red-candle 1.1.1 → 1.2.0

data/lib/candle/ner.rb

@@ -30,10 +30,10 @@ module Candle
       # Load a pre-trained NER model from HuggingFace
       #
       # @param model_id [String] HuggingFace model ID (e.g., "dslim/bert-base-NER")
-      # @param device [Device, nil] Device to run on (defaults to best available)
+      # @param device [Device] Device to run on (defaults to best available)
       # @param tokenizer [String, nil] Tokenizer model ID to use (defaults to same as model_id)
       # @return [NER] NER instance
-      def from_pretrained(model_id, device: nil, tokenizer: nil)
+      def from_pretrained(model_id, device: Candle::Device.best, tokenizer: nil)
         new(model_id, device, tokenizer)
       end
       
@@ -112,7 +112,7 @@ module Candle
     # @return [Array<Hash>] Filtered entities of the specified type
     def extract_entity_type(text, entity_type, confidence_threshold: 0.9)
       entities = extract_entities(text, confidence_threshold: confidence_threshold)
-      entities.select { |e| e["label"] == entity_type.upcase }
+      entities.select { |e| e[:label] == entity_type.upcase }
     end
     
     # Analyze text and return both entities and token predictions
@@ -137,12 +137,12 @@ module Candle
       return text if entities.empty?
       
       # Sort by start position (reverse for easier insertion)
-      entities.sort_by! { |e| -e["start"] }
+      entities.sort_by! { |e| -e[:start] }
       
       result = text.dup
       entities.each do |entity|
-        label = "[#{entity['label']}:#{entity['confidence'].round(2)}]"
-        result.insert(entity["end"], label)
+        label = "[#{entity[:label]}:#{entity[:confidence].round(2)}]"
+        result.insert(entity[:end], label)
       end
       
       result
@@ -152,7 +152,19 @@ module Candle
     #
     # @return [String] Model description
     def inspect
-      "#<Candle::NER #{model_info}>"
+      opts = options rescue {}
+      
+      parts = ["#<Candle::NER"]
+      parts << "model=#{opts["model_id"] || "unknown"}"
+      parts << "device=#{opts["device"] || "unknown"}"
+      parts << "labels=#{opts["num_labels"]}" if opts["num_labels"]
+      
+      if opts["entity_types"] && !opts["entity_types"].empty?
+        types = opts["entity_types"].sort.join(",")
+        parts << "types=#{types}"
+      end
+      
+      parts.join(" ") + ">"
     end
     
     alias to_s inspect
@@ -186,12 +198,12 @@ module Candle
           match_end = $~.offset(0)[1]
           
           entities << {
-            "text" => match_text,
-            "label" => @entity_type,
-            "start" => match_start,
-            "end" => match_end,
-            "confidence" => 1.0,
-            "source" => "pattern"
+            text: match_text,
+            label: @entity_type,
+            start: match_start,
+            end: match_end,
+            confidence: 1.0,
+            source: "pattern"
           }
         end
       end
@@ -242,12 +254,12 @@ module Candle
           
           if word_boundary?(prev_char) && word_boundary?(next_char)
             entities << {
-              "text" => text[idx, pattern.length],
-              "label" => @entity_type,
-              "start" => idx,
-              "end" => idx + pattern.length,
-              "confidence" => 1.0,
-              "source" => "gazetteer"
+              text: text[idx, pattern.length],
+              label: @entity_type,
+              start: idx,
+              end: idx + pattern.length,
+              confidence: 1.0,
+              source: "gazetteer"
             }
           end
           
@@ -327,19 +339,19 @@ module Candle
     
     def merge_entities(entities)
       # Sort by start position and confidence (descending)
-      sorted = entities.sort_by { |e| [e["start"], -e["confidence"]] }
+      sorted = entities.sort_by { |e| [e[:start], -e[:confidence]] }
       
       merged = []
       sorted.each do |entity|
         # Check if entity overlaps with any already merged
         overlaps = merged.any? do |existing|
-          entity["start"] < existing["end"] && entity["end"] > existing["start"]
+          entity[:start] < existing[:end] && entity[:end] > existing[:start]
         end
         
         merged << entity unless overlaps
       end
       
-      merged.sort_by { |e| e["start"] }
+      merged.sort_by { |e| e[:start] }
     end
   end
 end

data/lib/candle/reranker.rb

@@ -3,10 +3,20 @@ module Candle
     # Default model path for cross-encoder/ms-marco-MiniLM-L-12-v2
     DEFAULT_MODEL_PATH = "cross-encoder/ms-marco-MiniLM-L-12-v2"
     
