From 8f00a64547f64a33f8f2f0c1d8035c39a8f36aa7 Mon Sep 17 00:00:00 2001
From: James Taracevicz <james.taracevicz@system1.com>
Date: Fri, 30 Jun 2023 19:41:41 -0700
Subject: [PATCH 1/2] feat: added language_classifier.cc

---
 binding.gyp                | 11 ++++++
 src/language_classifier.cc | 70 ++++++++++++++++++++++++++++++++++++++
 2 files changed, 81 insertions(+)
 create mode 100644 src/language_classifier.cc

diff --git a/binding.gyp b/binding.gyp
index 9426bcc..737675b 100644
--- a/binding.gyp
+++ b/binding.gyp
@@ -25,6 +25,17 @@
                 "<!(node -e \"require('nan')\")",
                 "/usr/local/include"
             ]
+        },
+        {
+            "target_name": "language_classifier",
+            "sources": [ "src/language_classifier.cc" ],
+            "libraries": [
+                "-lpostal", "-L/usr/local/lib"
+            ],
+            "include_dirs": [
+                "<!(node -e \"require('nan')\")",
+                "/usr/local/include"
+            ]
         }
     ]
 }
diff --git a/src/language_classifier.cc b/src/language_classifier.cc
new file mode 100644
index 0000000..22782e7
--- /dev/null
+++ b/src/language_classifier.cc
@@ -0,0 +1,70 @@
+#include <libpostal/libpostal.h>
+#include <nan.h>
+#include <string.h>
+
+void LanguageClassifier(const Nan::FunctionCallbackInfo<v8::Value>& info) {
+    v8::Isolate *isolate = info.GetIsolate();
+
+    if (info.Length() < 1) {
+        Nan::ThrowTypeError("Usage: language_classifier(text)");
+        return;
+    }
+
+    if (!info[0]->IsString()) {
+        Nan::ThrowTypeError("First argument must be a string");
+        return;
+    }
+
+    Nan::Utf8String text_utf8(info[0]);
+    char *text = *text_utf8;
+
+    if (text == NULL) {
+        Nan::ThrowTypeError("Could not convert first argument to string");
+        return;
+    }
+
+    libpostal_language_classifier_response_t *response = libpostal_classify_language(text);
+
+    if (response != NULL) {
+        v8::Local<v8::Array> lang_array = Nan::New<v8::Array>(response->num_languages);
+        for (size_t i = 0; i < response->num_languages; i++) {
+            const char* language = response->languages[i]; // Directly access the array
+            
+            v8::Local<v8::Object> lang_obj = Nan::New<v8::Object>();
+            Nan::Set(lang_obj, Nan::New("language").ToLocalChecked(), Nan::New(language).ToLocalChecked());
+            
+            Nan::Set(lang_array, i, lang_obj);
+        }
+        libpostal_language_classifier_response_destroy(response);
+        info.GetReturnValue().Set(lang_array);
+    }
+}
+
+void cleanup(void*) {
+    libpostal_teardown();
+    libpostal_teardown_language_classifier();
+}
+
+void init(v8::Local<v8::Object> exports) {
+    if (!libpostal_setup() || !libpostal_setup_language_classifier()) {
+        Nan::ThrowError("Could not load libpostal");
+        return;
+    }
+
+    v8::Local<v8::Context> context = exports->CreationContext();
+
+    exports->Set(
+        context, 
+        Nan::New("language_classifier").ToLocalChecked(),
+        Nan::New<v8::FunctionTemplate>(LanguageClassifier)->GetFunction(context).ToLocalChecked()
+    );
+
+    #if NODE_MAJOR_VERSION >= 12
+        node::Environment* env = node::GetCurrentEnvironment(Nan::GetCurrentContext());
+        node::AtExit(env, cleanup, NULL);
+    #else
+        node::AtExit(cleanup);
+    #endif
+}
+
+NODE_MODULE(language_classifier, init)

From c4e4a4675b3125c210d0baebce0d09e8457e2272 Mon Sep 17 00:00:00 2001
From: James Taracevicz <james.taracevicz@system1.com>
Date: Fri, 30 Jun 2023 19:51:06 -0700
Subject: [PATCH 2/2] feat: included probability in language_classifier

---
 src/language_classifier.cc | 7 +++++--
 1 file changed, 5 insertions(+), 2 deletions(-)

diff --git a/src/language_classifier.cc b/src/language_classifier.cc
index 22782e7..45082d1 100644
--- a/src/language_classifier.cc
+++ b/src/language_classifier.cc
@@ -27,11 +27,14 @@ void LanguageClassifier(const Nan::FunctionCallbackInfo<v8::Value>& info) {
 
     if (response != NULL) {
         v8::Local<v8::Array> lang_array = Nan::New<v8::Array>(response->num_languages);
-        for (size_t i = 0; i < response->num_languages; i++) {
-            const char* language = response->languages[i]; // Directly access the array
+
+        for (size_t i = 0; i < response->num_languages; ++i) {
+            const char *language = response->languages[i];
+            const double probability = response->probs[i];
             
             v8::Local<v8::Object> lang_obj = Nan::New<v8::Object>();
             Nan::Set(lang_obj, Nan::New("language").ToLocalChecked(), Nan::New(language).ToLocalChecked());
+            Nan::Set(lang_obj, Nan::New("probability").ToLocalChecked(), Nan::New(probability));
             
             Nan::Set(lang_array, i, lang_obj);
         }