easy_sheet_io 0.1.2 → 0.2.1

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 2b4655b4dbf3e5a8bbb152f70018068103bc335f4f2d855b82bf0ecb27eb045c
-  data.tar.gz: a07f5bb08b09eb724b8a38d67dc7fe6e571ed040d558e76cff2a164850e28e09
+  metadata.gz: 3fe769a46f695624ce4e05905ed33313b749c5a77546ac545f723ff63950f430
+  data.tar.gz: abc63c91428ae99b41c80ed9846ae82970a1f43bc314a3b3d381f32cbf850a58
 SHA512:
-  metadata.gz: 883e6250a702be56a3a072129494bd8650fcb09db21ba2251fc9668dc078d0f3b7747889441702104a5772b2452ec85d00bd580c2f982db66bd33a496739f8db
-  data.tar.gz: 73d9b766d85d9489bfb5806fd0bb2e1ebba7333cb17511358ff0a0bf1541d664707193debf4097c394b35e126789fb1ec07ef37aefc4229c7cddb946cf36332f
+  metadata.gz: 8d75d2aa249909cd5310d2cd7ba360daced800dbcba7a57d3e53050d60a2318720289532c7cfe74f5ccf29fbe8b137debce92ec20d565105c436395b909fb0a3
+  data.tar.gz: 3ebddcd9ad90b22a43ab6a039029b79b91a897f6f8b4c3a6bbc547f5c878af6dc0e7e5287c379f055d3e0b0accbbffd460bb164b779b7a750f91148c1603c7cf

data/Gemfile.lock

@@ -1,7 +1,7 @@
 PATH
   remote: .
   specs:
-    easy_sheet_io (0.1.0)
+    easy_sheet_io (0.2.1)
 
 GEM
   remote: https://rubygems.org/

data/easy_sheet_io.gemspec

@@ -9,7 +9,7 @@ Gem::Specification.new do |spec|
   spec.email         = ["shun_yamaguchi_tc@live.jp"]
 
   spec.summary       = "A simple way to Open .csv, .xls, .xlsx files."
-  spec.description   = "A simple way to Open .csv, .xls, .xlsx files. You can convert it to 2D Array, Hash, Dataframe."
+  spec.description   = "A simple way to Open .csv, .xls, .xlsx files. You can convert it to 2D array, hash, data frame."
   spec.homepage      = "https://github.com/show-o-atakun/easy_sheet_io"
   spec.license       = "MIT"
   spec.required_ruby_version = Gem::Requirement.new(">= 2.3.0")

data/lib/easy_sheet_io/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module EasySheetIo
-  VERSION = "0.1.2"
+  VERSION = "0.2.1"
 end

data/lib/easy_sheet_io.rb

@@ -17,29 +17,66 @@ module EasySheetIo
 		return /csv$/ === path ? read_csv(path, **opt) : read_excel(path, **opt)
 	end
 	
-	# ##Generate DF from CSV File
+	# ##Generate Array from CSV File, and convert it to Hash or DataFrame.
 	# **opt candidate= line_from: 1, header: 0
-	def read_csv(path, format: :hash, **opt)
-		csv = CSV.parse(File.open path, &:read) # Get 2D Array
+	def read_csv(path, format: nil, encoding: "utf-8", col_sep: ",", **opt)
+		# Get 2D Array
+		begin
+			csv = CSV.parse(File.open(path, encoding: encoding, &:read), col_sep: col_sep)
+		rescue Encoding::InvalidByteSequenceError
+			# Try Another Encoding
+			puts "Fail Encoding #{encoding}. Trying cp932..."
+			csv = CSV.parse(File.open(path, encoding: "cp932", &:read), col_sep: col_sep)
+		end
+		
+		return csv if format.nil?
+
 		ans = to_hash(csv, **opt)
 		return format==:hash || format=="hash" ? ans : to_df(ans, format: format)
 	end
 
