libmf 0.3.0 → 0.5.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 510156ab4bccb39a99441002dd80f3540b643934294bc5724dca91987b4effba
-  data.tar.gz: 03a1f04e4679ec60cb2f1d2a96e2add0b9adba03fe5715bc4f10309ab795bcc1
+  metadata.gz: d82762549216fa7e42bc21e9450aaa7f9a423350915f8d1c6ed158cea7b4520d
+  data.tar.gz: 820677028926afecac55c270d74e572b70ce19e22c2d16710fa2c514f77e7154
 SHA512:
-  metadata.gz: eb862c0cf91a077ba2e9a0ed3f93491297cdc9801683e3587b91125c7e0b79123d0f942aea6bf3cef42fbbb43f5e8f822758a894f25a7eedf2fd9ee7b5ea4920
-  data.tar.gz: b4d74a6b3d4160f3b1165f90496a9f93c6b59ca99e5f3cc0094f022a0807fc8a34af3d9b0edeeff463264a60c19901ceaa171c3cea944d860103208e2d26dbf2
+  metadata.gz: 1278c9ad499069cd1287bda40878ad9e61c2e8d039bea27c439fcc5b1bcb32bcd5136b9ffde6bb967ba39ea49fbd97d832f765a3b6d5b90a8bd51a81d26843b1
+  data.tar.gz: fac4621e932e5eb07b4652ed0f7336275833a63e37814e853aa407e2f344076dec201cbff49b1a0271a115a715e2fd43ac0bdb25abdbd262ddfa5bcb14774ad8

data/CHANGELOG.md

@@ -1,3 +1,14 @@
+## 0.5.0 (2026-04-17)
+
+- Added checks for invalid data
+- Fixed memory leak in shared library
+- Dropped support for reading data directly from files
+- Dropped support for Ruby < 3.3
+
+## 0.4.0 (2024-10-22)
+
+- Dropped support for Ruby < 3.1
+
 ## 0.3.0 (2022-08-07)
 
 - Added metrics

data/LICENSE.txt

@@ -1,7 +1,7 @@
 BSD 3-Clause License
 
 Copyright (c) 2014-2015 The LIBMF Project.
-Copyright (c) 2019-2022 Andrew Kane.
+Copyright (c) 2019-2026 Andrew Kane.
 All rights reserved.
 
 Redistribution and use in source and binary forms, with or without

data/README.md

@@ -4,7 +4,7 @@
 
 Check out [Disco](https://github.com/ankane/disco) for higher-level collaborative filtering
 
-[![Build Status](https://github.com/ankane/libmf-ruby/workflows/build/badge.svg?branch=master)](https://github.com/ankane/libmf-ruby/actions)
+[![Build Status](https://github.com/ankane/libmf-ruby/actions/workflows/build.yml/badge.svg)](https://github.com/ankane/libmf-ruby/actions)
 
 ## Installation
 
@@ -170,21 +170,25 @@ Calculate AUC (for one-class MF)
 model.auc(data, transpose)
 ```
 
-## Performance
+## Example
 
-For performance, read data directly from files
+Download the [MovieLens 100K dataset](https://grouplens.org/datasets/movielens/100k/) and use:
 
 ```ruby
-model.fit("train.txt", eval_set: "validate.txt")
-model.cv("train.txt")
-```
+require "csv"
+
+train_set = Libmf::Matrix.new
+valid_set = Libmf::Matrix.new
+
+CSV.foreach("u.data", col_sep: "\t").with_index do |row, i|
+  data = i < 80000 ? train_set : valid_set
+  data.push(row[0].to_i, row[1].to_i, row[2].to_f)
+end
 
-Data should be in the format `row_index column_index value`:
+model = Libmf::Model.new(factors: 20)
+model.fit(train_set, eval_set: valid_set)
 
