textrepo 0.4.4 → 0.5.3

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 7415dcbc8ec9483a33c0027432a60473fd0b77ad67369b25b88fb0f399edd788
-  data.tar.gz: 9467336f5d93ec578d5693005a4737f4eef55aeb11e4c0a4ec12f197e81de240
+  metadata.gz: b8a737ed78449fe9562b18fa9a1e6f5a06e8adb7e460dd5fd35ac9f9a5127e6e
+  data.tar.gz: af06e6e6e6c004253452a8c8871de26b93e800ac090c0cf3d1c446ea498bd386
 SHA512:
-  metadata.gz: f989032a0f1db66aeb1392cf9b05fd89306152d1b76907953c90ec14839d7a34465a017a2be8017536270b9a071cd3027497305978fdf59681d3547c8a68a5e5
-  data.tar.gz: cfb126c03ce9720dd2f2e90453a5d2baf33160c7786a8a1d9b5ef02b2c4a4e2531de5744d255efd7979a39e87c9e93287de3397533404143b51be1eb8bb23a58
+  metadata.gz: e433bbb36d848f98c0f75999fecd01b29e69e9767e1d11d32897e78b50394c2787087deed5718e2266d81390736435d25e48285b47087179d64411a8a56ad21a
+  data.tar.gz: 414bdb550479be8dbb7dff29569b04064cb1a40704de8fad3d36c90ffd9ffcf73b69b4a651987245a481878e27a62e8251399ede6f81b5fb4c16c173a76da4e3

data/CHANGELOG.md

@@ -7,6 +7,29 @@ and this project adheres to [Semantic Versioning](https://semver.org/).
 ## [Unreleased]
 Nothing to record here.
 
+## [0.5.3] - 2020-11-03
+### Changed
+- Fix issue #38: fix typo in code for FileSystemRepository.
+
+## [0.5.2] - 2020-11-03
+### Changed
+- Fix issue #34:
+  - fix FileSystemRepository#entries to accept "yyyymo" pattern as a
+    Timestamp pattern.
+- Fix issue #33: fix typo in the doc for FileSystemRepository.new.
+- Fix issue #31: unfriendly error message of Timestamp.parse_s.
+
+## [0.5.1] - 2020-11-02
+### Changed
+- Fix issue #28.
+  - Modify `Repository#update` to do nothing when the given text is
+    identical to the one in the repository.
+
+## [0.5.0] - 2020-11-01
+### Added
+- Add a new API `Repository#search`.
+- Add a new API `Repository#exist?`. (0.4.3)
+
 ## [0.4.0] - 2020-10-14
 ### Added
 - Released to rubygems.org.

data/lib/textrepo/error.rb

@@ -1,12 +1,39 @@
 module Textrepo
+
+  ##
+  # Following errors might occur in repository operations:
+  #   +---------------------------------+-----------------------+
+  #   | operation (args)                | error type            |
+  #   +---------------------------------+-----------------------+
+  #   | create (timestamp, text)        | Duplicate timestamp   |
+  #   |                                 | Empty text            |
+  #   +---------------------------------+-----------------------+
+  #   | read   (timestamp)              | Missing timestamp     |
+  #   +---------------------------------+-----------------------+
+  #   | update (timestamp, text)        | Mssing timestamp      |
+  #   |                                 | Empty text            |
+  #   +---------------------------------+-----------------------+
+  #   | delete (timestamp)              | Missing timestamp     |
+  #   +---------------------------------+-----------------------+
+  #   | search (pattern, stamp_pattern) | Invalid search result |
+  #   +---------------------------------+-----------------------+
+
   class Error < StandardError; end
 
+  # :stopdoc:
   module ErrMsg
     UNKNOWN_REPO_TYPE   = 'unknown type for repository: %s'
     DUPLICATE_TIMESTAMP = 'duplicate timestamp: %s'
     EMPTY_TEXT          = 'empty text'
     MISSING_TIMESTAMP   = 'missing timestamp: %s'
+    INVALID_TIMESTAMP_STRING = "invalid string as timestamp: %s"
+    INVALID_SEARCH_RESULT = "invalid result by searcher: %s"
   end
+  # :startdoc:
+
+  ##
+  # An error raised if unknown type was specified as the repository
+  # type.
 
