Explicitly export library symbols

.gitignore

@@ -52,3 +52,8 @@ wheels/
 *.exe
 *.out
 *.app
+
+# Dev environment
+.vscode
+.vs
+CMakeSettings.json

CMakeLists.txt

@@ -56,6 +56,7 @@ target_compile_definitions(
   flashlight-text
   PUBLIC
   FL_TEXT_USE_KENLM=$<BOOL:${FL_TEXT_USE_KENLM}>
+  FL_TEXT_DLL
 )
 
 

flashlight/lib/text/Defines.h

@@ -0,0 +1,23 @@
+/*
+ * Copyright (c) Facebook, Inc. and its affiliates.
+ *
+ * This source code is licensed under the MIT-style license found in the
+ * LICENSE file in the root directory of this source tree.
+ */
+
+#pragma once
+
+#if defined(_WIN32) || defined(_MSC_VER)
+
+#ifdef FL_TEXT_DLL
+#define FL_TEXT_API __declspec(dllexport)
+#else // FL_TEXT_DLL
+#define FL_TEXT_API __declspec(dllimport)
+#endif // FL_TEXT_DLL
+
+#else // defined(_WIN32) || defined(_MSC_VER)
+
+#define FL_TEXT_API __attribute__((visibility("default")))
+#define FL_DEPRECATED(msg) __attribute__((deprecated(msg)))
+
+#endif // defined(_WIN32) || defined(_MSC_VER)

flashlight/lib/text/String.h

@@ -17,6 +17,8 @@
 #include <unordered_map>
 #include <vector>
 
+#include "flashlight/lib/text/Defines.h"
+
 namespace fl {
 namespace lib {
 
@@ -32,37 +34,39 @@ using EnableIfSame = typename std::enable_if<std::is_same<S, T>::value>::type;
 // ================================== Functions
 // ==================================
 
-std::string trim(const std::string& str);
+FL_TEXT_API std::string trim(const std::string& str);
 
-void replaceAll(
-    std::string& str,
-    const std::string& from,
-    const std::string& repl);
+FL_TEXT_API void
+replaceAll(std::string& str, const std::string& from, const std::string& repl);
 
-bool startsWith(const std::string& input, const std::string& pattern);
-bool endsWith(const std::string& input, const std::string& pattern);
+FL_TEXT_API bool startsWith(
+    const std::string& input,
+    const std::string& pattern);
+FL_TEXT_API bool endsWith(const std::string& input, const std::string& pattern);
 
-std::vector<std::string>
+FL_TEXT_API std::vector<std::string>
 split(char delim, const std::string& input, bool ignoreEmpty = false);
 
-std::vector<std::string> split(
+FL_TEXT_API std::vector<std::string> split(
     const std::string& delim,
     const std::string& input,
     bool ignoreEmpty = false);
 
-std::vector<std::string> splitOnAnyOf(
+FL_TEXT_API std::vector<std::string> splitOnAnyOf(
     const std::string& delim,
     const std::string& input,
     bool ignoreEmpty = false);
 
-std::vector<std::string> splitOnWhitespace(
+FL_TEXT_API std::vector<std::string> splitOnWhitespace(
     const std::string& input,
     bool ignoreEmpty = false);
 
 /**
  * Join a vector of `std::string` inserting `delim` in between.
  */
-std::string join(const std::string& delim, const std::vector<std::string>& vec);
+FL_TEXT_API std::string join(
+    const std::string& delim,
+    const std::vector<std::string>& vec);
 
 /**
  * Join a range of `std::string` specified by iterators.
@@ -126,5 +130,6 @@ void dedup(std::vector<T>& in) {
   auto it = std::unique(in.begin(), in.end());
   in.resize(std::distance(in.begin(), it));
 }
+
 } // namespace lib
 } // namespace fl

flashlight/lib/text/decoder/Decoder.h

@@ -72,6 +72,7 @@ class Decoder {
   /* Get all the final hypothesis */
   virtual std::vector<DecodeResult> getAllFinalHypothesis() const = 0;
 };
+
 } // namespace text
 } // namespace lib
 } // namespace fl

flashlight/lib/text/decoder/LexiconDecoder.h

@@ -9,6 +9,7 @@
 
 #include <unordered_map>
 
+#include "flashlight/lib/text/Defines.h"
 #include "flashlight/lib/text/decoder/Decoder.h"
 #include "flashlight/lib/text/decoder/Trie.h"
 #include "flashlight/lib/text/decoder/lm/LM.h"
@@ -111,7 +112,7 @@ struct LexiconDecoderState {
  * search space and all candidate words are generated from it if unkScore is
  * -inf, otherwise <UNK> will be generated for OOVs.
  */
-class LexiconDecoder : public Decoder {
+class FL_TEXT_API LexiconDecoder : public Decoder {
  public:
   LexiconDecoder(
       LexiconDecoderOptions opt,
@@ -182,6 +183,7 @@ class LexiconDecoder : public Decoder {
   int nDecodedFrames_; // Total number of decoded frames.
   int nPrunedFrames_; // Total number of pruned frames from hyp_.
 };
+
 } // namespace text
 } // namespace lib
 } // namespace fl

flashlight/lib/text/decoder/LexiconFreeDecoder.h

@@ -9,6 +9,7 @@
 
 #include <unordered_map>
 
+#include "flashlight/lib/text/Defines.h"
 #include "flashlight/lib/text/decoder/Decoder.h"
 #include "flashlight/lib/text/decoder/lm/LM.h"
 
