Class: CardIndex

Inherits:
Object
  • Object
show all
Defined in:
lib/zillacore/card_index.rb

Constant Summary collapse

SIMILARITY_THRESHOLD = 
0.65
SEMANTIC_THRESHOLD = 
0.65
SEMANTIC_COLLECTION = 
"card-titles"
QMD_DEBOUNCE =

seconds

30

Instance Attribute Summary collapse

Instance Method Summary collapse

Constructor Details

#initialize(index_file:, titles_dir:) ⇒ CardIndex

Returns a new instance of CardIndex.



28
29
30
31
32
33
34
35
36
 
# File 'lib/zillacore/card_index.rb', line 28

def initialize(index_file:, titles_dir:)
  @index_file = index_file
  @titles_dir = titles_dir
  @data = {}
  @mutex = Mutex.new
  @qmd_mutex = Mutex.new
  @qmd_last_run = nil
  @qmd_pending = false
end

Instance Attribute Details

#index_file ⇒ Object (readonly)

Returns the value of attribute index_file.



26
27
28
 
# File 'lib/zillacore/card_index.rb', line 26

def index_file
  @index_file
end

#titles_dir ⇒ Object (readonly)

Returns the value of attribute titles_dir.



26
27
28
 
# File 'lib/zillacore/card_index.rb', line 26

def titles_dir
  @titles_dir
end

Instance Method Details

#[](key) ⇒ Object

— Hash-like access (thread-safe) —



40
41
42
 
# File 'lib/zillacore/card_index.rb', line 40

def [](key)
  @mutex.synchronize { @data[key] }
end

#[]=(key, value) ⇒ Object 



44
45
46
 
# File 'lib/zillacore/card_index.rb', line 44

def []=(key, value)
  @mutex.synchronize { @data[key] = value }
end

#backfill ⇒ Object

— Backfill from Fizzy API on startup —



304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
 
# File 'lib/zillacore/card_index.rb', line 304

def backfill
  Thread.new do
    LOG.info "[CardIndex] Starting backfill from Fizzy API..."
    backfilled = 0
    seen_boards = Set.new

    PROJECTS.each do |project_key, config|
      result = backfill_project(project_key, config, seen_boards)
      backfilled += result if result
    end

    save
    LOG.info "[CardIndex] Backfill complete: #{backfilled} new cards indexed (#{size} total)"

    ensure_card_titles_collection
    schedule_qmd_reindex
  end
end

#backfill_project(project_key, config, seen_boards) ⇒ Object

Backfill cards for a single project. Returns count of new cards indexed, or nil if skipped.



324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
 
# File 'lib/zillacore/card_index.rb', line 324

def backfill_project(project_key, config, seen_boards)
  repo_path = config["repo_path"]
  return nil unless repo_path && File.directory?(repo_path)

  fizzy_yaml = File.join(repo_path, ".fizzy.yaml")
  unless File.exist?(fizzy_yaml)
    LOG.debug "[CardIndex] Skipping '#{project_key}' — no .fizzy.yaml"
    return nil
  end

  begin
    board_id = YAML.safe_load_file(fizzy_yaml)["board"]
  rescue StandardError => e
    LOG.warn "[CardIndex] Could not read .fizzy.yaml for '#{project_key}': #{e.message}"
    return nil
  end

  if seen_boards.include?(board_id)
    LOG.debug "[CardIndex] Skipping '#{project_key}' — board #{board_id} already fetched"
    return nil
  end
  seen_boards << board_id

  count = 0
  output = run_cmd("fizzy", "card", "list", "--all", chdir: repo_path, env: default_fizzy_env)
  cards = JSON.parse(output)["data"] || []
  cards.each do |card|
    num = card["number"]
    next unless num
    next if key?(num.to_s)

    index_card(
      number: num,
      title: card["title"] || card["description"]&.slice(0, 80) || "untitled",
      creator_name: card.dig("creator", "name"),
      creator_id: card.dig("creator", "id"),
      tags: card["tags"] || [],
      closed: card["closed"] || false
    )
    count += 1
  end
  count
rescue StandardError => e
  LOG.warn "[CardIndex] Backfill failed for project '#{project_key}': #{e.message}"
  0
end

