AudioWebService

A small service which can run the Sphinx latice demo, the ProsodyLab aligner and various Praat Scripts to detect utterances and syllables in any file which contains an audio track.

How to use

On the server

Install the module with: npm install fielddb-audio-service or by cloning this repository git clone https://github.com/FieldDB/AudioWebService.git

node audio-service.js &

Run the tests to see if your machine is set up:

npm test

Upload using curl or shell script

curl -F files=@$HOME/Documents/georgian/phrases/alo.mp3 
  -F files=@$HOME/Documents/georgian/phrases/ara.mp3 
  -F token=mytokengoeshere 
  -F username=testingupload 
  -F dbname=testingupload-firstcorpus 
  https://localhost:3184/upload/extract/utterances 

Upload using an HTML5 browser client

<form class="form-inline button-group" id="uploadAudioForTextGridform" enctype="multipart/form-data" action="{{audioServerUrl}}/upload/extract/utterances" method="post">
  <label>
    <span>Import long audio/video elicitation session(s) </span>
  </label>
  <div class="input-prepend">
    <span class="btn btn-default btn-file btn-info btn-mini">
      <span>
        <i class="icon-file"></i> 
        Choose file(s)
      </span>
      <input id="uploadAudioForTextGridformFiles" type="file" multiple="true" name="files" value="Audio/Video files to be imported"/>
    </span>
  </div>
  <div class="input-append">
    <button class="btn btn-info btn-mini" type="submit">
      <i class="icon-upload"></i>
      <span> Upload</span>
    </button>
  </div>
  <input class="hidden" type="text" name="token" value="{{audiouploadtoken}}"/>
  <input class="hidden" type="text" name="username" value="{{username}}"/>
  <input class="hidden" type="text" name="dbname" value="{{pouchname}}"/>
  <input class="hidden" type="text" name="returnTextGrid" value="true"/>
</form>

Upload using Javascript/jQuery/Backbone/AJAX browser client

In your code, you can also use jQuery or Backbone to perform the upload and do something with the resulting json.

(Backbone event)
"submit #uploadAudioForTextGridform": function(e) {
  if (e) {
    e.stopPropagation();
    e.preventDefault();
  }

  //get the action-url of the form
  var actionurl = e.currentTarget.action;
  var data = new FormData();
  jQuery.each($('#uploadAudioForTextGridformFiles')[0].files, function(i, file) {
    data.append(i, file);
  });
  data.append("token", "testinguploadtoken");
  data.append("pouchname", this.model.get("pouchname"));
  data.append("username", window.app.get("authentication").get("userPrivate").get("username"));
  data.append("returnTextGrid", true);
  this.model.get("audioVideo").reset();
  var self = this;
  $.ajax({
    url: actionurl,
    type: 'post',
    // dataType: 'json',
    cache: false,
    contentType: false,
    processData: false,
    data: data,
    success: function(results) {
      if (results && results.status === 200) {
        self.model.set("uploadDetails", results);
        self.model.set("files", results.files);
        self.model.set("status", "File(s) uploaded and utterances were extracted.");
        var messages = [];
        self.model.set("rawText","");
        /* Check for any textgrids which failed */
        for (var fileIndex = 0; fileIndex < results.files.length; fileIndex++) {
          if (results.files[fileIndex].textGridStatus >= 400) {
            console.log(results.files[fileIndex]);
            var instructions = instructions = results.files[fileIndex].textGridInfo;
            if(results.files[fileIndex].textGridStatus >= 500){
              instructions = " Please report this error to us at support@lingsync.org ";
            }
            messages.push("Generating the textgrid for " + results.files[fileIndex].fileBaseName + " seems to have failed. "+instructions);
          } else {
            self.model.addAudioVideoFile(audioUrl + "/" + self.model.get("pouchname") + "/" + results.files[fileIndex].fileBaseName + '.mp3');
            self.model.downloadTextGrid(results.files[fileIndex]);
          }
        }
        if (messages.length > 0) {
          self.model.set("status", messages.join(", "));
          $(self.el).find(".status").html(self.model.get("status"));
          window.appView.toastUser(messages.join(", "), "alert-danger", "Import:");
        }
      } else {
        console.log(results);
        var message = "Upload might have failed to complete processing on your file(s). Please report this error to us at support@lingsync.org ";
        self.model.set("status", message + ": " + JSON.stringify(results));
        window.appView.toastUser(message, "alert-danger", "Import:");
      }
      $(self.el).find(".status").html(self.model.get("status"));
    },
    error: function(response) {
      var reason = {};
      if (response && response.responseJSON) {
        reason = response.responseJSON;
      } else {
        var message = "Error contacting the server. ";
        if (response.status >= 500) {
          message = message + " Please report this error to us";
        } else if (response.status === 413) {
          message = message + " Your file is too big for upload, please try using FFMpeg to convert it to an mp3 for upload (you can still use your original video/audio in the app when the utterance chunking is done on an mp3.) ";
        } else {
          message = message + " Are you offline? If you are online and you still recieve this error, please report it to us: ";
        }
        reason = {
          status: response.status,
          userFriendlyErrors: [message + response.status]
        };
      }
      console.log(reason);
      if (reason && reason.userFriendlyErrors) {
        self.model.set("status", "Upload error: " + reason.userFriendlyErrors.join(" "));
        window.appView.toastUser(reason.userFriendlyErrors.join(" "), "alert-danger", "Import:");
        $(self.el).find(".status").html(self.model.get("status"));
      }
    }
  });
  this.model.set("status", "Contacting server...");
  $(this.el).find(".status").html(this.model.get("status"));
},