-	# ##Generate DF from Excel File
+	# ##Generate Array from EXCEL File, and convert it to Hash or DataFrame.
 	# **opt candidate= line_from: 1, header: 0)
-	def read_excel(path, sheet_i: 0, format: :hash, **opt)
-		a2d = open_excel(path, sheet_i) # Get 2D Array
+	def read_excel(path, sheet_i: 0, format: nil, encoding: "utf-8", **opt)
+		a2d = open_excel(path, sheet_i, encoding: encoding) # Get 2D Array
+		return a2d if format.nil?
+		
 		ans = to_hash(a2d, **opt)
 		return format==:hash || format=="hash" ? ans : to_df(ans, format: format)
 	end
 	
 	# Convert 2d Array to Hash
 	# ##header: nil -> Default Headers(:column1, column2,...) are generated.
-	def to_hash(array2d, line_from: 1, header: 0)
-		output = array2d[line_from..]
-		hd = header.nil? ? [*0...(output.longest_line)].map{"column#{_1}"} : array2d[header]
-		output_transpose = output[0].zip(*output[1..])
+	# line_until=nil means the data are picked up until the end line.
+	def to_hash(array2d, line_from: 1, line_until: nil, header: 0)
+		
+		# Define Read Range------------		
+		lfrom, luntil = line_from, line_until
+		lf_reg, lu_reg = line_from.kind_of?(Regexp), line_until.kind_of?(Regexp)
 		
+		if lf_reg || lu_reg
+			lines_ary = array2d.map{ _1.join "," }
+			lfrom = lines_ary.find_index{ line_from === _1 } if lf_reg
+			luntil = (lines_ary.length-1) - lines_ary.reverse.find_index{ line_until === _1 } if lu_reg
+		end
+		# -----------------------------
+			
+		# Define Data Array------------
+		output = array2d[lfrom...luntil]
+		output_transpose = output[0].zip(*output[1..])
+		# -----------------------------
+
+		# Define Header----------------
+		if header.nil? || header=="string" || header==:string
+			hd = [*0...(output.longest_line)].map{"column#{_1}"}
+		elsif header=="symbol" || header==:symbol
+			hd = [*0...(output.longest_line)].map{"column#{_1}".intern}
+		else
+			hd = array2d[header]
+		end
+		# -----------------------------
+
+		# Make Hash(Header => Data Array)  
 		return hd.each_with_object({}).with_index {|(hdr, hash), i| hash[hdr]=output_transpose[i]}
 	end
 	
@@ -53,18 +90,26 @@ module EasySheetIo
 	end
 	
 	# ##Genarate Hash from excel file
-	def open_excel(path, sheet_i)
-		begin
-			book = /xlsx$/ === path ? Roo::Excelx.new(path) : Roo::Excel.new(path)
+	def open_excel(path, sheet_i, encoding: "utf-8")
+		if /xlsx$/ === path
+			puts "Sorry, encoding option is not supported yet for xlsx file." if encoding != "utf-8"
+
+			book = Roo::Excelx.new(path)
 			s = book.sheet(sheet_i)
 			
-			## bottole neck===
+			## bottole neck
 			return s.to_a
-			
-		rescue Encoding::InvalidByteSequenceError
-		
-			Spreadsheet.client_encoding="Windows-31J"
-			ss = Spreadsheet.open(path)
+
+		# xls
+		else
+			begin
+				Spreadsheet.client_encoding = encoding
+				ss = Spreadsheet.open(path)
+			rescue Encoding::InvalidByteSequenceError
+				puts "Fail Encoding #{encoding}. Trying Windows-31J..."
+				Spreadsheet.client_encoding = "Windows-31J"
+				ss = Spreadsheet.open(path)
+			end
 
 			a2d = []
 			ss.worksheets[sheet_i].rows.each do |row|

metadata

@@ -1,17 +1,17 @@
 --- !ruby/object:Gem::Specification
 name: easy_sheet_io
 version: !ruby/object:Gem::Version
-  version: 0.1.2
+  version: 0.2.1
 platform: ruby
 authors:
 - show-o-atakun
 autorequire:
 bindir: exe
 cert_chain: []
-date: 2022-03-18 00:00:00.000000000 Z
+date: 2022-03-23 00:00:00.000000000 Z
 dependencies: []
 description: A simple way to Open .csv, .xls, .xlsx files. You can convert it to 2D
-  Array, Hash, Dataframe.
+  array, hash, data frame.
 email:
 - shun_yamaguchi_tc@live.jp
 executables: []