update fasttext to remove deprecated message

pommedeterresautee · Oct 27, 2019 · 7c1c7cd · 7c1c7cd
1 parent eaa4821
commit 7c1c7cd
Show file tree

Hide file tree

Showing 23 changed files with 900 additions and 182 deletions.
diff --git a/CRAN-RELEASE b/CRAN-RELEASE
@@ -1,2 +1,2 @@
-This package was submitted to CRAN on 2019-04-14.
-Once it is accepted, delete this file and tag the release (commit f935b7df52).
+This package was submitted to CRAN on 2019-05-30.
+Once it is accepted, delete this file and tag the release (commit eaa4821769).
diff --git a/DESCRIPTION b/DESCRIPTION
@@ -1,8 +1,8 @@
 Package: fastrtext
 Type: Package
 Title: 'fastText' Wrapper for Text Classification and Word Representation
-Version: 0.3.2
-Date: 2019-04-14
+Version: 0.3.4
+Date: 2019-10-27
 Authors@R: c(person("Michaël", "Benesty", role = c("aut", "cre", "cph"), email = "michael@benesty.fr"),
   person("Facebook, Inc", role = c("cph"), email = "bojanowski@fb.com"))
 Maintainer: Michaël Benesty <michael@benesty.fr>

diff --git a/NEWS.md b/NEWS.md
@@ -1,6 +1,9 @@
-# 0.3.3 ()
-
+# 0.3.3 (10/27/19)
+
+  * remove deprecated code to fix Cran warnings
+  * update to last FastText version
   * support one vs all loss
+  * less macro (remove possibility to stop learning with CTRL+C)
 
 # 0.3.2 (10.04.19)
 

diff --git a/cran-comments.md b/cran-comments.md
@@ -1,3 +1,10 @@
+## Version 0.3.4
+
+fix deprecated error messages.
+
+## Version 0.3.3
+small update
+
 ## Version 0.3.2
 Following Cran e-mail (from Prof Brian Ripley), remove strip-debug in Makevars file.
 Now the size of the package is > 10Mb and it generates a Warning regarding its size.

diff --git a/src/Makevars b/src/Makevars
@@ -8,7 +8,7 @@ PKG_CPPFLAGS = -pthread -include r_compliance.h -I$(PKGROOT)
 # pthread is used for multithreading by fastText
 PKG_LIBS = -pthread
 
-OBJECTS = add_prefix.o r_compliance.o $(PKGROOT)/args.o $(PKGROOT)/matrix.o $(PKGROOT)/dictionary.o $(PKGROOT)/loss.o $(PKGROOT)/productquantizer.o $(PKGROOT)/densematrix.o $(PKGROOT)/quantmatrix.o $(PKGROOT)/vector.o $(PKGROOT)/model.o $(PKGROOT)/utils.o $(PKGROOT)/meter.o $(PKGROOT)/fasttext.o $(PKGROOT)/main.o fastrtext.o RcppExports.o
+OBJECTS = add_prefix.o r_compliance.o $(PKGROOT)/autotune.o $(PKGROOT)/args.o $(PKGROOT)/matrix.o $(PKGROOT)/dictionary.o $(PKGROOT)/loss.o $(PKGROOT)/productquantizer.o $(PKGROOT)/densematrix.o $(PKGROOT)/quantmatrix.o $(PKGROOT)/vector.o $(PKGROOT)/model.o $(PKGROOT)/utils.o $(PKGROOT)/meter.o $(PKGROOT)/fasttext.o $(PKGROOT)/main.o fastrtext.o RcppExports.o
 
 # Reduce the size of the compiled library by removing unneeded debug information
 # Need to check if we are on Linux and if strip is installed

diff --git a/src/fastrtext.cpp b/src/fastrtext.cpp
@@ -283,7 +283,7 @@ class fastrtext{
     case loss_name::ova:
       return "one-vs-all";
     default:
-      stop("Unrecognized loss (ns / hs / softmax) name!");
+      stop("Unrecognized loss (ns / hs / softmax / ova / one-vs-all) name!");
     }
   }
 

diff --git a/src/fasttext/args.cc b/src/fasttext/args.cc
@@ -12,6 +12,7 @@
 
 #include <iostream>
 #include <stdexcept>
+#include <unordered_map>
 
 namespace fasttext {
 
@@ -36,12 +37,19 @@ Args::Args() {
   verbose = 2;
   pretrainedVectors = "";
   saveOutput = false;
+  seed = 0;
 
   qout = false;
   retrain = false;
   qnorm = false;
   cutoff = 0;
   dsub = 2;
+
+  autotuneValidationFile = "";
+  autotuneMetric = "f1";
+  autotunePredictions = 1;
+  autotuneDuration = 60 * 5; // 5 minutes
+  autotuneModelSize = "";
 }
 
 std::string Args::lossToString(loss_name ln) const {
@@ -78,6 +86,16 @@ std::string Args::modelToString(model_name mn) const {
   return "Unknown model name!"; // should never happen
 }
 
