sportdb-readers 0.2.3 → 0.3.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: 7a80a58bb3ddb33bdba8040a926546b548ba5436
-  data.tar.gz: 608847551d5a437ab847e91aa79ea82626d010cf
+  metadata.gz: 6d2044ec2c7a67046c429f4f52fa78a02d64f42b
+  data.tar.gz: 4e4e0102a368e5f9b53d8d40681470017fbc8579
 SHA512:
-  metadata.gz: fc388f76b6b1ec3ca85ea4bea73ace80f13af88ef05eb7601b04176aeca742c35e828ee1fa6d5ca674d6a71789007f5b785185e8b241db4ee14e89218ff624ca
-  data.tar.gz: aa7ed9979898a56656f16d9093687b3d4c7a1590add76d4d5a866966e5bd2169e545fabcbc0fdc11f7c79b2ac8677667c9ca52d7a57e0dfb23963becbaa2d71f
+  metadata.gz: bf9db195f63e40f376a5bf9a63a7c73c79a6fcaf8dc42bbcc2fd6e18813ec275ccdab4ba5209ff52adba42b98ca7c211e2b172ac844fe2c721c349ceaf3685d7
+  data.tar.gz: afe05f74d798eb1a0ea04d0d2ed207f858f1d3135a7bfd27eeefadbc55dc8afd8d7eb290dd2652f01394f4a9af335bbc83a7b7f715467e681b2618a86b8c3fca

data/Manifest.txt

@@ -5,6 +5,7 @@ Rakefile
 lib/sportdb/readers.rb
 lib/sportdb/readers/conf_linter.rb
 lib/sportdb/readers/conf_reader.rb
+lib/sportdb/readers/datafile.rb
 lib/sportdb/readers/league_outline_reader.rb
 lib/sportdb/readers/match_linter.rb
 lib/sportdb/readers/match_parser.rb

data/lib/sportdb/readers.rb

@@ -1,5 +1,8 @@
 # encoding: utf-8
 
+require 'zip'     ## todo/check: if zip is alreay included in a required module?
+
+
 
 require 'sportdb/config'
 require 'sportdb/models'   ## add sql database support
@@ -15,6 +18,7 @@ require 'sportdb/readers/conf_linter'
 require 'sportdb/readers/match_parser'
 require 'sportdb/readers/match_reader'
 require 'sportdb/readers/match_linter'
+require 'sportdb/readers/datafile'
 
 
 
@@ -22,42 +26,55 @@ require 'sportdb/readers/match_linter'
 ##  add convenience shortcut helpers
 module SportDb
 
-  def self.read_conf( path, sync: true, season: nil )    ## note: sync is dry run (for lint checking)
+  ## note: sync is dry run (for lint checking)
+  def self.read_conf( path, season: nil, sync: true )
     sync ? ConfReaderV2.read( path, season: season )
          : ConfLinter.read( path, season: season )
   end
 
-  ### todo: add alias read_matches - why? why not?
-  def self.read_match( path, sync: true, season: nil )     ## note: sync is dry run (for lint checking)
+  def self.read_match( path, season: nil, sync: true )  ### todo/check: add alias read_matches - why? why not?
     sync ? MatchReaderV2.read( path, season: season )
          : MatchLinter.read( path, season: season )
   end
 
-  def self.lint_conf( path, season: nil )   read_conf( path, sync: false, season: season ); end
-  def self.lint_match( path, season: nil )  read_match( path, sync: false, season: season ); end
 
+  def self.parse_conf( txt, season: nil, sync: true )
+    sync ? ConfReaderV2.parse( txt, season: season )
+         : ConfLinter.parse( txt, season: season )
+  end
 
-  def self.read( path, sync: true, season: nil )
-    ## step 1: collect all datafiles
-    if File.directory?( path )   ## if directory read complete package
-      datafiles_conf = Datafile.find_conf( path )
-      datafiles      = Datafile.find( path, %r{/\d{4}-\d{2}    ## season folder e.g. /2019-20
-                                               /[a-z0-9_-]+\.txt$    ## txt e.g /1-premierleague.txt
-                                              }x )
+  def self.parse_match( txt, season: nil, sync: true )  ### todo/check: add alias read_matches - why? why not?
+    sync ? MatchReaderV2.parse( txt, season: season )
+         : MatchLinter.parse( txt, season: season )
+  end
 
-      datafiles_conf.each { |datafile| read_conf( datafile, sync: sync, season: season ) }
-      datafiles.each { |datafile| read_match( datafile, sync: sync, season: season ) }
+
+  def self.read( path, season: nil, sync: true )
+    pack = if File.directory?( path )          ## if directory assume "unzipped" package
+              Datafile::DirPackage.new( path )
+           elsif Datafile.match_zip( path )    ## check if file is a .zip (archive) file
+              Datafile::ZipPackage.new( path )
+           else                                ## no package; assume single (standalone) datafile
+             nil
+           end
+
+    if pack
+       pack.read_conf( season: season, sync: sync )
+       pack.read_match( season: season, sync: sync )
     else
-      ## check if datafile matches conf(iguration) naming (e.g. .conf.txt)
-      if Datafile.match_conf( path )
-        read_conf( path, sync: sync, season: season )
-      else   ## assume "regular" match datafile
-        read_match( path, sync: sync, season: season )
+      if Datafile.match_conf( path )      ## check if datafile matches conf(iguration) naming (e.g. .conf.txt)
+        read_conf( path, season: season, sync: sync )
+      else                                ## assume "regular" match datafile
+        read_match( path, season: season, sync: sync )
       end
     end
   end  # method read
 
