Merge pull request #7 from grain-team/ryan/support-voice-spans

README.md

@@ -35,16 +35,33 @@ iex> vtt = """
            Hello world!
            """
 ...> Vttyl.parse(vtt) |> Enum.into([])
-[%Vttyl.Part{end: 17609, part: 1, start: 15450, text: "Hello world!"}]
+[%Vttyl.Part{end: 17609, part: 1, start: 15450, text: "Hello world!", voice: nil}]
 ```
 
 #### Stream Parsing
 
 ```elixir
 iex> "same_text.vtt" |> File.stream!([], 2048) |> Vttyl.parse_stream() |> Enum.into([])
-[%Vttyl.Part{end: 17609, part: 1, start: 15450, text: "Hello world!"}]
+[%Vttyl.Part{end: 17609, part: 1, start: 15450, text: "Hello world!", voice: nil}]
 ```
 
+#### Simple Voice Spans
+
+(Closing voice spans are currently not supported)
+
+```elixir
+iex> vtt = """
+           WEBVTT
+
+           1
+           00:00:15.450 --> 00:00:17.609
+           <v Andy>Hello world!
+           """
+...> Vttyl.parse(vtt) |> Enum.into([])
+[%Vttyl.Part{end: 17609, part: 1, start: 15450, text: "Hello world!", voice: "Andy"}]
+```
+
+
 ### Encoding
 
 Vttyl also supports encoding parts.
@@ -60,6 +77,17 @@ Hello world!
 """
 ```
 
+```elixir
+iex> parts = [%Vttyl.Part{end: 17609, part: 1, start: 15450, text: "Hello world!", voice: "Andy"}]
+...> Vttyle.encode(parts)
+"""
+WEBVTT
+1
+00:00:15.450 --> 00:00:17.609
+<v Andy>Hello world!
+"""
+```
+
 ## License
 
 Vttyl is Copyright © 2019 Grain Intelligence, Inc. It is free software, and may be

config/config.exs

@@ -1,30 +1 @@
-# This file is responsible for configuring your application
-# and its dependencies with the aid of the Mix.Config module.
-use Mix.Config
-
-# This configuration is loaded before any dependency and is restricted
-# to this project. If another project depends on this project, this
-# file won't be loaded nor affect the parent project. For this reason,
-# if you want to provide default values for your application for
-# third-party users, it should be done in your "mix.exs" file.
-
-# You can configure your application as:
-#
-#     config :vttyl, key: :value
-#
-# and access this configuration in your application as:
-#
-#     Application.get_env(:vttyl, :key)
-#
-# You can also configure a third-party app:
-#
-#     config :logger, level: :info
-#
-
-# It is also possible to import configuration files, relative to this
-# directory. For example, you can emulate configuration per environment
-# by uncommenting the line below and defining dev.exs, test.exs and such.
-# Configuration from the imported file will override the ones defined
-# here (which is why it is important to import them last).
-#
-#     import_config "#{Mix.env()}.exs"
+import Config

lib/vttyl/decode.ex

@@ -25,7 +25,8 @@ defmodule Vttyl.Decode do
 
         # Text content should be on one line and the other stuff should have appeared
         not is_nil(acc.part) and not is_nil(acc.start) and not is_nil(acc.end) and line != "" ->
-          %Part{acc | text: line}
+          {voice, text} = parse_text(line)
+          %Part{acc | text: text, voice: voice}
 
         true ->
           acc
@@ -59,6 +60,15 @@ defmodule Vttyl.Decode do
     Regex.match?(@line_regex, line)
   end
 
+  @annotation_space_regex ~r/[ \t]/
+  defp parse_text("<v" <> line) do
+    [voice, text] = String.split(line, ">", parts: 2)
+    [_, voice] = String.split(voice, @annotation_space_regex, parts: 2)
+    {voice, text}
+  end
+
+  defp parse_text(text), do: {nil, text}
+
   defp parse_timestamps(line) do
     line
     |> String.split("-->")

lib/vttyl/encode.ex

@@ -1,10 +1,19 @@
 defmodule Vttyl.Encode do
   @moduledoc false
+  alias Vttyl.Part
 
   @spec encode_part(Part.t(), :vtt | :srt) :: String.t()
   def encode_part(part, type) do
     ts = fmt_timestamp(part.start, type) <> " --> " <> fmt_timestamp(part.end, type)
-    Enum.join([part.part, ts, part.text], "\n")
+
+    text =
+      if type == :vtt && part.voice do
+        "<v #{part.voice}>" <> part.text
+      else
+        part.text
+      end
+
+    Enum.join([part.part, ts, text], "\n")
   end
 
   @hour_ms 3_600_000

lib/vttyl/part.ex

@@ -3,7 +3,8 @@ defmodule Vttyl.Part do
           part: non_neg_integer(),
           start: millisecond :: integer,
           end: millisecond :: integer,
-          text: String.t()
+          text: String.t(),
+          voice: String.t() | nil
         }
-  defstruct [:start, :end, :text, :part]
+  defstruct [:start, :end, :text, :part, :voice]
 end

priv/samples/voice.vtt

@@ -0,0 +1,17 @@
+WEBVTT
+
+1
+00:00.000 --> 00:02.000
+<v.first.loud Esme>It’s a blue apple tree!
+
+2
+00:02.000 --> 00:04.000
+<v	Mary>No way!
+
+3
+00:04.000 --> 00:06.000
+<v Esme F>Hee!
+
+4
+00:06.000 --> 00:08.000
+<v.loud Mary>That’s awesome!

test/vttyl_test.exs

@@ -79,6 +79,18 @@ defmodule VttylTest do
 
       assert length(parsed) == 20
     end
+
+    test "voice spans" do
+      parsed =
+        "voice.vtt"
+        |> get_vtt_file()
+        |> File.stream!([], 2048)
+        |> Vttyl.parse_stream()
+        |> Stream.map(& &1.voice)
+        |> Enum.into([])
+
+      assert ["Esme", "Mary", "Esme F", "Mary"] == parsed
+    end
   end
 
   describe "encode_vtt/1" do