   class UnknownRepoTypeError < Error
     def initialize(type)
@@ -14,20 +41,9 @@ module Textrepo
     end
   end
 
-  # Following errors might occur in repository operations:
-  # +--------------------------+---------------------+
-  # | operation (args)         | error type          |
-  # +--------------------------+---------------------+
-  # | create (timestamp, text) | Duplicate timestamp |
-  # |                          | Empty text          |
-  # +--------------------------+---------------------+
-  # | read   (timestamp)       | Missing timestamp   |
-  # +--------------------------+---------------------+
-  # | update (timestamp, text) | Mssing timestamp    |
-  # |                          | Empty text          |
-  # +--------------------------+---------------------+
-  # | delete (timestamp)       | Missing timestamp   |
-  # +--------------------------+---------------------+
+  ##
+  # An error raised if the specified timestamp has already exist in
+  # the repository.
 
   class DuplicateTimestampError < Error
     def initialize(timestamp)
@@ -35,16 +51,43 @@ module Textrepo
     end
   end
 
+  ##
+  # An error raised if the given text is empty.
+
   class EmptyTextError < Error
     def initialize
       super(ErrMsg::EMPTY_TEXT)
     end
   end
 
+  ##
+  # An error raised if the given timestamp has not exist in the
+  # repository.
+
   class MissingTimestampError < Error
     def initialize(timestamp)
       super(ErrMsg::MISSING_TIMESTAMP % timestamp)
     end
   end
 
+  ##
+  # An error raised if an argument is invalid to convert a
+  # Textrepo::Timestamp object.
+
+  class InvalidTimestampStringError < Error
+    def initialize(str)
+      super(ErrMsg::INVALID_TIMESTAMP_STRING % str)
+    end
+  end
+
+  ##
+  # An error raise if the search result is not suitable to use.
+  #
+
+  class InvalidSearchResultError < Error
+    def initialize(str)
+      super(ErrMsg::INVALID_SEARCH_RESULT % str)
+    end
+  end
+
 end

data/lib/textrepo/file_system_repository.rb

@@ -1,4 +1,5 @@
 require 'fileutils'
+require "open3"
 
 module Textrepo
 
@@ -19,6 +20,16 @@ module Textrepo
 
     attr_reader :extname
 
+    ##
+    # Searcher program name.
+
+    attr_reader :searcher
+
+    ##
+    # An array of options to pass to the searcher program.
+
+    attr_reader :searcher_options
+
     ##
     # Default name for the repository which uses when no name is
     # specified in the configuration settings.
@@ -31,6 +42,11 @@ module Textrepo
 
     FAVORITE_EXTNAME = 'md'
 
+    ##
+    # Default searcher program to search text in the repository.
+
+    FAVORITE_SEARCHER = 'grep'
+
     ##
     # Creates a new repository object.  The argument, `conf` must be a
     # Hash object.  It should hold the follwoing values:
@@ -41,15 +57,33 @@ module Textrepo
     # - OPTIONAL: (if not specified, default values are used)
     #   - :repository_name => basename of the root path for the repository
     #   - :default_extname => extname for a file stored into in the repository
+    #   - :searcher => a program to search like `grep`
+    #   - :searcher_options => an Array of option to pass to the searcher
     #
     # The root path of the repository looks like the following:
     # - conf[:repository_base]/conf[:repository_name]
     # 
-    # Default values are set when `repository_name` and `default_extname`
+    # Default values are set when `:repository_name` and `:default_extname`
     # were not defined in `conf`.
     #
+    # Be careful to set `:searcher_options`, it must be to specify the
+    # searcher behavior equivalent to `grep` with "-inRE".  The
+    # default values for the searcher options is defined for BSD grep
+    # (default grep on macOS), GNU grep, and ripgrep (aka rg).  They
+    # are:
+    #
+    #   "grep"   => ["-i", "-n", "-R", "-E"]
+    #   "egrep"  => ["-i", "-n", "-R"]
+    #   "ggrep"  => ["-i", "-n", "-R", "-E"]
+    #   "gegrep" => ["-i", "-n", "-R"]
+    #   "rg"     => ["-S", "-n", "--no-heading", "--color", "never"]
+    #
+    # If use those searchers, it is not recommended to set
+    # `:searcher_options`.  The default value works well in
+    # `textrepo`.
+    #
     # :call-seq:
-    #     new(Rbnotes::Conf or Hash) -> FileSystemRepository
+    #     new(Hash or Hash like object) -> FileSystemRepository
 
     def initialize(conf)
       super
@@ -58,6 +92,8 @@ module Textrepo
       @path = File.expand_path("#{name}", base)
       FileUtils.mkdir_p(@path)
       @extname = conf[:default_extname] || FAVORITE_EXTNAME
+      @searcher = find_searcher(conf[:searcher])
+      @searcher_options = conf[:searcher_options]
     end
 