+std::string Args::metricToString(metric_name mn) const {
+  switch (mn) {
+    case metric_name::f1score:
+      return "f1score";
+    case metric_name::labelf1score:
+      return "labelf1score";
+  }
+  return "Unknown metric name!"; // should never happen
+}
+
 void Args::parseArgs(const std::vector<std::string>& args) {
   std::string command(args[1]);
   if (command == "supervised") {
@@ -97,6 +115,8 @@ void Args::parseArgs(const std::vector<std::string>& args) {
       exit(EXIT_FAILURE);
     }
     try {
+      setManual(args[ai].substr(1));
+
       if (args[ai] == "-h") {
         std::cerr << "Here is the help! Usage:" << std::endl;
         printHelp();
@@ -157,6 +177,8 @@ void Args::parseArgs(const std::vector<std::string>& args) {
       } else if (args[ai] == "-saveOutput") {
         saveOutput = true;
         ai--;
+      } else if (args[ai] == "-seed") {
+        seed = std::stoi(args.at(ai + 1));
       } else if (args[ai] == "-qnorm") {
         qnorm = true;
         ai--;
@@ -170,12 +192,24 @@ void Args::parseArgs(const std::vector<std::string>& args) {
         cutoff = std::stoi(args.at(ai + 1));
       } else if (args[ai] == "-dsub") {
         dsub = std::stoi(args.at(ai + 1));
+      } else if (args[ai] == "-autotune-validation") {
+        autotuneValidationFile = std::string(args.at(ai + 1));
+      } else if (args[ai] == "-autotune-metric") {
+        autotuneMetric = std::string(args.at(ai + 1));
+        getAutotuneMetric(); // throws exception if not able to parse
+        getAutotuneMetricLabel(); // throws exception if not able to parse
+      } else if (args[ai] == "-autotune-predictions") {
+        autotunePredictions = std::stoi(args.at(ai + 1));
+      } else if (args[ai] == "-autotune-duration") {
+        autotuneDuration = std::stoi(args.at(ai + 1));
+      } else if (args[ai] == "-autotune-modelsize") {
+        autotuneModelSize = std::string(args.at(ai + 1));
       } else {
         std::cerr << "Unknown argument: " << args[ai] << std::endl;
         printHelp();
         exit(EXIT_FAILURE);
       }
-    } catch (std::out_of_range&) {
+    } catch (std::out_of_range) {
       std::cerr << args[ai] << " is missing an argument" << std::endl;
       printHelp();
       exit(EXIT_FAILURE);
@@ -195,6 +229,7 @@ void Args::printHelp() {
   printBasicHelp();
   printDictionaryHelp();
   printTrainingHelp();
+  printAutotuneHelp();
   printQuantizationHelp();
 }
 
@@ -235,11 +270,27 @@ void Args::printTrainingHelp() {
       << "  -neg                number of negatives sampled [" << neg << "]\n"
       << "  -loss               loss function {ns, hs, softmax, one-vs-all} ["
       << lossToString(loss) << "]\n"
-      << "  -thread             number of threads [" << thread << "]\n"
+      << "  -thread             number of threads (set to 1 to ensure reproducible results) ["
+      << thread << "]\n"
       << "  -pretrainedVectors  pretrained word vectors for supervised learning ["
       << pretrainedVectors << "]\n"
       << "  -saveOutput         whether output params should be saved ["
-      << boolToString(saveOutput) << "]\n";
+      << boolToString(saveOutput) << "]\n"
+      << "  -seed               random generator seed  [" << seed << "]\n";
+}
+
+void Args::printAutotuneHelp() {
+  std::cerr
+      << "\nThe following arguments are for autotune:\n"
+      << "  -autotune-validation            validation file to be used for evaluation\n"
+      << "  -autotune-metric                metric objective {f1, f1:labelname} ["
+      << autotuneMetric << "]\n"
+      << "  -autotune-predictions           number of predictions used for evaluation  ["
+      << autotunePredictions << "]\n"
+      << "  -autotune-duration              maximum duration in seconds ["
+      << autotuneDuration << "]\n"
+      << "  -autotune-modelsize             constraint model file size ["
+      << autotuneModelSize << "] (empty = do not quantize)\n";
 }
 
 void Args::printQuantizationHelp() {
@@ -317,4 +368,74 @@ void Args::dump(std::ostream& out) const {
       << " " << t << std::endl;
 }
 
