claude_memory 0.4.0 → 0.5.0

data/lib/claude_memory/resolve/resolver.rb

@@ -50,61 +50,91 @@ module ClaudeMemory
       end
 
       def resolve_fact(fact_data, entity_ids, content_item_id, occurred_at)
-        subject_id = entity_ids[fact_data[:subject]] ||
+        subject_id = resolve_subject(fact_data, entity_ids)
+        existing_facts = @store.facts_for_slot(subject_id, fact_data[:predicate])
+        resolution = determine_resolution(existing_facts, fact_data, entity_ids)
+
+        apply_resolution(resolution, fact_data, subject_id, entity_ids, content_item_id, occurred_at, existing_facts)
+      end
+
+      def resolve_subject(fact_data, entity_ids)
+        entity_ids[fact_data[:subject]] ||
           @store.find_or_create_entity(type: "repo", name: fact_data[:subject])
+      end
 
-        predicate = fact_data[:predicate]
-        object_val = fact_data[:object]
-        object_entity_id = entity_ids[object_val]
-
-        outcome = {created: 0, superseded: 0, conflicts: 0, provenance: 0}
-
-        existing_facts = @store.facts_for_slot(subject_id, predicate)
-
-        # No transaction wrapper needed - handled by apply method
-        # This allows all facts to be processed in a single transaction
-        if PredicatePolicy.single?(predicate) && existing_facts.any?
-          matching = existing_facts.find { |f| values_match?(f, object_val, object_entity_id) }
-          if matching
-            add_provenance(matching[:id], content_item_id, fact_data)
-            outcome[:provenance] = 1
-            return outcome
-          elsif supersession_signal?(fact_data)
-            supersede_facts(existing_facts, occurred_at)
-            outcome[:superseded] = existing_facts.size
-          else
-            create_conflict(existing_facts.first[:id], fact_data, subject_id, content_item_id, occurred_at)
-            outcome[:conflicts] = 1
-            return outcome
-          end
+      def determine_resolution(existing_facts, fact_data, entity_ids)
+        return :insert unless PredicatePolicy.single?(fact_data[:predicate]) && existing_facts.any?
+
+        object_entity_id = entity_ids[fact_data[:object]]
+        matching = existing_facts.find { |f| values_match?(f, fact_data[:object], object_entity_id) }
+
+        if matching
+          :reinforce
+        elsif supersession_signal?(fact_data)
+          :supersede
+        else
+          :conflict
+        end
+      end
+
+      def apply_resolution(resolution, fact_data, subject_id, entity_ids, content_item_id, occurred_at, existing_facts)
+        case resolution
+        when :reinforce
+          apply_reinforcement(existing_facts, fact_data, entity_ids, content_item_id)
+        when :conflict
+          apply_conflict(existing_facts, fact_data, subject_id, content_item_id, occurred_at)
+        else
+          apply_insert(fact_data, subject_id, entity_ids, content_item_id, occurred_at, existing_facts, resolution)
+        end
+      end
+
+      def apply_reinforcement(existing_facts, fact_data, entity_ids, content_item_id)
+        object_entity_id = entity_ids[fact_data[:object]]
+        matching = existing_facts.find { |f| values_match?(f, fact_data[:object], object_entity_id) }
+        add_provenance(matching[:id], content_item_id, fact_data)
+        {created: 0, superseded: 0, conflicts: 0, provenance: 1}
+      end
+
+      def apply_conflict(existing_facts, fact_data, subject_id, content_item_id, occurred_at)
+        create_conflict(existing_facts.first[:id], fact_data, subject_id, content_item_id, occurred_at)
+        {created: 0, superseded: 0, conflicts: 1, provenance: 0}
+      end
+
+      def apply_insert(fact_data, subject_id, entity_ids, content_item_id, occurred_at, existing_facts, resolution)
+        superseded_count = 0
+        if resolution == :supersede
+          supersede_facts(existing_facts, occurred_at)
+          superseded_count = existing_facts.size
         end
 
