+        end
+        output_data[key] = [] if output_data[key].class != Array
+        output_data[key] << hash
+        true
+      end
+    end
+    if sheet_ids.count != 0
+      logger.error "I CANT READ ALL SHEETS. NOT READABLE SHEETS:"
+      sheet_ids.each do |sid|
+        logger.error " - #{xls_file.sheets[sid].name}"
+      end
+      return false
+    end
+    true
+  end
+
+  def file_has_sheets?
+    if self.xls_file.sheets.count != 0
+      logger.debug "  [+] FILE HAS #{xls_file.sheets.count} SHEETS"
+      return true
+    else
+      logger.error "  [ERROR] FILE HAS NO SHEETS"
+      return false
+    end
+  end
+
+  def validate_sheet_count(count)
+    if xls_file.sheets.count == value
+      logger.debug "    [+] THE FILE HAS THE CORRECT COUNT OF SHEETS(#{value}/#{xls_file.sheets.count})"
+      return true
+    else
+      logger.error "    [ERROR] THE FILE HAS A WRONG COUNT OF SHEETS(#{value}/#{xls_file.sheets.count})"
+      return false
+    end
+  end
+
+  def field_exists?(sheet, row, column)
+    return false unless row_exists?(sheet, row)
+    unless hash[column].class == Integer
+      logger.error "    [ERROR] COLUMN NUMBER IS NOT AN INTEGER!"
+      return false
+    end
+    if xls_file.sheets[sheet].rows[row].count <= column
+      logger.error "    [ERROR] COLUMN NUMBER IS HIGHER THAN THE ROWS IN THE FILE!"
+      return false
+    end
+    true
+  end
+
+  def row_exists?(sheet, row)
+    return false unless sheet_exists?(sheet)
+    unless row.class == Integer
+      logger.error "    [ERROR] ROW IS NOT AN INTEGER!"
+      return false
+    end
+    if xls_file.sheets[sheet].rows.count <= row
+      logger.error "    [ERROR] ROW NUMBER IS HIGHER THAN THE ROWS IN FILE!"
+      return false
+    end
+    true
+  end
+
+  def sheet_exists?(sheet)
+    logger.debug "[+] CHECKING IF THE SHEET EXISTS"
+    unless sheet.class == Integer
+      logger.error "  [ERROR] SHEET NUMBER IS NOT AN INTEGER!"
+      return false
+    end
+    if xls_file.sheets.count <= sheet
+      logger.error "  [ERROR] SHEET NUMBER IS HIGHER THAN THE SHEET COUNT!"
+      return false
+    end
+    logger.debug "[-] SHEET IS VALID"
+    true
+  end
+
+  def validate_content(validator_hash, sid)
+    # SYNTAX: 
+    #   {
+    #     :field_content => [
+    #       {
+    #         :row => 0,
+    #         :column => 0,
+    #         :content => "TEXT_IN_FIELD"
+    #        }
+    #     ]
+    #   }
+    logger.debug "[+] VALIDATE CONTENT OF EXCEL_SHEET WITH ID #{sid} AND VALIDATOR_HASH #{validator_hash}"
+    return false unless file_has_sheets?
+    return false unless sheet_exists?(sid)
+    validator_hash.keys.each do |key, value|
+      if key == :field_content
+        logger.debug "  [+] VALIDATE THE CONTENT OF SOME FIELDS"
+        if value.class == Array
+          value.each do |hash|
+            logger.debug "    [+] VALIDATING FIELD #{self.xls_file.sheets[sid].name}[#{hash[:row]}][#{hash[:column]}]"
+            return false unless field_exists?(sid, hash[:row], hash[:column])
+            if xls_file.sheets[sid].rows[hash[:row]][hash[:column]] == hash[:content]
+              logger.debug "      [OK] FIELD IS VALID!"
+            else
+              logger.error "      [ERROR] FIELD IS NOT VALID! SHEET: #{self.xls_file.sheets[sid].name}[#{hash[:row]}][#{hash[:column]}] FIELD-CONTENT: #{xls_file.sheets[sid].rows[hash[:row]][hash[:column]]} VALIDATOR: #{hash}"
+              return false
+            end
+            logger.debug "    [-] FIELD IS VALID"
+          end
+        else
+          logger.error "    [ERROR] validator_hash[:field_content] IS NOT AN ARRAY: validator_hash(#{validator_hash})"
+          return false
+        end
+        logger.debug "  [-] FIELDS ARE VALID"
+      else
+        logger.error "  [ERROR] validator_hash: #{key} IS NOT IMPLEMENTED"
+        return false
+      end
+    end
+    logger.debug "[-] EXCEL_SHEET IS VALID"
+    true
+  end
+
+  def validate_file_type(file_path)
+    logger.debug "[+] VALIDATING FILE TYPE OF #{file_path}"
+    if File.exists?(file_path)
+      logger.debug "  [OK] THE FILE EXISTS"
+    else
+      logger.error "  [ERROR] THE FILE DOES NOT EXISTS: function(validate_file_type) file_path(#{file_path})"
+      return false
+    end
+    mime_type = `file --brief --mime-type "#{file_path}"`.strip
+    if mime_type == "application/zip" or mime_type == "application/vnd.openxmlformats-officedocument.spreadsheetml.sheet"
+      logger.debug "  [OK] MIME TYPE IS #{mime_type} READABLE"
+    else
+      logger.error "  [ERROR] MIME TYPE IS NOT COMPATIBLE: function(validate_file_type) file_path(#{file_path})"
+      return false
+    end
+    logger.debug "[-] FILE TYPE IS VALID"
+    true
+  end
+
+  def read_file(file_path)
+    logger.debug "[+] TRYING TO READ FILE #{file_path}"
+    self.xls_file = SimpleXlsxReader.open(file_path)
+    if self.xls_file.kind_of?(SimpleXlsxReader::Document)
+      logger.debug "  [OK] FILE IS READABLE WITH THE SimpleXlsxReader"
+    else
+      logger.error "  [ERROR] FILE IS NOT READABLE WITH The SimpleXlsxReader: function(read_file) file_path(#{file_path})"
+      return false
+    end
+    logger.debug "[-] END TRYING TO READ FILE"
+    true
+  end
+end

metadata

@@ -0,0 +1,57 @@
+--- !ruby/object:Gem::Specification
+name: universal_excel_parser
+version: !ruby/object:Gem::Version
+  version: 1.1.0
+platform: ruby
+authors:
+- cdt
+autorequire: 
+bindir: bin
+cert_chain: []
+date: 2017-08-10 00:00:00.000000000 Z
+dependencies:
+- !ruby/object:Gem::Dependency
+  name: simple_xlsx_reader
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: '1.0'
+  type: :runtime
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: '1.0'
+description: 
+email: codingdt@gmx.de
+executables: []
+extensions: []
+extra_rdoc_files: []
+files:
+- lib/universal_excel_parser.rb
+homepage: http://rubygems.org/gems/universal_excel_parser
+licenses: []
+metadata: {}
+post_install_message: 
+rdoc_options: []
+require_paths:
+- lib
+required_ruby_version: !ruby/object:Gem::Requirement
+  requirements:
+  - - ">="
+    - !ruby/object:Gem::Version
+      version: '0'
+required_rubygems_version: !ruby/object:Gem::Requirement
+  requirements:
+  - - ">="
+    - !ruby/object:Gem::Version
+      version: '0'
+requirements: []
+rubyforge_project: 
+rubygems_version: 2.5.2
+signing_key: 
+specification_version: 4
+summary: Hash based excel parser
+test_files: []