-  def self.lint( path, season: nil )  read( path, sync: false, season: season ); end
+
+  ## (more) convenience helpers for lint(ing)
+  def self.lint( path, season: nil )        read( path, season: season, sync: false ); end
+  def self.lint_conf( path, season: nil )   read_conf( path, season: season, sync: false ); end
+  def self.lint_match( path, season: nil )  read_match( path, season: season, sync: false ); end
 
 end # module SportDb
 

data/lib/sportdb/readers/datafile.rb

@@ -0,0 +1,116 @@
+
+
+module Datafile
+
+
+  ## todo/fix: move regex patterns upstream to datafile in sportdb-formats!!
+
+  CONF_RE  = %r{ /\.conf\.txt$
+               }x
+
+  MATCH_RE = %r{ /\d{4}-\d{2}        ## season folder e.g. /2019-20
+                 /[a-z0-9_-]+\.txt$  ## txt e.g /1-premierleague.txt
+              }x
+
+  ZIP_RE = %r{ \.zip$
+            }x
+  def self.match_zip( path, pattern: ZIP_RE ) pattern.match( path ); end
+
+
+
+class PackageBase
+
+  ## note: "abstract" methods - each and read required in derived class !!!!
+
+  def each_conf( &blk )   each( pattern: CONF_RE, &blk ); end
+  def each_match( &blk )  each( pattern: MATCH_RE, &blk ); end
+
+  def read_conf( season:, sync: )
+    read( pattern: CONF_RE ) do |name,txt|
+      SportDb.parse_conf( season: season, sync: sync )
+    end
+  end
+  def read_match( season:, sync: )
+    read( pattern: MATCH_RE ) do |name, txt|
+      SportDb.parse_match( season: season, sync: sync )
+    end
+  end
+end   # class PackageBase
+
+
+
+class DirPackage < PackageBase    ## todo/check: find a better name e.g. UnzippedPackage, FilesystemPackage, etc. - why? why not?
+
+  def initialize( path )
+    @path = path   ## rename to root_path or base_path or somehting - why? why not?
+  end
+
+
+  def each_file( pattern: )    ## todo/check: rename to glob or something - why? why not?
+    ## note: incl. files starting with dot (.)) as candidates (normally excluded with just *)
+    Dir.glob( "#{@path}/**/{*,.*}.txt" ).each do |path|
+      ## todo/fix: (auto) skip and check for directories
+      if pattern.match( path )
+        yield( path)
+      else
+        ## puts "  skipping >#{path}<"
+      end
+    end
+  end
+
+
+  Entry = Struct.new( :name )
+
+  def each( pattern: )   ## todo/check: rename to each_entry - why? why not?
+    each_file( pattern: pattern ) do |path|
+      ## fix: split path like a "virtual" zip like entry
+      yield( Entry.new( path ) )
+    end
+  end
+
+  def read( pattern: )
+    each_file( pattern: pattern ) do |path|
+      txt = File.open( path, 'r:utf-8').read
+      yield( path, txt )  ## only pass along txt - why? why not? or pass along entry and not just entry.name?
+    end
+  end
+end  # class DirPackage
+
+
+
+## helper wrapper for datafiles in zips
+class ZipPackage < PackageBase
+  def initialize( path )
+    @path = path
+  end
+
+
+  def each( pattern: )
+    Zip::File.open( @path ) do |zipfile|
+      zipfile.each do |entry|
+        if entry.directory?
+          next ## skip
+        elsif entry.file?
+          if pattern.match( entry.name )
+            yield( entry )
+          else
+            ## puts "  skipping >#{entry.name}<"
+          end
+        else
+          puts "** !! ERROR !! #{entry.name} is unknown zip file type, sorry"
+          exit 1
+        end
+      end
+    end
+  end
+
+  def read( pattern: )
+    each( pattern: pattern ) do |entry|
+      txt = entry.get_input_stream.read
+      ##  puts "** encoding: #{txt.encoding}"  #=> encoding: ASCII-8BIT
+      txt = txt.force_encoding( Encoding::UTF_8 )
+      yield( "#{@path}!/#{entry.name}", txt )    ## only pass along txt - why? why not? or pass along entry and not just entry.name?
+    end
+  end
+end  # class ZipPackage
+end  # module Datafile

data/lib/sportdb/readers/version.rb

@@ -5,8 +5,8 @@ module SportDb
 module Readers
 
   MAJOR = 0    ## todo: namespace inside version or something - why? why not??
-  MINOR = 2
-  PATCH = 3
+  MINOR = 3
+  PATCH = 0
   VERSION = [MAJOR,MINOR,PATCH].join('.')
 
   def self.version

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: sportdb-readers
 version: !ruby/object:Gem::Version
-  version: 0.2.3
+  version: 0.3.0
 platform: ruby
 authors:
 - Gerald Bauer
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2019-11-14 00:00:00.000000000 Z
+date: 2019-11-16 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: sportdb-config
@@ -97,6 +97,7 @@ files:
 - lib/sportdb/readers.rb
 - lib/sportdb/readers/conf_linter.rb
 - lib/sportdb/readers/conf_reader.rb
+- lib/sportdb/readers/datafile.rb
 - lib/sportdb/readers/league_outline_reader.rb
 - lib/sportdb/readers/match_linter.rb
 - lib/sportdb/readers/match_parser.rb