+        fact_id = insert_new_fact(fact_data, subject_id, entity_ids, occurred_at)
+        link_superseded_facts(fact_id, existing_facts) if superseded_count > 0
+        add_provenance(fact_id, content_item_id, fact_data)
+
+        {created: 1, superseded: superseded_count, conflicts: 0, provenance: 1}
+      end
+
+      def insert_new_fact(fact_data, subject_id, entity_ids, occurred_at)
         fact_scope = fact_data[:scope_hint] || @current_scope
         fact_project = (fact_scope == "global") ? nil : @current_project_path
 
-        fact_id = @store.insert_fact(
+        @store.insert_fact(
           subject_entity_id: subject_id,
-          predicate: predicate,
-          object_entity_id: object_entity_id,
-          object_literal: object_val,
+          predicate: fact_data[:predicate],
+          object_entity_id: entity_ids[fact_data[:object]],
+          object_literal: fact_data[:object],
           polarity: fact_data[:polarity] || "positive",
           confidence: fact_data[:confidence] || 1.0,
           valid_from: occurred_at,
           scope: fact_scope,
           project_path: fact_project
         )
-        outcome[:created] = 1
+      end
 
-        if existing_facts.any? && outcome[:superseded] > 0
-          existing_facts.each do |old_fact|
-            @store.insert_fact_link(from_fact_id: fact_id, to_fact_id: old_fact[:id], link_type: "supersedes")
-          end
+      def link_superseded_facts(new_fact_id, old_facts)
+        old_facts.each do |old_fact|
+          @store.insert_fact_link(from_fact_id: new_fact_id, to_fact_id: old_fact[:id], link_type: "supersedes")
         end
-
-        add_provenance(fact_id, content_item_id, fact_data)
-        outcome[:provenance] = 1
-
-        outcome
       end
 
       def supersession_signal?(fact_data)
@@ -151,7 +181,9 @@ module ClaudeMemory
           fact_id: fact_id,
           content_item_id: content_item_id,
           quote: fact_data[:quote],
-          strength: fact_data[:strength] || "stated"
+          strength: fact_data[:strength] || "stated",
+          line_start: fact_data[:line_start],
+          line_end: fact_data[:line_end]
         )
       end
     end

data/lib/claude_memory/store/sqlite_store.rb

@@ -2,6 +2,7 @@
 
 require "sequel"
 require "sequel/extensions/migration"
+require "digest"
 require "json"
 require "extralite"
 require "sequel/adapters/extralite"
@@ -9,7 +10,7 @@ require "sequel/adapters/extralite"
 module ClaudeMemory
   module Store
     class SQLiteStore
-      SCHEMA_VERSION = 7
+      SCHEMA_VERSION = 10
 
       attr_reader :db
 
@@ -17,32 +18,20 @@ module ClaudeMemory
         @db_path = db_path
         @db = connect_database(db_path)
 
-        configure_pragmas
-
         ensure_schema!
       end
 
       private
 
       def connect_database(db_path)
-        # Extralite adapter: Better performance and GVL release
-        Sequel.connect("extralite:#{db_path}")
-      end
-
-      def configure_pragmas
-        # Enable WAL mode for better concurrency
-        # - Multiple readers don't block each other
-        # - Writers don't block readers
-        # - Safer concurrent hook execution
-        @db.run("PRAGMA journal_mode = WAL")
-        @db.run("PRAGMA synchronous = NORMAL")
-
-        # Set busy timeout to 30 seconds (increased from 5s)
-        # - Allows much longer wait times before raising BusyException
-        # - Critical for concurrent hook execution with MCP server
-        # - Combined with ingester retry logic, provides ~5 minutes total wait
-        # - Extralite releases GVL for better threading performance
-        @db.run("PRAGMA busy_timeout = 30000")
+        Sequel.connect(
+          "extralite:#{db_path}",
+          connect_sqls: [
+            "PRAGMA journal_mode = WAL",
+            "PRAGMA synchronous = NORMAL",
+            "PRAGMA busy_timeout = 30000"
+          ]
+        )
       end
 
       public
@@ -110,6 +99,10 @@ module ClaudeMemory
         @db[:ingestion_metrics]
       end
 