@@ -96,7 +97,7 @@ struct LexiconFreeDecoderState {
  * score of the transcription W. We are allowed to generate words from all the
  * possible combination of tokens.
  */
-class LexiconFreeDecoder : public Decoder {
+class FL_TEXT_API LexiconFreeDecoder : public Decoder {
  public:
   LexiconFreeDecoder(
       LexiconFreeDecoderOptions opt,
@@ -165,6 +166,7 @@ class LexiconFreeDecoder : public Decoder {
   int nDecodedFrames_; // Total number of decoded frames.
   int nPrunedFrames_; // Total number of pruned frames from hyp_.
 };
+
 } // namespace text
 } // namespace lib
 } // namespace fl

flashlight/lib/text/decoder/LexiconFreeSeq2SeqDecoder.h

@@ -11,6 +11,7 @@
 #include <optional>
 #include <unordered_map>
 
+#include "flashlight/lib/text/Defines.h"
 #include "flashlight/lib/text/decoder/Decoder.h"
 #include "flashlight/lib/text/decoder/Utils.h"
 #include "flashlight/lib/text/decoder/lm/LM.h"
@@ -93,7 +94,7 @@ struct LexiconFreeSeq2SeqDecoderState {
  * TODO: Doesn't support online decoding now.
  *
  */
-class LexiconFreeSeq2SeqDecoder : public Decoder {
+class FL_TEXT_API LexiconFreeSeq2SeqDecoder : public Decoder {
  public:
   LexiconFreeSeq2SeqDecoder(
       LexiconFreeSeq2SeqDecoderOptions opt,
@@ -136,6 +137,7 @@ class LexiconFreeSeq2SeqDecoder : public Decoder {
 
   std::unordered_map<int, std::vector<LexiconFreeSeq2SeqDecoderState>> hyp_;
 };
+
 } // namespace text
 } // namespace lib
 } // namespace fl

flashlight/lib/text/decoder/LexiconSeq2SeqDecoder.h

@@ -11,6 +11,7 @@
 #include <optional>
 #include <unordered_map>
 
+#include "flashlight/lib/text/Defines.h"
 #include "flashlight/lib/text/decoder/Decoder.h"
 #include "flashlight/lib/text/decoder/Trie.h"
 #include "flashlight/lib/text/decoder/Utils.h"
@@ -111,7 +112,7 @@ struct LexiconSeq2SeqDecoderState {
  * TODO: Doesn't support online decoding now.
  *
  */
-class LexiconSeq2SeqDecoder : public Decoder {
+class FL_TEXT_API LexiconSeq2SeqDecoder : public Decoder {
  public:
   LexiconSeq2SeqDecoder(
       LexiconSeq2SeqDecoderOptions opt,
@@ -160,6 +161,7 @@ class LexiconSeq2SeqDecoder : public Decoder {
 
   std::unordered_map<int, std::vector<LexiconSeq2SeqDecoderState>> hyp_;
 };
+
 } // namespace text
 } // namespace lib
 } // namespace fl

flashlight/lib/text/decoder/Trie.h

@@ -10,6 +10,8 @@
 #include <unordered_map>
 #include <vector>
 
+#include "flashlight/lib/text/Defines.h"
+
 namespace fl {
 namespace lib {
 namespace text {
@@ -59,7 +61,7 @@ using TrieNodePtr = std::shared_ptr<TrieNode>;
  * the search space in deocder and quickly look up scores for a given token
  * (completed word) or make prediction for incompleted ones based on smearing.
  */
-class Trie {
+class FL_TEXT_API Trie {
  public:
   Trie(int maxChildren, int rootIdx)
       : root_(std::make_shared<TrieNode>(rootIdx)), maxChildren_(maxChildren) {}
@@ -90,6 +92,7 @@ class Trie {
 };
 
 using TriePtr = std::shared_ptr<Trie>;
+
 } // namespace text
 } // namespace lib
 } // namespace fl

flashlight/lib/text/decoder/Utils.cpp

@@ -7,9 +7,6 @@
 
 namespace fl {
 namespace lib {
-namespace text {
-
-// Place holder
-} // namespace text
+namespace text {} // namespace text
 } // namespace lib
 } // namespace fl

flashlight/lib/text/decoder/lm/ConvLM.cpp

@@ -234,6 +234,7 @@ void ConvLM::updateCache(std::vector<LMStatePtr> states) {
     }
   }
 }
+
 } // namespace text
 } // namespace lib
 } // namespace fl

flashlight/lib/text/decoder/lm/ConvLM.h

@@ -9,6 +9,7 @@
 
 #include <functional>
 
+#include "flashlight/lib/text/Defines.h"
 #include "flashlight/lib/text/decoder/lm/LM.h"
 #include "flashlight/lib/text/dictionary/Defines.h"
 #include "flashlight/lib/text/dictionary/Dictionary.h"
@@ -29,7 +30,7 @@ struct ConvLMState : LMState {
       : tokens(std::vector<int>(size)), length(size) {}
 };
 
