Class: Opencdd::Reader

Inherits:
Object
  • Object
show all
Defined in:
lib/opencdd/reader.rb

Constant Summary collapse

XLSX_EXTENSIONS = 
%w[.xlsx .xlsm .xltx].freeze
SHARDED_CLASS_CODE =

Layout detection is owned by Opencdd::Parcel::LayoutDetector (single source of truth). These regexes are exposed here as backward-compat constants for external callers; new code should call LayoutDetector directly.

Opencdd::Parcel::LayoutDetector::CLASS_CODE_PATTERN
UNID_PATTERN = 
Opencdd::Parcel::LayoutDetector::UNID_PATTERN
LEGACY_XLS_RE = 
Opencdd::Parcel::LayoutDetector::FILE_PATTERN

Instance Attribute Summary collapse

Class Method Summary collapse

Instance Method Summary collapse

Constructor Details

#initialize(path) ⇒ Reader

Returns a new instance of Reader.



79
80
81
 
# File 'lib/opencdd/reader.rb', line 79

def initialize(path)
  @path = path
end

Instance Attribute Details

#path ⇒ Object (readonly)

Returns the value of attribute path.



77
78
79
 
# File 'lib/opencdd/reader.rb', line 77

def path
  @path
end

Class Method Details

.detect(path) ⇒ Object 



20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
 
# File 'lib/opencdd/reader.rb', line 20

def detect(path)
  case File.basename(path)
  when LEGACY_XLS_RE then :legacy_single
  else
    if File.directory?(path)
      children = Dir.children(path)
      if children.any? { |f| f =~ LEGACY_XLS_RE }
        :legacy_dir
      elsif children.any? { |f| sharded_class_subdir?(File.join(path, f)) }
        :sharded_dir
      else
        :unknown_dir
      end
    elsif XLSX_EXTENSIONS.include?(File.extname(path).downcase)
      :xlsx
    elsif File.extname(path).downcase == ".xls"
      :legacy_single
    else
      :unknown
    end
  end
end

.load_database(path) ⇒ Object 



16
17
18
 
# File 'lib/opencdd/reader.rb', line 16

def load_database(path)
  new(path).load
end

.sharded_class_subdir?(path) ⇒ Boolean

A class-code subdir is "sharded" if it contains export_.xls directly (legacy flat layout) OR exposes the per-version layout (+entity.json+ pointing at a UNID subfolder, or a single UNID subfolder holding export.xls).

Returns:

  • (Boolean) 


48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
 
# File 'lib/opencdd/reader.rb', line 48

def self.sharded_class_subdir?(path)
  return false unless File.directory?(path)
  return false unless File.basename(path) =~ SHARDED_CLASS_CODE

  children = Dir.children(path)
  return true if children.any? { |f| f =~ LEGACY_XLS_RE }

  # Per-version layout: _entity.json names a UNID subfolder.
  entity_idx = File.join(path, "_entity.json")
  if File.file?(entity_idx)
    require "json"
    begin
      data = JSON.parse(File.read(entity_idx))
      current = data["current_version_dir"]
      return true if current && current =~ UNID_PATTERN &&
                     File.directory?(File.join(path, current))
    rescue JSON::ParserError
      # fall through
    end
  end

  # Per-version layout without _entity.json: a UNID subfolder holds XLS.
  unid_subdirs = children.select { |n| n =~ UNID_PATTERN }
                          .map { |n| File.join(path, n) }
  unid_subdirs.any? do |p|
    File.directory?(p) && Dir.children(p).any? { |f| f =~ LEGACY_XLS_RE }
  end
end

Instance Method Details

#load ⇒ Object 



83
84
85
86
87
88
 
# File 'lib/opencdd/reader.rb', line 83

def load
  db = Opencdd::Database.new
  load_into(db)
  db.finalize!
  db
end

#load_into(db) ⇒ Object 



90
91
92
93
94
95
96
97
98
99
100
101
 
# File 'lib/opencdd/reader.rb', line 90

def load_into(db)
  case Opencdd::Reader.detect(@path)
  when :xlsx, :legacy_single
    Opencdd::Parcel::WorkbookReader.new(@path).load_into(db)
  when :legacy_dir
    Opencdd::Parcel::FlatDirReader.new(@path).load_into(db)
  when :sharded_dir
    Opencdd::Parcel::ShardedDirReader.new(@path).load_into(db)
  else
    raise ArgumentError, "Cannot detect Parcel/Excel format at #{@path.inspect}"
  end
end