+      def llm_cache
+        @db[:llm_cache]
+      end
+
       def upsert_content_item(source:, text_hash:, byte_len:, session_id: nil, transcript_path: nil,
         project_path: nil, occurred_at: nil, raw_text: nil, metadata: nil,
         git_branch: nil, cwd: nil, claude_version: nil, thinking_level: nil, source_mtime: nil)
@@ -194,6 +187,7 @@ module ClaudeMemory
         datatype: nil, polarity: "positive", valid_from: nil, status: "active",
         confidence: 1.0, created_from: nil, scope: "project", project_path: nil)
         now = Time.now.utc.iso8601
+        docid = generate_docid(subject_entity_id, predicate, object_literal, now)
         facts.insert(
           subject_entity_id: subject_entity_id,
           predicate: predicate,
@@ -207,10 +201,15 @@ module ClaudeMemory
           created_from: created_from,
           created_at: now,
           scope: scope,
-          project_path: project_path
+          project_path: project_path,
+          docid: docid
         )
       end
 
+      def find_fact_by_docid(docid)
+        facts.where(docid: docid).first
+      end
+
       def update_fact(fact_id, status: nil, valid_to: nil, scope: nil, project_path: nil, embedding: nil)
         updates = {}
         updates[:status] = status if status
@@ -253,13 +252,16 @@ module ClaudeMemory
           .all
       end
 
-      def insert_provenance(fact_id:, content_item_id: nil, quote: nil, attribution_entity_id: nil, strength: "stated")
+      def insert_provenance(fact_id:, content_item_id: nil, quote: nil, attribution_entity_id: nil, strength: "stated",
+        line_start: nil, line_end: nil)
         provenance.insert(
           fact_id: fact_id,
           content_item_id: content_item_id,
           quote: quote,
           attribution_entity_id: attribution_entity_id,
-          strength: strength
+          strength: strength,
+          line_start: line_start,
+          line_end: line_end
         )
       end
 
@@ -343,6 +345,65 @@ module ClaudeMemory
         }
       end
 
+      # Look up a cached LLM response by cache key
+      #
+      # @param cache_key [String] SHA256 hex digest of operation+model+input
+      # @return [Hash, nil] Cached result row or nil
+      def llm_cache_lookup(cache_key)
+        llm_cache.where(cache_key: cache_key).first
+      end
+
+      # Store an LLM response in the cache
+      #
+      # @param operation [String] Operation type (e.g., "distill", "extract")
+      # @param model [String] Model identifier
+      # @param input_hash [String] SHA256 of input content
+      # @param result_json [String] JSON response to cache
+      # @param input_tokens [Integer, nil] Tokens in request
+      # @param output_tokens [Integer, nil] Tokens in response
+      # @return [Integer] The created cache entry ID
+      def llm_cache_store(operation:, model:, input_hash:, result_json:, input_tokens: nil, output_tokens: nil)
+        cache_key = Digest::SHA256.hexdigest("#{operation}:#{model}:#{input_hash}")
+
+        llm_cache
+          .insert_conflict(target: :cache_key, update: {
+            result_json: result_json,
+            input_tokens: input_tokens,
+            output_tokens: output_tokens,
+            created_at: Time.now.utc.iso8601
+          })
+          .insert(
+            cache_key: cache_key,
+            operation: operation,
+            model: model,
+            input_hash: input_hash,
+            result_json: result_json,
+            input_tokens: input_tokens,
+            output_tokens: output_tokens,
+            created_at: Time.now.utc.iso8601
+          )
+      end
+
+      # Generate a cache key for LLM response lookup
+      #
+      # @param operation [String] Operation type
+      # @param model [String] Model identifier
+      # @param input [String] Raw input content
+      # @return [String] SHA256 hex digest cache key
+      def llm_cache_key(operation, model, input)
+        input_hash = Digest::SHA256.hexdigest(input)
+        Digest::SHA256.hexdigest("#{operation}:#{model}:#{input_hash}")
+      end
+
+      # Prune cache entries older than the given age
+      #
+      # @param max_age_seconds [Integer] Maximum age in seconds (default: 7 days)
+      # @return [Integer] Number of entries pruned
+      def llm_cache_prune(max_age_seconds: 604_800)
+        cutoff = (Time.now - max_age_seconds).utc.iso8601
+        llm_cache.where { created_at < cutoff }.delete
+      end
+
       private
 
       def ensure_schema!
@@ -398,6 +459,20 @@ module ClaudeMemory
         @db[:meta].where(key: key).get(:value)
       end
 