Upload using an Android client

HttpURLConnection urlConnection;
try {
  url = new URL(urlStringAuthenticationSession);
  urlConnection = (HttpURLConnection) url.openConnection();
  urlConnection.setRequestMethod("POST");
  urlConnection
  .setRequestProperty("Content-Type", "application/json");
  urlConnection.setDoInput(true);
  urlConnection.setDoOutput(true);
  urlConnection.connect();
} catch (MalformedURLException e) {
  e.printStackTrace();
  this.userFriendlyErrorMessage = "Problem determining which server to contact, please report this error.";
  return null;
} catch (ProtocolException e) {
  this.userFriendlyErrorMessage = "Problem using POST, please report this error.";
  e.printStackTrace();
  return null;
} catch (IOException e) {
  this.userFriendlyErrorMessage = "Problem opening connection to server, please report this error.";
  e.printStackTrace();
  return null;
}
JsonObject jsonParam = new JsonObject();
jsonParam.addProperty("token", token);
jsonParam.addProperty("username", username);
jsonParam.addProperty("dbname", dbname);
jsonParam.addProperty("returnTextGrid", returnTextGrid);

DataOutputStream printout;
try {
  printout = new DataOutputStream(urlConnection.getOutputStream());
  String jsonString = jsonParam.toString();
  Log.d(Config.TAG, jsonString);
  printout.write(jsonString.getBytes());
  printout.flush();
  printout.close();
} catch (IOException e) {
  e.printStackTrace();
  this.userFriendlyErrorMessage = "Problem writing to the server connection.";
  return null;
}
String JSONResponse = this.processResponse(url, urlConnection);

Formal Documentation

http://opensourcefieldlinguistics.github.io/FieldDB/

Examples

See the test for current examples.

Contributing

In lieu of a formal styleguide, take care to maintain the existing coding style. Add unit tests for any new or changed functionality. Lint and test your code using Jasmine Node.

npm test

Release History

• v0.1 Sept 16 2011 Audio upload and sphinx execution for Android client
• v1.56 May 26 2013 Run ProsodyLab Aligner
• v1.70 Aug 26 2013 Detect syllables using Praat
• v1.102.3 April 22 2014 Long audio import support
• v2.2.0 May 19 2014 Support for 1.5GB movies

License

Copyright (c) 2014 OpenSourceFieldLinguistics Contribs
Licensed under the Apache 2.0 license.