libmf 0.4.0 → 0.5.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: b264767ceb43c4a0407896e2bcf2373a47296e1e0580c65ec2ef08471589f661
-  data.tar.gz: addfb47cc44989620f847eac0392aef7431e596af08095ddd56ed227c4b99274
+  metadata.gz: d82762549216fa7e42bc21e9450aaa7f9a423350915f8d1c6ed158cea7b4520d
+  data.tar.gz: 820677028926afecac55c270d74e572b70ce19e22c2d16710fa2c514f77e7154
 SHA512:
-  metadata.gz: b22bcab875b5c2e8b7432c137f90b3ea37a8c16c3e64aa0fb4c15f824f8345a3105ca981a4ea1ea091396dcd0ce10320eab52f56e6ede75601ea2192e5bc61ac
-  data.tar.gz: 2e946522f35f77a46ab3880d74da38a0efdcaf432eabc1945d35af344d5cafb2c7712a3e0660352f8cf84fbd44af5c9c4e1cc6110ef5b908cd66c3393e124a99
+  metadata.gz: 1278c9ad499069cd1287bda40878ad9e61c2e8d039bea27c439fcc5b1bcb32bcd5136b9ffde6bb967ba39ea49fbd97d832f765a3b6d5b90a8bd51a81d26843b1
+  data.tar.gz: fac4621e932e5eb07b4652ed0f7336275833a63e37814e853aa407e2f344076dec201cbff49b1a0271a115a715e2fd43ac0bdb25abdbd262ddfa5bcb14774ad8

data/CHANGELOG.md

@@ -1,3 +1,10 @@
+## 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

data/LICENSE.txt

@@ -1,7 +1,7 @@
 BSD 3-Clause License
 
 Copyright (c) 2014-2015 The LIBMF Project.
-Copyright (c) 2019-2024 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

@@ -191,23 +191,6 @@ model.fit(train_set, eval_set: valid_set)
 puts model.rmse(valid_set)
 ```
 
-## Performance
-
-For performance, read data directly from files
-
-```ruby
-model.fit("train.txt", eval_set: "validate.txt")
-model.cv("train.txt")
-```
-
-Data should be in the format `row_index column_index value`:
-
-```txt
-0 0 5.0
-0 2 3.5
-1 1 4.0
-```
-
 ## Numo
 
 Get latent factors as Numo arrays

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?
 
@@ -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 = "0.4.0"
+  VERSION = "0.5.0"
 end

metadata

@@ -1,14 +1,13 @@
 --- !ruby/object:Gem::Specification
 name: libmf
 version: !ruby/object:Gem::Version
-  version: 0.4.0
+  version: 0.5.0
 platform: ruby
 authors:
 - Andrew Kane
-autorequire:
 bindir: bin
 cert_chain: []
-date: 2024-10-23 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: '3.1'
+      version: '3.3'
 required_rubygems_version: !ruby/object:Gem::Requirement
   requirements:
   - - ">="
     - !ruby/object:Gem::Version
       version: '0'
 requirements: []
-rubygems_version: 3.5.16
-signing_key:
+rubygems_version: 4.0.6
 specification_version: 4
 summary: Large-scale sparse matrix factorization for Ruby
 test_files: []