+      def generate_docid(subject_entity_id, predicate, object_literal, created_at)
+        input = "#{subject_entity_id}:#{predicate}:#{object_literal}:#{created_at}"
+        docid = Digest::SHA256.hexdigest(input)[0, 8]
+
+        # Handle unlikely collisions by rehashing with a counter
+        counter = 0
+        while facts.where(docid: docid).any?
+          counter += 1
+          docid = Digest::SHA256.hexdigest("#{input}:#{counter}")[0, 8]
+        end
+
+        docid
+      end
+
       def slugify(type, name)
         "#{type}:#{name.downcase.gsub(/[^a-z0-9]+/, "_").gsub(/^_|_$/, "")}"
       end

data/lib/claude_memory/sweep/sweeper.rb

@@ -35,6 +35,12 @@ module ClaudeMemory
 
         @stats[:elapsed_seconds] = Time.now - @start_time
         @stats[:budget_honored] = @stats[:elapsed_seconds] <= budget
+        ClaudeMemory.logger.info("sweep",
+          message: "Sweep complete",
+          elapsed_seconds: @stats[:elapsed_seconds].round(3),
+          budget_honored: @stats[:budget_honored],
+          proposed_expired: @stats[:proposed_facts_expired],
+          disputed_expired: @stats[:disputed_facts_expired])
         @stats
       end
 

data/lib/claude_memory/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module ClaudeMemory
-  VERSION = "0.4.0"
+  VERSION = "0.5.0"
 end

