Class: Wp2txt::IndexCache

Inherits:
Object
  • Object
show all
Defined in:
lib/wp2txt/index_cache.rb

Overview

SQLite-based cache for multistream index Dramatically speeds up repeated index access by storing parsed entries in SQLite

Constant Summary collapse

CACHE_VERSION = 
1
CACHE_SUFFIX = 
".sqlite3"

Instance Attribute Summary collapse

Instance Method Summary collapse

Constructor Details

#initialize(source_path, cache_dir: nil) ⇒ IndexCache

Returns a new instance of IndexCache.



16
17
18
19
20
21
 
# File 'lib/wp2txt/index_cache.rb', line 16

def initialize(source_path, cache_dir: nil)
  @source_path = source_path
  @cache_dir = cache_dir || default_cache_dir
  @cache_path = build_cache_path
  @db = nil
end

Instance Attribute Details

#cache_path ⇒ Object (readonly)

Returns the value of attribute cache_path.



14
15
16
 
# File 'lib/wp2txt/index_cache.rb', line 14

def cache_path
  @cache_path
end

#source_path ⇒ Object (readonly)

Returns the value of attribute source_path.



14
15
16
 
# File 'lib/wp2txt/index_cache.rb', line 14

def source_path
  @source_path
end

Instance Method Details

#clear! ⇒ Object

Delete cache file



183
184
185
 
# File 'lib/wp2txt/index_cache.rb', line 183

def clear!
  FileUtils.rm_f(@cache_path)
end

#find_by_titles(titles) ⇒ Hash

Find entries by titles (efficient batch lookup)

Parameters:

  • titles (Array<String>) —

    titles to look up

Returns:

  • (Hash) —

    title => entry or nil



135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
 
# File 'lib/wp2txt/index_cache.rb', line 135

def find_by_titles(titles)
  return {} if titles.empty?
  return {} unless valid?

  results = {}
  open_db
  begin
    # Use IN clause with placeholders for batch lookup
    placeholders = titles.map { "?" }.join(",")
    sql = "SELECT title, page_id, byte_offset FROM index_entries WHERE title IN (#{placeholders})"

    # SQLite3 2.x requires bind variables as an array
    @db.execute(sql, titles) do |row|
      title, page_id, offset = row
      results[title] = { offset: offset, page_id: page_id, title: title }
    end

    results
  ensure
    close_db
  end
end

#load ⇒ Hash

Load index entries from cache

Returns:

  • (Hash) —

    { entries_by_title: {}, entries_by_id: {}, stream_offsets: [] }



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
 
# File 'lib/wp2txt/index_cache.rb', line 52

def load
  return nil unless valid?

  entries_by_title = {}
  entries_by_id = {}
  stream_offsets = []

  open_db
  begin
    # Load all entries
    @db.execute("SELECT title, page_id, byte_offset FROM index_entries") do |row|
      title, page_id, offset = row
      entry = { offset: offset, page_id: page_id, title: title }
      entries_by_title[title] = entry
      entries_by_id[page_id] = entry
    end

    # Load stream offsets
    @db.execute("SELECT byte_offset FROM stream_offsets ORDER BY byte_offset") do |row|
      stream_offsets << row[0]
    end

    { entries_by_title: entries_by_title, entries_by_id: entries_by_id, stream_offsets: stream_offsets }
  ensure
    close_db
  end
end

#save(entries_by_title, stream_offsets) ⇒ Object

Save index entries to cache

Parameters:

  • entries_by_title (Hash) —

    title => entry hash

  • stream_offsets (Array<Integer>) —

    sorted stream offsets



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
128
129
130
 
# File 'lib/wp2txt/index_cache.rb', line 83

def save(entries_by_title, stream_offsets)
  FileUtils.mkdir_p(File.dirname(@cache_path))

  # Remove old cache if exists
  FileUtils.rm_f(@cache_path)

  open_db
  begin
    create_schema

    # Use transaction for better performance
    @db.execute("BEGIN TRANSACTION")

    # Save metadata
    source_stat = File.stat(@source_path)
    save_metadata(
      source_path: @source_path,
      source_mtime: source_stat.mtime.to_i,
      source_size: source_stat.size,
      cache_version: CACHE_VERSION,
      entry_count: entries_by_title.size
    )

    # Save entries in batches for performance
    stmt = @db.prepare("INSERT INTO index_entries (title, page_id, byte_offset) VALUES (?, ?, ?)")
    entries_by_title.each do |title, entry|
      stmt.execute([title, entry[:page_id], entry[:offset]])
    end
    stmt.close

    # Save stream offsets
    stmt = @db.prepare("INSERT INTO stream_offsets (byte_offset) VALUES (?)")
    stream_offsets.each do |offset|
      stmt.execute([offset])
    end
    stmt.close

    @db.execute("COMMIT")

    true
  rescue SQLite3::Exception => e
    @db.execute("ROLLBACK") rescue nil
    FileUtils.rm_f(@cache_path)
    raise e
  ensure
    close_db
  end
end

#stats ⇒ Object

Get cache statistics



159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
 
# File 'lib/wp2txt/index_cache.rb', line 159

def stats
  return nil unless File.exist?(@cache_path)

  open_db
  begin
    meta = load_metadata
    entry_count = @db.get_first_value("SELECT COUNT(*) FROM index_entries")
    stream_count = @db.get_first_value("SELECT COUNT(*) FROM stream_offsets")

    {
      cache_path: @cache_path,
      cache_size: File.size(@cache_path),
      entry_count: entry_count,
      stream_count: stream_count,
      source_path: meta[:source_path],
      source_mtime: meta[:source_mtime] ? Time.at(meta[:source_mtime].to_i) : nil,
      cache_version: meta[:cache_version]
    }
  ensure
    close_db
  end
end

#valid? ⇒ Boolean

Check if cache exists and is valid for the current source file

Returns:

  • (Boolean) —

    true if cache is usable



25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
 
# File 'lib/wp2txt/index_cache.rb', line 25

def valid?
  return false unless File.exist?(@cache_path)
  return false unless File.exist?(@source_path)

  begin
    open_db
    meta = load_metadata
    return false unless meta

    # Check cache version
    return false if meta[:cache_version].to_i != CACHE_VERSION

    # Check source file hasn't changed
    source_stat = File.stat(@source_path)
    return false if meta[:source_mtime].to_i != source_stat.mtime.to_i
    return false if meta[:source_size].to_i != source_stat.size

    true
  rescue SQLite3::Exception
    false
  ensure
    close_db
  end
end