transcribe audio file with deepgram

ai-conversation/audio_capture.c

@@ -14,30 +14,109 @@ void print_pcm_state(snd_pcm_t *pcm_handle) {
     printf("State of pcm handle: %s\n", state_name);
 }
 
+struct file_data {
+    FILE *file;
+    size_t size;
+};
+
+struct response_data {
+    char *data;
+    size_t size;
+};
+
+size_t read_callback(void *ptr, size_t size, size_t nmemb, void *userdata) {
+    struct file_data *data = (struct file_data *)userdata;
+    size_t bytes_to_read = size * nmemb;
+    return fread(ptr, 1, bytes_to_read, data->file);
+}
+
+// Returns string that caller must free
+char* find_transcript(char* json_response)
+{
+  char* start = strstr(json_response, "\"transcript\":\"");
+  if (!start) return NULL;
+
+  start += 14; // Skip past "transcript":"
+  char* end = strchr(start, '"'); // Find closing quote
+  // Copy substring between start and end
+  size_t len = end - start;
+  char* result = malloc(len + 1);
+  strncpy(result, start, len);
+  result[len] = '\0';
+  return result;
+}
+
 size_t write_callback(void *contents, size_t size, size_t nmemb, void *userp) {
     size_t total_size = size * nmemb;
-    printf("%.*s", (int)total_size, (char*)contents);
+    struct response_data *response = (struct response_data *)userp;
+    
+    // Reallocate buffer to fit new data
+    response->data = realloc(response->data, response->size + total_size + 1);
+    if (response->data == NULL) {
+        return 0; // Error
+    }
+    
+    // Copy new data to buffer
+    memcpy(&response->data[response->size], contents, total_size);
+    response->size += total_size;
+    response->data[response->size] = '\0'; // Null terminate
+    
     return total_size;
 }
 
-void test_request() {
-  CURL *curl;
-  CURLcode res;
-  curl = curl_easy_init();
+// Returns transcript of audio
+char* transcribe(FILE* audio_file) {
+    CURL *curl;
+    CURLcode res;
+    struct curl_slist *headers = NULL;
+    
+    // Initialize response buffer
+    struct response_data response = {0};
+    response.data = malloc(1);
+    response.size = 0;
 
-  if (curl) {
-    curl_easy_setopt(curl, CURLOPT_URL, "https://www.google.com");
-    curl_easy_setopt(curl, CURLOPT_WRITEFUNCTION, write_callback);
+    // Get file size
+    fseek(audio_file, 0, SEEK_END);
+    long file_size = ftell(audio_file);
+    fseek(audio_file, 0, SEEK_SET);
 
-    res = curl_easy_perform(curl);
-    if (res != CURLE_OK)
-    {
-      fprintf(stderr, "curl failed: %s\n", curl_easy_strerror(res));
-      return;
+    struct file_data data;
+    data.file = audio_file;
+    data.size = file_size;
+
+    curl = curl_easy_init();
+    if (curl) {
+        // Set headers for Deepgram API
+        headers = curl_slist_append(headers, "Authorization: Token YOUR_DEEPGRAM_API_KEY");
+        headers = curl_slist_append(headers, "Content-Type: audio/*");
+
+        curl_easy_setopt(curl, CURLOPT_URL, "https://api.deepgram.com/v1/listen");
+        curl_easy_setopt(curl, CURLOPT_HTTPHEADER, headers);
+        curl_easy_setopt(curl, CURLOPT_POST, 1L);
+        curl_easy_setopt(curl, CURLOPT_READFUNCTION, read_callback);
+        curl_easy_setopt(curl, CURLOPT_READDATA, &data);
+        curl_easy_setopt(curl, CURLOPT_POSTFIELDSIZE, file_size);
+        curl_easy_setopt(curl, CURLOPT_WRITEFUNCTION, write_callback);
+        curl_easy_setopt(curl, CURLOPT_WRITEDATA, &response);
+
+        res = curl_easy_perform(curl);
+        if (res != CURLE_OK) {
+            fprintf(stderr, "curl failed: %s\n", curl_easy_strerror(res));
+            free(response.data);
+            curl_slist_free_all(headers);
+            curl_easy_cleanup(curl);
+            return NULL;
+        }
+
+        curl_slist_free_all(headers);
+        curl_easy_cleanup(curl);
     }
 
-    curl_easy_cleanup(curl);
-  }
+    // Parse transcript from complete response
+    char* transcript = find_transcript(response.data);
+    free(response.data);
+    
+    return transcript;
 }
 
 int main() {
@@ -150,9 +229,19 @@ int main() {
         periods_read++;
     }
 
-    // TODO: keep appending above frames into one buffer, and send to deepgram
-
     fclose(file);
+    
+    // Reopen file for reading and transcribe
+    file = fopen("output.pcm", "rb");
+    if (file != NULL) {
+        char* transcript = transcribe(file);
+        if (transcript) {
+            printf("Transcript: %s\n", transcript);
+            free(transcript);
+        }
+        fclose(file);
+    }
+    
     free((void *)buf);
     snd_pcm_close(pcm_handle);
     snd_pcm_hw_params_free(params);