data/lib/claude_memory.rb

@@ -13,6 +13,7 @@ require_relative "claude_memory/core/null_explanation"
 require_relative "claude_memory/core/token_estimator"
 require_relative "claude_memory/core/batch_loader"
 require_relative "claude_memory/core/fact_ranker"
+require_relative "claude_memory/core/rr_fusion"
 require_relative "claude_memory/core/concept_ranker"
 require_relative "claude_memory/core/scope_filter"
 require_relative "claude_memory/core/result_builder"
@@ -20,7 +21,10 @@ require_relative "claude_memory/core/fact_collector"
 require_relative "claude_memory/core/embedding_candidate_builder"
 require_relative "claude_memory/core/fact_query_builder"
 require_relative "claude_memory/core/result_sorter"
+require_relative "claude_memory/core/relative_time"
 require_relative "claude_memory/core/text_builder"
+require_relative "claude_memory/core/snippet_extractor"
+require_relative "claude_memory/core/fact_graph"
 require_relative "claude_memory/commands/base_command"
 require_relative "claude_memory/commands/checks/database_check"
 require_relative "claude_memory/commands/checks/snapshot_check"
@@ -75,17 +79,22 @@ require_relative "claude_memory/ingest/privacy_tag"
 require_relative "claude_memory/ingest/content_sanitizer"
 require_relative "claude_memory/ingest/metadata_extractor"
 require_relative "claude_memory/ingest/tool_extractor"
+require_relative "claude_memory/ingest/tool_filter"
 require_relative "claude_memory/ingest/ingester"
 require_relative "claude_memory/ingest/transcript_reader"
+require_relative "claude_memory/logging/logger"
 require_relative "claude_memory/infrastructure/file_system"
 require_relative "claude_memory/infrastructure/in_memory_file_system"
 require_relative "claude_memory/infrastructure/operation_tracker"
 require_relative "claude_memory/infrastructure/schema_validator"
+require_relative "claude_memory/mcp/query_guide"
+require_relative "claude_memory/mcp/text_summary"
 require_relative "claude_memory/mcp/tool_helpers"
 require_relative "claude_memory/mcp/server"
 require_relative "claude_memory/mcp/tools"
 require_relative "claude_memory/publish"
 require_relative "claude_memory/recall/dual_query_template"
+require_relative "claude_memory/recall/expansion_detector"
 require_relative "claude_memory/recall"
 require_relative "claude_memory/shortcuts"
 require_relative "claude_memory/resolve/predicate_policy"
@@ -103,4 +112,16 @@ module ClaudeMemory
   def self.project_db_path(project_path = Dir.pwd)
     Configuration.new.project_db_path(project_path)
   end
+
+  # Module-level logger instance, shared across components
+  # @return [Logging::Logger, Logging::NullLogger]
+  def self.logger
+    @logger ||= Logging::Logger.new
+  end
+
+  # Replace the module-level logger (useful for testing)
+  # @param logger [Logging::Logger, Logging::NullLogger]
+  def self.logger=(logger)
+    @logger = logger
+  end
 end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: claude_memory
 version: !ruby/object:Gem::Version
-  version: 0.4.0
+  version: 0.5.0
 platform: ruby
 authors:
 - Valentino Stoll
@@ -88,6 +88,9 @@ files:
 - db/migrations/005_add_incremental_sync.rb
 - db/migrations/006_add_operation_tracking.rb
 - db/migrations/007_add_ingestion_metrics.rb
+- db/migrations/008_add_provenance_line_range.rb
+- db/migrations/009_add_docid.rb
+- db/migrations/010_add_llm_cache.rb
 - docs/EXAMPLES.md
 - docs/GETTING_STARTED.md
 - docs/RELEASE_NOTES_v0.2.0.md
@@ -103,6 +106,7 @@ files:
 - docs/expert_review.md
 - docs/improvements.md
 - docs/influence/.gitkeep
+- docs/influence/claude-supermemory.md
 - docs/influence/grepai.md
 - docs/influence/qmd.md
 - docs/organizational_memory_playbook.md
@@ -112,7 +116,6 @@ files:
 - docs/plans/updated_plan.md
 - docs/plugin.md
 - docs/quality_review.md
-- docs/remaining_improvements.md
 - docs/review_summary.md
 - exe/claude-memory
 - hooks/hooks.json
@@ -157,16 +160,20 @@ files:
 - lib/claude_memory/core/concept_ranker.rb
 - lib/claude_memory/core/embedding_candidate_builder.rb
 - lib/claude_memory/core/fact_collector.rb
+- lib/claude_memory/core/fact_graph.rb
 - lib/claude_memory/core/fact_id.rb
 - lib/claude_memory/core/fact_query_builder.rb
 - lib/claude_memory/core/fact_ranker.rb
 - lib/claude_memory/core/null_explanation.rb
 - lib/claude_memory/core/null_fact.rb
+- lib/claude_memory/core/relative_time.rb
 - lib/claude_memory/core/result.rb
 - lib/claude_memory/core/result_builder.rb
 - lib/claude_memory/core/result_sorter.rb
+- lib/claude_memory/core/rr_fusion.rb
 - lib/claude_memory/core/scope_filter.rb
 - lib/claude_memory/core/session_id.rb
+- lib/claude_memory/core/snippet_extractor.rb
 - lib/claude_memory/core/text_builder.rb
 - lib/claude_memory/core/token_estimator.rb
 - lib/claude_memory/core/transcript_path.rb
@@ -196,16 +203,21 @@ files:
 - lib/claude_memory/ingest/metadata_extractor.rb
 - lib/claude_memory/ingest/privacy_tag.rb
 - lib/claude_memory/ingest/tool_extractor.rb
+- lib/claude_memory/ingest/tool_filter.rb
 - lib/claude_memory/ingest/transcript_reader.rb
+- lib/claude_memory/logging/logger.rb
+- lib/claude_memory/mcp/query_guide.rb
 - lib/claude_memory/mcp/response_formatter.rb
 - lib/claude_memory/mcp/server.rb
 - lib/claude_memory/mcp/setup_status_analyzer.rb
+- lib/claude_memory/mcp/text_summary.rb
 - lib/claude_memory/mcp/tool_definitions.rb
 - lib/claude_memory/mcp/tool_helpers.rb
 - lib/claude_memory/mcp/tools.rb
 - lib/claude_memory/publish.rb
 - lib/claude_memory/recall.rb
 - lib/claude_memory/recall/dual_query_template.rb
+- lib/claude_memory/recall/expansion_detector.rb
 - lib/claude_memory/resolve/predicate_policy.rb
 - lib/claude_memory/resolve/resolver.rb
 - lib/claude_memory/shortcuts.rb