#build_scope_map ⇒ Object

— Scope extraction for cross-project duplicate filtering —



201
202
203
204
205
206
207
208
209
210
 
# File 'lib/zillacore/card_index.rb', line 201

def build_scope_map
  return if @scope_map_built

  @scope_map ||= {}
  PROJECTS.each do |key, cfg|
    (cfg["fizzy_tags"] || []).each { |t| @scope_map[t.downcase] = key }
    (cfg["scope_tags"] || {}).each { |tag, scope| @scope_map[tag.downcase] = scope }
  end
  @scope_map_built = true
end

#card_scopes(tags) ⇒ Object 



212
213
214
215
216
217
218
219
220
 
# File 'lib/zillacore/card_index.rb', line 212

def card_scopes(tags)
  return Set.new if tags.nil? || tags.empty?

  build_scope_map
  tag_names = tags.map { |t| (t.is_a?(Hash) ? t["name"] : t).to_s.downcase }
  scopes = Set.new
  tag_names.each { |t| scopes << @scope_map[t] if @scope_map[t] }
  scopes
end

#delete(key) ⇒ Object 



48
49
50
 
# File 'lib/zillacore/card_index.rb', line 48

def delete(key)
  @mutex.synchronize { @data.delete(key) }
end

#different_scopes?(tags_a, tags_b) ⇒ Boolean

Returns:

  • (Boolean) 


222
223
224
225
226
 
# File 'lib/zillacore/card_index.rb', line 222

def different_scopes?(tags_a, tags_b)
  scopes_a = card_scopes(tags_a)
  scopes_b = card_scopes(tags_b)
  scopes_a.any? && scopes_b.any? && !scopes_a.intersect?(scopes_b)
end

#dig(*keys) ⇒ Object 



64
65
66
 
# File 'lib/zillacore/card_index.rb', line 64

def dig(*keys)
  @mutex.synchronize { @data.dig(*keys) }
end

#each ⇒ Object 



60
61
62
 
# File 'lib/zillacore/card_index.rb', line 60

def each(&)
  @mutex.synchronize { @data.each(&) }
end

#ensure_card_titles_collection ⇒ Object

Ensure the qmd collection exists, create if not



112
113
114
115
116
117
118
119
120
121
 
# File 'lib/zillacore/card_index.rb', line 112

def ensure_card_titles_collection
  FileUtils.mkdir_p(@titles_dir)
  output, _, status = Open3.capture3("qmd", "collection", "list")
  return if status.success? && output.include?(SEMANTIC_COLLECTION)

  LOG.info "[CardIndex] Creating qmd collection '#{SEMANTIC_COLLECTION}'"
  _, stderr, s = Open3.capture3("qmd", "collection", "add", @titles_dir,
                                "--name", SEMANTIC_COLLECTION, "--mask", "*.md")
  LOG.warn "[CardIndex] Failed to create qmd collection: #{stderr}" unless s.success?
end

#evict_card(number) ⇒ Object 



194
195
196
197
 
# File 'lib/zillacore/card_index.rb', line 194

def evict_card(number)
  delete(number.to_s)
  remove_card_title_file(number)
end

#find_semantic_similar_cards(title, exclude_number: nil) ⇒ Object

— Semantic search via qmd vsearch —



244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
 
# File 'lib/zillacore/card_index.rb', line 244

def find_semantic_similar_cards(title, exclude_number: nil)
  output, stderr, status = Open3.capture3("qmd", "vsearch", title, "-c", SEMANTIC_COLLECTION,
                                          "--json", "--min-score", SEMANTIC_THRESHOLD.to_s, "--all")
  unless status.success?
    LOG.warn "[CardIndex] qmd vsearch failed: #{stderr.lines.last&.strip}"
    return []
  end

  clean = output.lines.reject { |l| l.start_with?("[node-llama-cpp]") }.join
  json_start = clean.index("[")
  return [] unless json_start

  results = JSON.parse(clean[json_start..])
  results.filter_map do |r|
    num = r["file"]&.match(%r{/(\d+)\.md$})&.[](1)
    next unless num
    next if num == exclude_number.to_s

    entry = self[num]
    next if entry&.dig("closed")