     ##
@@ -97,21 +133,25 @@ module Textrepo
     # Updates the file content in the repository.  A new timestamp
     # will be attached to the text.
     #
+    # See the documentation of Repository#update to know about errors
+    # and constraints of this method.
+    #
     # :call-seq:
     #     update(Timestamp, Array) -> Timestamp
 
     def update(timestamp, text)
       raise EmptyTextError if text.empty?
-      org_abs = abspath(timestamp)
-      raise MissingTimestampError, timestamp unless FileTest.exist?(org_abs)
+      raise MissingTimestampError, timestamp unless exist?(timestamp)
+
+      # does nothing if given text is the same in the repository one
+      return timestamp if read(timestamp) == text
 
       # the text must be stored with the new timestamp
       new_stamp = Timestamp.new(Time.now)
-      new_abs = abspath(new_stamp)
-      write_text(new_abs, text)
+      write_text(abspath(new_stamp), text)
 
-      # delete the original file in the repository
-      FileUtils.remove_file(org_abs)
+      # delete the original text file in the repository
+      FileUtils.remove_file(abspath(timestamp))
 
       new_stamp
     end
@@ -147,7 +187,7 @@ module Textrepo
         if exist?(stamp)
           results << stamp
         end
-      when 0, "yyyymoddhhmiss".size, "yyyymodd".size
+      when 0, "yyyymoddhhmiss".size, "yyyymodd".size, "yyyymo".size
         results += find_entries(stamp_pattern)
       when 4                    # "yyyy" or "modd"
         pat = nil
@@ -179,6 +219,27 @@ module Textrepo
       FileTest.exist?(abspath(timestamp))
     end
 
+    ##
+    # Searches a pattern in all text.  The given pattern is a word to
+    # search or a regular expression.  The pattern would be passed to
+    # a searcher program as it passed.
+    #
+    # See the document for Textrepo::Repository#search to know about
+    # the search result.
+    #
+    # :call-seq:
+    #     search(String for pattern, String for Timestamp pattern) -> Array
+
+    def search(pattern, stamp_pattern = nil)
+      result = nil
+      if stamp_pattern.nil?
+        result = invoke_searcher_at_repo_root(@searcher, pattern)
+      else
+        result = invoke_searcher_for_entries(@searcher, pattern, entries(stamp_pattern))
+      end
+      construct_search_result(result)
+    end
+
     # :stopdoc:
 
