easy_sheet_io 0.1.1 → 0.2.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 8cf3f5ae0805f655b09de03abc6465131b4d9f4eba1b00bed205ef023cd0d811
-  data.tar.gz: 8862de3262a47dcce8ed470f3e03ddcd1059a452b11556977dcb0109fd184dce
+  metadata.gz: fbf69dee727f658d602e0343e142c5a09ddcbd5b810980d1bdb068591e5a1088
+  data.tar.gz: 239442681c1e6cc6a7a492e48ba8d07032f3bfcfe2cb59009ab1d0ca8604aa6d
 SHA512:
-  metadata.gz: b420128080ca86db35e08fdc9a72227016c16673118e3c73508388f9c223686066dc585ff6676dcde4633e0f4b13a0e0a194079c50dc7d57a780a7262773923f
-  data.tar.gz: 5950173a0b12a60272f5db537577f2ccc14c5f01d9b7921f38752d0089582245fcfafada97ca13d5ad95dc73d5ff1753a3f7a83a96218a40b9b95f528a64f417
+  metadata.gz: d88d861cc3982416bb1066e24d497cc22d9f3ee4498d35853745d4bdbc7e38d8f3aeb9a267aa6669d62b26d3a266a2598dcaec0c9a58921b5a9b3e3f9bde0b4e
+  data.tar.gz: 45ed813fd422636c8ef9c1e1b5323397430a785680502956762e2fd28055361b61bdf5540171401e1684bae606e32439bd745caa870976182ebed69daafdc9e5

data/Gemfile.lock

@@ -1,7 +1,7 @@
 PATH
   remote: .
   specs:
-    easy_sheet_io (0.1.0)
+    easy_sheet_io (0.1.4)
 
 GEM
   remote: https://rubygems.org/

data/lib/easy_sheet_io/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module EasySheetIo
-  VERSION = "0.1.1"
+  VERSION = "0.2.0"
 end

data/lib/easy_sheet_io.rb

@@ -4,9 +4,8 @@ require "roo-xls"
 require "spreadsheet"
 require "rover"
 require "daru"
-require "./to_csv"
-require "./longest_line"
-
+require_relative "./to_csv"
+require_relative "./longest_line"
 require_relative "easy_sheet_io/version"
 
 module EasySheetIo
@@ -18,29 +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
 	
@@ -54,18 +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

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: easy_sheet_io
 version: !ruby/object:Gem::Version
-  version: 0.1.1
+  version: 0.2.0
 platform: ruby
 authors:
 - show-o-atakun
 autorequire:
 bindir: exe
 cert_chain: []
-date: 2022-03-17 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.