    { number: num.to_i, title: entry&.dig("title") || r["snippet"]&.strip || "", score: r["score"], method: :semantic }
  end
rescue JSON::ParserError => e
  LOG.warn "[CardIndex] Failed to parse qmd vsearch output: #{e.message}"
  []
end

#find_similar_cards(title, exclude_number: nil, tags: nil) ⇒ Object

— Merged search: trigram + semantic in parallel —



274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
 
# File 'lib/zillacore/card_index.rb', line 274

def find_similar_cards(title, exclude_number: nil, tags: nil)
  trigram_thread = Thread.new { find_trigram_similar_cards(title, exclude_number: exclude_number) }
  semantic_thread = Thread.new { find_semantic_similar_cards(title, exclude_number: exclude_number) }

  trigram_results = trigram_thread.value
  semantic_results = semantic_thread.value

  merged = {}
  (trigram_results + semantic_results).each do |match|
    key = match[:number]
    existing = merged[key]
    if existing.nil? || match[:score] > existing[:score]
      merged[key] = match
    elsif match[:score] == existing[:score] && existing[:method] != match[:method]
      merged[key] = existing.merge(method: :both)
    end
  end

  if tags && card_scopes(tags).any?
    merged.reject! do |num, _match|
      match_tags = dig(num.to_s, "tags")
      different_scopes?(tags, match_tags)
    end
  end

  merged.values.sort_by { |m| -m[:score] }
end

#find_trigram_similar_cards(title, exclude_number: nil) ⇒ Object

— Trigram search —



230
231
232
233
234
235
236
237
238
239
240
 
# File 'lib/zillacore/card_index.rb', line 230

def find_trigram_similar_cards(title, exclude_number: nil)
  matches = []
  each do |num, entry|
    next if num == exclude_number.to_s
    next if entry["closed"]

    score = trigram_similarity(title, entry["title"])
    matches << { number: num.to_i, title: entry["title"], score: score, method: :trigram } if score >= SIMILARITY_THRESHOLD
  end
  matches
end

#index_card(number:, title:, creator_name: nil, creator_id: nil, tags: [], closed: false) ⇒ Object 



180
181
182
183
184
185
186
187
188
189
190
191
192
 
# File 'lib/zillacore/card_index.rb', line 180

def index_card(number:, title:, creator_name: nil, creator_id: nil, tags: [], closed: false)
  @mutex.synchronize do
    @data[number.to_s] = {
      "title" => title,
      "creator_name" => creator_name,
      "creator_id" => creator_id,
      "tags" => tags.map { |t| t.is_a?(Hash) ? t["name"] : t.to_s },
      "closed" => closed,
      "indexed_at" => Time.now.iso8601
    }
  end
  sync_card_title_file(number, title, closed: closed)
end

#key?(key) ⇒ Boolean

Returns:

  • (Boolean) 


56
57
58
 
# File 'lib/zillacore/card_index.rb', line 56

def key?(key)
  @mutex.synchronize { @data.key?(key) }
end

#load ⇒ Object

— Index operations —



161
162
163
164
165
166
167
168
169
170
171
172
 
# File 'lib/zillacore/card_index.rb', line 161

def load
  data = if File.exist?(@index_file)
           JSON.parse(File.read(@index_file))
         else
           {}
         end
  @mutex.synchronize { @data.replace(data) }
  LOG.info "[CardIndex] Loaded #{size} cards from disk"
rescue JSON::ParserError => e
  LOG.error "Failed to parse card index: #{e.message}"
  @mutex.synchronize { @data.replace({}) }
end

#remove_card_title_file(number) ⇒ Object 



107
108
109
 
# File 'lib/zillacore/card_index.rb', line 107

def remove_card_title_file(number)
  FileUtils.rm_f(File.join(@titles_dir, "#{number}.md"))
end

#save ⇒ Object 



174
175
176
177
178
 
# File 'lib/zillacore/card_index.rb', line 174

def save
  @mutex.synchronize do
    File.write(@index_file, JSON.generate(@data))
  end
end

#schedule_qmd_reindex ⇒ Object

Debounced qmd update + embed. Runs in background thread.