-class ConvLM : public LM {
+class FL_TEXT_API ConvLM : public LM {
  public:
   ConvLM(
       const GetConvLmScoreFunc& getConvLmScoreFunc,
@@ -68,6 +69,7 @@ class ConvLM : public LM {
       const LMStatePtr& state,
       const int tokenIdx);
 };
+
 } // namespace text
 } // namespace lib
 } // namespace fl

flashlight/lib/text/decoder/lm/KenLM.cpp

@@ -11,12 +11,12 @@
 
 #ifdef USE_KENLM_FROM_LANGTECH
 #include "language_technology/jedi/lm/model.hh"
-#else
+#else // USE_KENLM_FROM_LANGTECH
 #if __has_include(<kenlm/lm/model.hh>)
 #include <kenlm/lm/model.hh>
 #elif __has_include(<lm/model.hh>)
 #include <lm/model.hh>
-#else
+#else // __has_include
 #error "KenLM header not found (kenlm/lm/model.hh)."
 #endif // __has_include
 #endif // USE_KENLM_FROM_LANGTECH
@@ -27,6 +27,8 @@ namespace text {
 
 KenLMState::KenLMState() : ken_(std::make_unique<lm::ngram::State>()) {}
 
+KenLMState::~KenLMState() {}
+
 KenLM::KenLM(const std::string& path, const Dictionary& usrTknDict) {
   // Load LM
   model_.reset(lm::ngram::LoadVirtual(path.c_str()));
@@ -79,6 +81,7 @@ std::pair<LMStatePtr, float> KenLM::finish(const LMStatePtr& state) {
       model_->BaseScore(inState->ken(), vocab_->EndSentence(), outState->ken());
   return std::make_pair(std::move(outState), score);
 }
+
 } // namespace text
 } // namespace lib
 } // namespace fl

flashlight/lib/text/decoder/lm/KenLM.h

@@ -9,6 +9,7 @@
 
 #include <memory>
 
+#include "flashlight/lib/text/Defines.h"
 #include "flashlight/lib/text/decoder/lm/LM.h"
 #include "flashlight/lib/text/dictionary/Dictionary.h"
 
@@ -36,8 +37,9 @@ namespace text {
  * indicies and compare functions
  * https://github.com/kpu/kenlm/blob/master/lm/state.hh.
  */
-struct KenLMState : LMState {
+struct FL_TEXT_API KenLMState : LMState {
   KenLMState();
+  ~KenLMState();
   std::unique_ptr<lm::ngram::State> ken_;
   lm::ngram::State* ken() {
     return ken_.get();
@@ -47,7 +49,7 @@ struct KenLMState : LMState {
 /**
  * KenLM extends LM by using the toolkit https://kheafield.com/code/kenlm/.
  */
-class KenLM : public LM {
+class FL_TEXT_API KenLM : public LM {
  public:
   KenLM(const std::string& path, const Dictionary& usrTknDict);
 
@@ -65,6 +67,7 @@ class KenLM : public LM {
 };
 
 using KenLMPtr = std::shared_ptr<KenLM>;
+
 } // namespace text
 } // namespace lib
 } // namespace fl

flashlight/lib/text/decoder/lm/ZeroLM.h

@@ -7,6 +7,7 @@
 
 #pragma once
 
+#include "flashlight/lib/text/Defines.h"
 #include "flashlight/lib/text/decoder/lm/LM.h"
 
 namespace fl {
@@ -17,7 +18,7 @@ namespace text {
  * ZeroLM is a dummy language model class, which mimics the behavious of a
  * uni-gram language model but always returns 0 as score.
  */
-class ZeroLM : public LM {
+class FL_TEXT_API ZeroLM : public LM {
  public:
   LMStatePtr start(bool startWithNothing) override;
 

flashlight/lib/text/dictionary/Dictionary.h

@@ -12,12 +12,15 @@
 #include <unordered_map>
 #include <vector>
 
+#include "flashlight/lib/text/Defines.h"
+
 namespace fl {
 namespace lib {
 namespace text {
+
 // A simple dictionary class which holds a bidirectional map
 // entry (strings) <--> integer indices. Not thread-safe !
-class Dictionary {
+class FL_TEXT_API Dictionary {
  public:
   // Creates an empty dictionary
   Dictionary() {}