Class: SportDb::Import::LeagueReader

Inherits:
Object
  • Object
show all
Includes:
LogUtils::Logging
Defined in:
lib/sportdb/helpers/league_reader.rb

Class Method Summary collapse

Instance Method Summary collapse

Constructor Details

#initialize(txt) ⇒ LeagueReader

Returns a new instance of LeagueReader.



22
23
24
 
# File 'lib/sportdb/helpers/league_reader.rb', line 22

def initialize( txt )
  @txt = txt
end

Class Method Details

.parse(txt) ⇒ Object 



14
15
16
 
# File 'lib/sportdb/helpers/league_reader.rb', line 14

def self.parse( txt )
  new( txt ).parse
end

.read(path) ⇒ Object

use - rename to read_file or from_file etc. - why? why not?



9
10
11
12
 
# File 'lib/sportdb/helpers/league_reader.rb', line 9

def self.read( path )   ## use - rename to read_file or from_file etc. - why? why not?
  txt = File.open( path, 'r:utf-8' ) { |f| f.read }
  parse( txt )
end

Instance Method Details

#_norm(str) ⇒ Object

norm(alize) helper - squish (spaces)

and remove dollars ($$$)
and remove leading and trailing spaces



137
138
139
140
 
# File 'lib/sportdb/helpers/league_reader.rb', line 137

def _norm( str )
  ## only extra clean-up of dollars for now ($$$)
  _squish( str.gsub( '$', '' ) )
end

#_squish(str) ⇒ Object 



142
143
144
 
# File 'lib/sportdb/helpers/league_reader.rb', line 142

def _squish( str )
  str.gsub( /[ \t\u00a0]+/, ' ' ).strip
end

#parse ⇒ Object 



26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
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
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
 
# File 'lib/sportdb/helpers/league_reader.rb', line 26

def parse
  recs = []
  last_rec = nil

  country  = nil    # last country
  intl     = false  # is international (league/tournament/cup/competition)
  clubs    = true   # or clubs|national teams

  OutlineReader.parse( @txt ).each do |node|
    if [:h1,:h2,:h3,:h4,:h5,:h6].include?( node[0] )
      heading_level  = node[0][1].to_i
      heading        = node[1]

      logger.debug "heading #{heading_level} >#{heading}<"

      if heading_level != 1
        puts "** !!! ERROR !!! unsupported heading level; expected heading 1 for now only; sorry"
        pp line
        exit 1
      else
        logger.debug "heading (#{heading_level}) >#{heading}<"
        last_heading = heading
        ## map to country or international / int'l or national teams
        if heading =~ /national team/i   ## national team tournament
          country = nil
          intl    = true
          clubs   = false
        elsif heading =~ /international|int'l/i  ## int'l club tournament
          country = nil
          intl    = true
          clubs   = true
        else
          ## assume country in heading; allow all "formats" supported by parse e.g.
          ##   Österreich • Austria (at)
          ##   Österreich • Austria
          ##   Austria
          ##   Deutschland (de) • Germany
          country = Country.parse_heading( heading )
          intl    = false
          clubs   = true

          ## check country code - MUST exist for now!!!!
          if country.nil?
            puts "!!! error [league reader] - unknown country >#{heading}< - sorry - add country to config to fix"
            exit 1
          end
        end
      end
    elsif node[0] == :p   ## paragraph with (text) lines
      lines = node[1]
      lines.each do |line|

      if line.start_with?( '|' )
          ## assume continuation with line of alternative names
          ##  note: skip leading pipe
          values = line[1..-1].split( '|' )   # team names - allow/use pipe(|)
          values = values.map {|value| _norm(value) }  ## squish/strip etc.

          logger.debug "alt_names: #{values.join( '|' )}"

          last_rec.alt_names += values
      else
        ## assume "regular" line
        ##  check if starts with id  (todo/check: use a more "strict"/better regex capture pattern!!!)
        if line =~ /^([a-z0-9][a-z0-9.]*)[ ]+(.+)$/
          league_key  = $1
          ## 1) strip (commercial) sponsor markers/tags e.g $$
          ## 2) strip and squish (white)spaces
          league_name = _norm( $2 )

          logger.debug "key: >#{league_key}<, name: >#{league_name}<"


          ## prepend country key/code if country present
          ##   todo/fix: only auto-prepend country if key/code start with a number (level) or incl. cup
          ##    why? lets you "overwrite" key if desired - use it - why? why not?
          if country
            league_key = "#{country.key}.#{league_key}"
          end

          rec = League.new( key:            league_key,
                            name:           league_name,
                            country:        country,
                            intl:           intl,
                            clubs:          clubs)
          recs << rec
          last_rec = rec
        else
          puts "** !!! ERROR !!! missing key for (canonical) league name"
          exit 1
        end
      end
      end  # each line
    else
      puts "** !!! ERROR !!! [league reader] - unknown line type:"
      pp node
      exit 1
    end
    ## pp line
  end
  recs
end