124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
 
# File 'lib/zillacore/card_index.rb', line 124

def schedule_qmd_reindex
  @qmd_mutex.synchronize do
    @qmd_pending = true
    return if @qmd_last_run && (Time.now - @qmd_last_run) < QMD_DEBOUNCE

    @qmd_last_run = Time.now
    @qmd_pending = false
  end

  Thread.new do
    LOG.info "[CardIndex] Running qmd update for card titles..."
    _, stderr, s = Open3.capture3("qmd", "update")
    LOG.warn "[CardIndex] qmd update failed: #{stderr}" unless s.success?

    LOG.info "[CardIndex] Running qmd embed for card titles..."
    _, stderr, s = Open3.capture3("qmd", "embed")
    LOG.warn "[CardIndex] qmd embed failed: #{stderr}" unless s.success?

    LOG.info "[CardIndex] qmd reindex complete"

    needs_rerun = @qmd_mutex.synchronize do
      if @qmd_pending
        @qmd_pending = false
        @qmd_last_run = Time.now
        true
      else
        false
      end
    end
    schedule_qmd_reindex if needs_rerun
  rescue StandardError => e
    LOG.warn "[CardIndex] qmd reindex failed: #{e.message}"
  end
end

#size ⇒ Object 



52
53
54
 
# File 'lib/zillacore/card_index.rb', line 52

def size
  @mutex.synchronize { @data.size }
end

#sync_card_title_file(number, title, closed: false) ⇒ Object

— Card title files for qmd collection —



97
98
99
100
101
102
103
104
105
 
# File 'lib/zillacore/card_index.rb', line 97

def sync_card_title_file(number, title, closed: false)
  FileUtils.mkdir_p(@titles_dir)
  path = File.join(@titles_dir, "#{number}.md")
  if closed
    FileUtils.rm_f(path)
  else
    File.write(path, title)
  end
end

#sync_title_files ⇒ Object

— Startup —



373
374
375
376
377
378
 
# File 'lib/zillacore/card_index.rb', line 373

def sync_title_files
  FileUtils.mkdir_p(@titles_dir)
  each do |num, entry|
    sync_card_title_file(num, entry["title"], closed: entry["closed"])
  end
end

#to_h ⇒ Object 



72
73
74
 
# File 'lib/zillacore/card_index.rb', line 72

def to_h
  @mutex.synchronize { @data.dup }
end

#to_json ⇒ Object 



68
69
70
 
# File 'lib/zillacore/card_index.rb', line 68

def to_json(...)
  @mutex.synchronize { @data.to_json(...) }
end

#trigram_similarity(str_a, str_b) ⇒ Object 



85
86
87
88
89
90
91
92
93
 
# File 'lib/zillacore/card_index.rb', line 85

def trigram_similarity(str_a, str_b)
  ta = trigrams(str_a)
  tb = trigrams(str_b)
  return 0.0 if ta.empty? || tb.empty?

  intersection = (ta & tb).size.to_f
  union = (ta | tb).size.to_f
  intersection / union
end

#trigrams(str) ⇒ Object

— Trigram similarity —



78
79
80
81
82
83
 
# File 'lib/zillacore/card_index.rb', line 78

def trigrams(str)
  normalized = str.downcase.gsub(/[^a-z0-9\s]/, "").strip
  return Set.new if normalized.length < 3

  Set.new((0..(normalized.length - 3)).map { |i| normalized[i, 3] })
end