+    # Load a pre-trained reranker model from HuggingFace
+    # @param model_id [String] HuggingFace model ID (defaults to cross-encoder/ms-marco-MiniLM-L-12-v2)
+    # @param device [Candle::Device] The device to use for computation (defaults to best available)
+    # @return [Reranker] A new Reranker instance
+    def self.from_pretrained(model_id = DEFAULT_MODEL_PATH, device: Candle::Device.best)
+      _create(model_id, device)
+    end
+    
     # Constructor for creating a new Reranker with optional parameters
+    # @deprecated Use {.from_pretrained} instead
     # @param model_path [String, nil] The path to the model on Hugging Face
     # @param device [Candle::Device, Candle::Device.cpu] The device to use for computation
-    def self.new(model_path: DEFAULT_MODEL_PATH, device: Candle::Device.cpu)
+    def self.new(model_path: DEFAULT_MODEL_PATH, device: Candle::Device.best)
+      $stderr.puts "[DEPRECATION] `Reranker.new` is deprecated. Please use `Reranker.from_pretrained` instead."
       _create(model_path, device)
     end
 
@@ -20,5 +30,14 @@ module Candle
         { doc_id: doc_id, score: score, text: doc }
       }
     end
+    
+    # Improved inspect method
+    def inspect
+      opts = options rescue {}
+      parts = ["#<Candle::Reranker"]
+      parts << "model=#{opts["model_id"] || "unknown"}"
+      parts << "device=#{opts["device"] || "unknown"}"
+      parts.join(" ") + ">"
+    end
   end
 end

data/lib/candle/tensor.rb

@@ -51,6 +51,21 @@ module Candle
       to_f.to_i
     end
     
+    # Improved inspect method showing shape, dtype, and device
+    def inspect
+      shape_str = shape.join("x")
+      
+      parts = ["#<Candle::Tensor"]
+      parts << "shape=#{shape_str}"
+      parts << "dtype=#{dtype}"
+      parts << "device=#{device}"
+      
+      # Add element count for clarity
+      parts << "elements=#{elem_count}"
+      
+      parts.join(" ") + ">"
+    end
+    
     
     # Override class methods to support keyword arguments for device
     class << self

data/lib/candle/version.rb

@@ -1,5 +1,5 @@
 # :nocov:
 module Candle
-  VERSION = "1.1.1"
+  VERSION = "1.2.0"
 end
 # :nocov:

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: red-candle
 version: !ruby/object:Gem::Version
-  version: 1.1.1
+  version: 1.2.0
 platform: ruby
 authors:
 - Christopher Petersen
@@ -9,7 +9,7 @@ authors:
 autorequire:
 bindir: bin
 cert_chain: []
-date: 2025-07-28 00:00:00.000000000 Z
+date: 2025-08-10 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: rb_sys
@@ -137,6 +137,20 @@ dependencies:
     - - ">="
       - !ruby/object:Gem::Version
         version: '0'
+- !ruby/object:Gem::Dependency
+  name: rspec
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: '3.13'
+  type: :development
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: '3.13'
 description: huggingface/candle for Ruby
 email:
 - chris@petersen.io
@@ -169,14 +183,14 @@ files:
 - ext/candle/src/llm/quantized_gguf.rs
 - ext/candle/src/llm/qwen.rs
 - ext/candle/src/llm/text_generation.rs
-- ext/candle/src/ner.rs
-- ext/candle/src/reranker.rs
 - ext/candle/src/ruby/device.rs
 - ext/candle/src/ruby/dtype.rs
 - ext/candle/src/ruby/embedding_model.rs
 - ext/candle/src/ruby/errors.rs
 - ext/candle/src/ruby/llm.rs
 - ext/candle/src/ruby/mod.rs
+- ext/candle/src/ruby/ner.rs
+- ext/candle/src/ruby/reranker.rs
 - ext/candle/src/ruby/result.rs
 - ext/candle/src/ruby/structured.rs
 - ext/candle/src/ruby/tensor.rs
@@ -196,6 +210,8 @@ files:
 - ext/candle/target/release/build/clang-sys-cac31d63c4694603/out/macros.rs
 - ext/candle/target/release/build/pulp-1b95cfe377eede97/out/x86_64_asm.rs
 - ext/candle/target/release/build/rb-sys-f8ac4edc30ab3e53/out/bindings-0.9.116-mri-arm64-darwin24-3.3.0.rs
+- ext/candle/tests/device_tests.rs
+- ext/candle/tests/tensor_tests.rs
 - lib/candle.rb
 - lib/candle/build_info.rb
 - lib/candle/device_utils.rb