cmfrec 0.1.0 → 0.1.1

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 2e9f45e0c3826b90788782ac8a0838476fc7849e75d91f2d5949b08b382a28c3
-  data.tar.gz: 2f44b988001bf2b23e3c5938b4c5e7b9089e8943d86aa42c86e72be0ff558ea1
+  metadata.gz: 2091fae505c3d98468c6cfa08e0f80c6e97804fe8558bea97191336fa0179423
+  data.tar.gz: b66cece9f8659a6f79ac6ee302a1d2068d9a8b89b8903b2182f0fa8be6838869
 SHA512:
-  metadata.gz: 592ef4363bc016da1a35a958f99dd13a0fc7e900ecbe77b3ba95fc4b4cbbe151397924ff90aab2e326ac3369ac07c0380c76b1d3a2ac71b390664118de0f8611
-  data.tar.gz: 27d3e1ce80e88af9e8b062837f68329b10718372be15c08d4cd2e056619d1cd9e5ed906e6070798c32993ccea2f617b706af5e6afb7cba4f3cfc70ea22393ba6
+  metadata.gz: 5feed4c89f6249646b61d0713bcbc83725561942437ba3abdf2ca70a82a8f76d4a614b4f47a2c07e158a7155442f31b730b1d759234520bd34e50cae7e723b16
+  data.tar.gz: 649ce14693b2b7cfcc6039d4e39d94c3e3dce695e32b62f6c8c94c2751621622c519e5b536c917446e781b1acbadd3a511789efaae3ffc4aa2062738697f975f

data/CHANGELOG.md

@@ -1,3 +1,7 @@
+## 0.1.1 (2020-11-28)
+
+- Added `predict` method
+
 ## 0.1.0 (2020-11-27)
 
 - First release

data/README.md

@@ -6,8 +6,6 @@
 - Works with explicit and implicit feedback
 - Uses high-performance matrix factorization
 
-Not available for Windows yet
-
 [![Build Status](https://github.com/ankane/cmfrec/workflows/build/badge.svg?branch=master)](https://github.com/ankane/cmfrec/actions)
 
 ## Installation
@@ -18,6 +16,8 @@ Add this line to your application’s Gemfile:
 gem 'cmfrec'
 ```
 
+Not available for Windows yet
+
 ## Getting Started
 
 Create a recommender
@@ -48,7 +48,7 @@ recommender.fit([
 
 > Use `value` instead of rating for implicit feedback
 
-Get recommendations - “users like you also liked”
+Get recommendations for a user in the training data
 
 ```ruby
 recommender.user_recs(user_id)
@@ -69,10 +69,10 @@ Use the `count` option to specify the number of recommendations (default is 5)
 recommender.user_recs(user_id, count: 3)
 ```
 
-Get predicted ratings for specific items
+Get predicted ratings for specific users and items
 
 ```ruby
-recommender.user_recs(user_id, item_ids: [1, 2, 3])
+recommender.predict([{user_id: 1, item_id: 2}, {user_id: 2, item_id: 4}])
 ```
 
 ## Side Information
@@ -81,12 +81,12 @@ Add side information about users, items, or both
 
 ```ruby
 user_info = [
-  {user_id: 1, a: 1, b: 1},
-  {user_id: 2, a: 1, b: 1},
+  {user_id: 1, cats: 1, dogs: 0},
+  {user_id: 2, cats: 2, dogs: 1},
 ]
 item_info = [
-  {item_id: 1, c: 1, d: 1},
-  {item_id: 2, c: 1, d: 1},
+  {item_id: 1, genre_comedy: 1, genre_drama: 0},
+  {item_id: 2, genre_comedy: 0, genre_drama: 1},
 ]
 recommender.fit(ratings, user_info: user_info, item_info: item_info)
 ```
@@ -98,13 +98,13 @@ ratings = [
   {item_id: 1, rating: 5},
   {item_id: 2, rating: 3}
 ]
-recommender.new_user_recs(ratings, user_info: {a: 1, b: 1})
+recommender.new_user_recs(ratings, user_info: {cats: 0, dogs: 2})
 ```
 
 Get recommendations with only side information
 
 ```ruby
-recommender.new_user_recs([], user_info: {a: 1, b: 1})
+recommender.new_user_recs([], user_info: {cats: 0, dogs: 2})
 ```
 
 ## Options

data/lib/cmfrec/recommender.rb

@@ -172,6 +172,49 @@ module Cmfrec
       self
     end
 
+    def predict(data)
+      check_fit
+
+      data = to_dataset(data)
+
+      u = data.map { |v| @user_map[v[:user_id]] || -1 }
+      i = data.map { |v| @item_map[v[:item_id]] || -1 }
+
+      pred_a = int_ptr(u)
+      pred_b = int_ptr(i)
+      nnz = data.size
+      outp = Fiddle::Pointer.malloc(nnz * Fiddle::SIZEOF_DOUBLE)
+
+      FFI.predict_multiple(
+        @a, @k_user,
+        @b, @k_item,
+        @bias_a, @bias_b,
+        @global_mean,
+        @k, @k_main,
+        @m, @n,
+        pred_a, pred_b, nnz,
+        outp,
+        @nthreads
+      )
+
+      predictions = real_array(outp)
+
+      nan_index = predictions.each_index.select { |j| predictions[j].nan? }
+      if nan_index.any?
+        # TODO improve performance
+        user_bias = send(:user_bias)
+        item_bias = send(:item_bias)
+        nan_index.each do |j|
+          v = @global_mean
+          v += user_bias[u[j]] if user_bias && u[j] != -1
+          v += item_bias[i[j]] if item_bias && i[j] != -1
+          predictions[j] = v
+        end
+      end
+
+      predictions
+    end
+
     def user_recs(user_id, count: 5, item_ids: nil)
       check_fit
       user = @user_map[user_id]
@@ -181,24 +224,9 @@ module Cmfrec
           # remove missing ids
           item_ids = item_ids.select { |v| @item_map[v] }
 
-          pred_a = int_ptr([@user_map[user_id]] * item_ids.size)
-          pred_b = int_ptr(item_ids.map { |v| @item_map[v] })
-          nnz = item_ids.size
-          outp = Fiddle::Pointer.malloc(nnz * Fiddle::SIZEOF_DOUBLE)
-
-          FFI.predict_multiple(
-            @a, @k_user,
-            @b, @k_item,
-            @bias_a, @bias_b,
-            @global_mean,
-            @k, @k_main,
-            @m, @n,
-            pred_a, pred_b, nnz,
-            outp,
-            @nthreads
-          )
-
-          scores = real_array(outp)
+          data = item_ids.map { |v| {user_id: user_id, item_id: v} }
+          scores = predict(data)
+
           item_ids.zip(scores).map do |item_id, score|
             {item_id: item_id, score: score}
           end

data/lib/cmfrec/version.rb

@@ -1,3 +1,3 @@
 module Cmfrec
-  VERSION = "0.1.0"
+  VERSION = "0.1.1"
 end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: cmfrec
 version: !ruby/object:Gem::Version
-  version: 0.1.0
+  version: 0.1.1
 platform: ruby
 authors:
 - Andrew Kane