     private
@@ -208,9 +269,154 @@ module Textrepo
 
     def find_entries(stamp_pattern)
       Dir.glob("#{@path}/**/#{stamp_pattern}*.#{@extname}").map { |e|
-        Timestamp.parse_s(timestamp_str(e))
+        begin
+          Timestamp.parse_s(timestamp_str(e))
+        rescue InvalidTimestampStringError => _
+          # Just ignore the erroneous entry, since it is not a text in
+          # the repository.  It may be a garbage, or some kind of
+          # hidden stuff of the repository, ... etc.
+          nil
+        end
+      }.compact
+    end
+
+    ##
+    # The upper limit of files to search at one time.  The value has
+    # no reason to select.  It seems to me that not too much, not too
+    # little to handle in one process to search.
+
+    LIMIT_OF_FILES = 20
+
+    ##
+    # When no timestamp pattern was given, invoke the searcher with
+    # the repository root path as its argument and the recursive
+    # searching option.  The search could be done in only one process.
+
+    def invoke_searcher_at_repo_root(searcher, pattern)
+      o, s = Open3.capture2(searcher, *find_searcher_options(searcher),
+                            pattern, @path)
+      output = []
+      output += o.lines.map(&:chomp) if s.success? && (! o.empty?)
+      output
+    end
+
+    ##
+    # When a timestamp pattern was given, at first, list target files,
+    # then invoke the searcher for those files.  Since the number of
+    # target files may be so much, it seems to be dangerous to pass
+    # all of them to a single search process at one time.
+    #
+    # One more thing to mention, the searcher, like `grep`, does not
+    # add the filename at the beginning of the search result line, if
+    # the target is one file.  This behavior is not suitable in this
+    # purpose.  The code below adds the filename when the target is
+    # one file.
+
+    def invoke_searcher_for_entries(searcher, pattern, entries)
+      output = []
+
+      num_of_entries = entries.size
+      if num_of_entries == 1
+        # If the search taget is one file, the output needs special
+        # treatment.
+        file = abspath(entries[0])
+        o, s = Open3.capture2(searcher, *find_searcher_options(searcher),
+                              pattern, file)
+        if s.success? && (! o.empty?)
+          output += o.lines.map { |line|
+            # add filename at the beginning of the search result line
+            [file, line.chomp].join(":")
+          }
+        end
+      elsif num_of_entries > LIMIT_OF_FILES
+        output += invoke_searcher_for_entries(searcher, pattern, entries[0..(LIMIT_OF_FILES - 1)])
+        output += invoke_searcher_for_entries(searcher, pattern, entries[LIMIT_OF_FILES..-1])
+      else
+        # When the number of target is less than the upper limit,
+        # invoke the searcher with all of target files as its
+        # arguments.
+        files = find_files(entries)
+        o, s = Open3.capture2(searcher, *find_searcher_options(searcher),
+                              pattern, *files)
+        if s.success? && (! o.empty?)
+          output += o.lines.map(&:chomp)
+        end
+      end
+
+      output
+    end
+
+    SEARCHER_OPTS = {
+      # case insensitive, print line number, recursive search, work as egrep
+      "grep"   => ["-i", "-n", "-R", "-E"],
+      # case insensitive, print line number, recursive search
+      "egrep"  => ["-i", "-n", "-R"],
+      # case insensitive, print line number, recursive search, work as gegrep
+      "ggrep"  => ["-i", "-n", "-R", "-E"],
+      # case insensitive, print line number, recursive search
+      "gegrep" => ["-i", "-n", "-R"],
+      # smart case, print line number, no color
+      "rg"     => ["-S", "-n", "--no-heading", "--color", "never"],
+    }
+
+    def find_searcher_options(searcher)
+      @searcher_options || SEARCHER_OPTS[File.basename(searcher)] || ""
+    end
+
+    def find_files(timestamps)
+      timestamps.map{|stamp| abspath(stamp)}
+    end
+
+    ##
+    # The argument must be an Array contains the searcher output.
+    # Each item is constructed from 3 parts:
+    #   "<pathname>:<integer>:<text>"
+    #
+    # For example, it may looks like:
+    #
+    #   "/somewhere/2020/11/20201101044300.md:18:foo is foo"
+    #
+    # Or it may contains more ":" in the text part as:
+    #
+    #   "/somewhere/2020/11/20201101044500.md:119:apple:orange:grape"
+    #
+    # In the latter case, `split(":")` will split it too much.  That is,
+    # the result will be:
+    #
+    #  ["/somewhere/2020/11/20201101044500.md", "119", "apple", "orange", "grape"]
+    #
+    # Text part must be joined with ":".
+
+    def construct_search_result(output)
+      output.map { |line|
+        begin
+          pathname, num, *match_text = line.split(":")
+          [Timestamp.parse_s(timestamp_str(pathname)),
+           num.to_i,
+           match_text.join(":")]
+        rescue InvalidTimestampStringError, TypeError => _
+          raise InvalidSearchResultError, [@searcher, @searcher_options.join(" ")].join(" ")
+        end
+      }.compact
+    end
+
+    def find_searcher(program = nil)
+      candidates = [FAVORITE_SEARCHER]
+      candidates.unshift(program) unless program.nil? || candidates.include?(program)
+      search_paths = ENV["PATH"].split(":")
+      candidates.map { |prog|
+        find_in_paths(prog, search_paths)
+      }[0]
+    end
+
+    def find_in_paths(prog, paths)
+      paths.each { |p|
+        abspath = File.expand_path(prog, p)
+        return abspath if FileTest.exist?(abspath) && FileTest.executable?(abspath)
       }
+      nil
     end
+    # :startdoc:
 