+bool Args::hasAutotune() const {
+  return !autotuneValidationFile.empty();
+}
+
+bool Args::isManual(const std::string& argName) const {
+  return (manualArgs_.count(argName) != 0);
+}
+
+void Args::setManual(const std::string& argName) {
+  manualArgs_.emplace(argName);
+}
+
+metric_name Args::getAutotuneMetric() const {
+  if (autotuneMetric.substr(0, 3) == "f1:") {
+    return metric_name::labelf1score;
+  } else if (autotuneMetric == "f1") {
+    return metric_name::f1score;
+  }
+  throw std::runtime_error("Unknown metric : " + autotuneMetric);
+}
+
+std::string Args::getAutotuneMetricLabel() const {
+  if (getAutotuneMetric() == metric_name::labelf1score) {
+    std::string label = autotuneMetric.substr(3);
+    if (label.empty()) {
+      throw std::runtime_error("Empty metric label : " + autotuneMetric);
+    }
+    return label;
+  }
+  return std::string();
+}
+
+int64_t Args::getAutotuneModelSize() const {
+  std::string modelSize = autotuneModelSize;
+  if (modelSize.empty()) {
+    return Args::kUnlimitedModelSize;
+  }
+  std::unordered_map<char, int> units = {
+      {'k', 1000},
+      {'K', 1000},
+      {'m', 1000000},
+      {'M', 1000000},
+      {'g', 1000000000},
+      {'G', 1000000000},
+  };
+  uint64_t multiplier = 1;
+  char lastCharacter = modelSize.back();
+  if (units.count(lastCharacter)) {
+    multiplier = units[lastCharacter];
+    modelSize = modelSize.substr(0, modelSize.size() - 1);
+  }
+  uint64_t size = 0;
+  size_t nonNumericCharacter = 0;
+  bool parseError = false;
+  try {
+    size = std::stol(modelSize, &nonNumericCharacter);
+  } catch (std::invalid_argument&) {
+    parseError = true;
+  }
+  if (!parseError && nonNumericCharacter != modelSize.size()) {
+    parseError = true;
+  }
+  if (parseError) {
+    throw std::invalid_argument(
+        "Unable to parse model size " + autotuneModelSize);
+  }
+
+  return size * multiplier;
+}
+
 } // namespace fasttext
diff --git a/src/fasttext/args.h b/src/fasttext/args.h
@@ -11,18 +11,21 @@
 #include <istream>
 #include <ostream>
 #include <string>
+#include <unordered_set>
 #include <vector>
 
 namespace fasttext {
 
 enum class model_name : int { cbow = 1, sg, sup };
 enum class loss_name : int { hs = 1, ns, softmax, ova };
+enum class metric_name : int { f1score = 1, labelf1score };
 
 class Args {
  protected:
-  std::string lossToString(loss_name) const;
   std::string boolToString(bool) const;
   std::string modelToString(model_name) const;
+  std::string metricToString(metric_name) const;
+  std::unordered_set<std::string> manualArgs_;
 
  public:
   Args();
@@ -48,21 +51,38 @@ class Args {
   int verbose;
   std::string pretrainedVectors;
   bool saveOutput;
+  int seed;
 
   bool qout;
   bool retrain;
   bool qnorm;
   size_t cutoff;
   size_t dsub;
 
+  std::string autotuneValidationFile;
+  std::string autotuneMetric;
+  int autotunePredictions;
+  int autotuneDuration;
+  std::string autotuneModelSize;
+
   void parseArgs(const std::vector<std::string>& args);
   void printHelp();
   void printBasicHelp();
   void printDictionaryHelp();
   void printTrainingHelp();
+  void printAutotuneHelp();
   void printQuantizationHelp();
   void save(std::ostream&);
   void load(std::istream&);
   void dump(std::ostream&) const;
+  bool hasAutotune() const;
+  bool isManual(const std::string& argName) const;
+  void setManual(const std::string& argName);
+  std::string lossToString(loss_name) const;
+  metric_name getAutotuneMetric() const;
+  std::string getAutotuneMetricLabel() const;
+  int64_t getAutotuneModelSize() const;
+
+  static constexpr double kUnlimitedModelSize = -1.0;
 };
 } // namespace fasttext