-```txt
-0 0 5.0
-0 2 3.5
-1 1 4.0
+puts model.rmse(valid_set)
 ```
 
 ## Numo

data/lib/libmf/ffi.rb

@@ -53,7 +53,6 @@ module Libmf
     end
 
     attach_function :mf_get_default_param, [], Parameter.by_value
-    attach_function :mf_read_problem, [:string], Problem.by_value
     attach_function :mf_save_model, [Model.by_ref, :string], :int
     attach_function :mf_load_model, [:string], Model.auto_ptr
     attach_function :mf_destroy_model, [:pointer], :void

data/lib/libmf/model.rb

@@ -10,9 +10,22 @@ module Libmf
       @model =
         if eval_set
           eval_set = create_problem(eval_set)
+          param = self.param
+
+          # LIBMF does not handle these cases
+          if param[:fun] == 12
+            if eval_set[:m] > train_set[:m]
+              raise ArgumentError, "Eval set cannot have extra rows for one_class_l2 loss"
+            end
+
+            if eval_set[:n] > train_set[:n]
+              raise ArgumentError, "Eval set cannot have extra columns for one_class_l2 loss"
+            end
+          end
+
           FFI.mf_train_with_validation(train_set, eval_set, param)
         else
-          FFI.mf_train(train_set, param)
+          FFI.mf_train(train_set, self.param)
         end
       raise Error, "fit failed" if @model.null?
 
@@ -141,13 +154,13 @@ module Libmf
       options[:bins] ||= 25 unless options[:nr_bins]
       options[:copy_data] = false unless options.key?(:copy_data)
       options_map = {
-        :loss => :fun,
-        :factors => :k,
-        :threads => :nr_threads,
-        :bins => :nr_bins,
-        :iterations => :nr_iters,
-        :learning_rate => :eta,
-        :nmf => :do_nmf
+        loss: :fun,
+        factors: :k,
+        threads: :nr_threads,
+        bins: :nr_bins,
+        iterations: :nr_iters,
+        learning_rate: :eta,
+        nmf: :do_nmf
       }
       options.each do |k, v|
         k = options_map[k] if options_map[k]
@@ -160,8 +173,7 @@ module Libmf
 
     def create_problem(data)
       if data.is_a?(String)
-        # need to expand path so it's absolute
-        return FFI.mf_read_problem(File.expand_path(data))
+        raise Error, "Reading data directly from files is no longer supported"
       end
 
       if data.is_a?(Matrix)
@@ -175,7 +187,23 @@ module Libmf
       # and write directly to C string
       buffer = String.new
       pack_format = "iif"
+      int_max = 2**31 - 1
+      umax = -1
+      vmax = -1
       data.each do |row|
+        u = row[0]
+        if u < 0 || u >= int_max
+          raise ArgumentError, "Invalid row index"
+        end
+
+        v = row[1]
+        if v < 0 || v >= int_max
+          raise ArgumentError, "Invalid column index"
+        end
+
+        umax = u if u > umax
+        vmax = v if v > vmax
+
         row.pack(pack_format, buffer: buffer)
       end
 
@@ -186,12 +214,9 @@ module Libmf
       # FFI will throw an error above if too long
       raise Error, "Bad buffer size" if r.size != buffer.bytesize
 
-      m = data.max_by { |r| r[0] }[0] + 1
-      n = data.max_by { |r| r[1] }[1] + 1
-
       prob = FFI::Problem.new
-      prob[:m] = m
-      prob[:n] = n
+      prob[:m] = umax + 1
+      prob[:n] = vmax + 1
       prob[:nnz] = data.size
       prob[:r] = r
       prob

data/lib/libmf/version.rb

@@ -1,3 +1,3 @@
 module Libmf
-  VERSION = "0.3.0"
+  VERSION = "0.5.0"
 end

data/lib/libmf.rb

@@ -2,9 +2,9 @@
 require "ffi"
 
 # modules
-require "libmf/matrix"
-require "libmf/model"
-require "libmf/version"
+require_relative "libmf/matrix"
+require_relative "libmf/model"
+require_relative "libmf/version"
 
 module Libmf
   class Error < StandardError; end

metadata

@@ -1,14 +1,13 @@
 --- !ruby/object:Gem::Specification
 name: libmf
 version: !ruby/object:Gem::Version
-  version: 0.3.0
+  version: 0.5.0
 platform: ruby
 authors:
 - Andrew Kane
-autorequire:
 bindir: bin
 cert_chain: []
-date: 2022-08-08 00:00:00.000000000 Z
+date: 1980-01-02 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: ffi
@@ -24,7 +23,6 @@ dependencies:
     - - ">="
       - !ruby/object:Gem::Version
         version: '0'
-description:
 email: andrew@ankane.org
 executables: []
 extensions: []
@@ -50,7 +48,6 @@ homepage: https://github.com/ankane/libmf-ruby
 licenses:
 - BSD-3-Clause
 metadata: {}
-post_install_message:
 rdoc_options: []
 require_paths:
 - lib
@@ -58,15 +55,14 @@ required_ruby_version: !ruby/object:Gem::Requirement
   requirements:
   - - ">="
     - !ruby/object:Gem::Version
-      version: '2.7'
+      version: '3.3'
 required_rubygems_version: !ruby/object:Gem::Requirement
   requirements:
   - - ">="
     - !ruby/object:Gem::Version
       version: '0'
 requirements: []
-rubygems_version: 3.3.7
-signing_key:
+rubygems_version: 4.0.6
 specification_version: 4
 summary: Large-scale sparse matrix factorization for Ruby
 test_files: []