   end
 end

data/lib/textrepo/repository.rb

@@ -43,8 +43,17 @@ module Textrepo
     def read(timestamp); []; end
 
     ##
-    # Updates the content with text in the repository, which is
-    # associated to the timestamp.  Returns the timestamp.
+    # Updates the content with given text in the repository, which is
+    # associated to the given timestamp.  Returns the timestamp newly
+    # generated during the execution.
+    #
+    # If the given Timestamp is not existed as a Timestamp attached to
+    # text in the repository, raises MissingTimestampError.
+    #
+    # If the given text is empty, raises EmptyTextError.
+    #
+    # If the given text is identical to the text in the repository,
+    # does nothing.  Returns the given timestamp itself.
     #
     # :call-seq:
     #     update(Timestamp, Array) -> Timestamp
@@ -92,6 +101,25 @@ module Textrepo
     #     exist?(Timestamp) -> true or false
 
     def exist?(timestamp); false; end
+
+    ##
+    # Searches a pattern (word or regular expression) in text those
+    # matches to a given timestamp pattern.  Returns an Array of
+    # search results.  If no match, returns an empty Array.
+    #
+    # See the document for Repository#entries about a timestamp
+    # pattern.  When nil is passed as a timestamp pattern, searching
+    # applies to all text in the repository.
+    #
+    # Each entry of the result Array is constructed from 3 items, (1)
+    # timestamp (Timestamp), (2) line number (Integer), (3) matched
+    # line (String).
+    #
+    # :call-seq:
+    #     search(String for pattern, String for Timestamp pattern) -> Array
+
+    def search(pattern, stamp_pattern = nil); []; end
+
   end
 
   require_relative 'file_system_repository'

data/lib/textrepo/timestamp.rb

@@ -72,8 +72,11 @@ module Textrepo
       #    yyyymoddhhmiss sfx      yyyy    mo    dd    hh    mi    ss    sfx
       #   "20201230123456"     -> "2020", "12", "30", "12", "34", "56"
       #   "20201230123456_789" -> "2020", "12", "30", "12", "34", "56", "789"
+      #
+      # Raises InvalidTimestampStringError if nil was passed as an arguemnt.
 
       def split_stamp(stamp_str)
+        raise InvalidTimestampStringError, stamp_str if stamp_str.nil?
         #    yyyy  mo    dd    hh    mi      ss      sfx
         a = [0..3, 4..5, 6..7, 8..9, 10..11, 12..13, 15..17].map {|r| stamp_str[r]}
         a[-1].nil? ? a[0..-2] : a
@@ -83,13 +86,27 @@ module Textrepo
       # Generate a Timestamp object from a string which represents a
       # timestamp, such "20201028163400".
       #
+      # Raises InvalidTimestampStringError if cannot convert the
+      # argument into a Timestamp object.
+      #
       # :call-seq:
       #     parse_s("20201028163400") -> Timestamp
       #     parse_s("20201028163529_034") -> Timestamp
 
       def parse_s(stamp_str)
-        year, mon, day, hour, min, sec , sfx = split_stamp(stamp_str).map(&:to_i)
-        Timestamp.new(Time.new(year, mon, day, hour, min, sec), sfx)
+        begin
+          ye, mo, da, ho, mi, se, sfx = split_stamp(stamp_str).map(&:to_i)
+          Timestamp.new(Time.new(ye, mo, da, ho, mi, se), sfx)
+        rescue InvalidTimestampStringError, ArgumentError => _
+          emsg = if stamp_str.nil?
+            "(nil)"
+          elsif stamp_str.empty?
+            "(empty string)"
+          else
+            stamp_str
+          end
+          raise InvalidTimestampStringError, emsg
+        end
       end
 
     end

data/lib/textrepo/version.rb

@@ -1,3 +1,3 @@
 module Textrepo
-  VERSION = '0.4.4'
+  VERSION = '0.5.3'
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: textrepo
 version: !ruby/object:Gem::Version
-  version: 0.4.4
+  version: 0.5.3
 platform: ruby
 authors:
 - mnbi
 autorequire:
 bindir: exe
 cert_chain: []
-date: 2020-10-28 00:00:00.000000000 Z
+date: 2020-11-03 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: bundler