README, word-phoneme dict, phoneme progress script

- Updated README a little
- Added a word-to-phoneme dictionary (CMU open source)
- Added a Python script to track progress on phonemes

README.md

@@ -9,3 +9,30 @@ Running the server
 1. Install Rust >= 1.1.0
 2. `cargo run`
 
+Phoneme
+-------
+- [CMU Dictionary](http://svn.code.sf.net/p/cmusphinx/code/trunk/cmudict/), 
+  which uses [Arpabet](https://en.wikipedia.org/wiki/Arpabet).
+  - 54 units in the Arpabet:
+    - Vowels:
+      - 10 monophthongs
+      - 5 diphthongs
+      - 8 R-colored vowels
+    - Consonants:
+      - 6 Stops
+      - 2 Affricates
+      - 9 Fricatives
+      - 6 Nasals
+      - 5 Liquids
+      - 3 Semivowels
+
+Diphones
+--------
+Generate a list of phoneme 2-tuples.
+
+Unit Selection
+--------------
+- Combine multiple sources into final output (words, phonemes, diphones)
+- Have alternatives (words, phonemes) that pair better with others.
+- Optimize joining on a score (start/ending pitch, tempo, etc.)
+

script/arpabet_progress.py

@@ -0,0 +1,62 @@
+#!/usr/bin/env python2
+# Show how many phonemes remain
+
+from os import path
+import os
+import re
+
+DICTIONARY_FILE = '/home/bt/dev/trumpet/dictionary/cmudict-0.7b'
+PHONE_DIR = '/home/bt/Dropbox/jungle.horse/sounds/trump/_phonemes'
+
+def current_phonemes():
+    """
+    Gets the current set of phoneme wav files.
+    Returns a frozenset of strings.
+    """
+    phonemes = []
+    for p in os.listdir(PHONE_DIR):
+        if not p.endswith('.wav'):
+            continue
+        phonemes.append(p.replace('.wav', ''))
+
+    return frozenset(phonemes)
+
+def all_phonemes():
+    """
+    Read the CMU Word-Phoneme dictionary and extract the phoneme set.
+    Returns a frozenset of strings.
+    """
+    lines = []
+    with open(DICTIONARY_FILE) as f:
+        lines = f.readlines()
+
+    lines = [l for l in lines if re.match("^\w+", l)]
+    phonemes = set()
+
+    for l in lines:
+        word_phones = l.split()[1:]
+        for w in word_phones:
+            phonemes.add(w)
+
+    return frozenset(phonemes)
+
+current = current_phonemes()
+phonemes = all_phonemes()
+
+print "\nAll Phonemes: {}\n".format(len(phonemes))
+li = list(phonemes)
+li.sort()
+print li
+
+print "\nCurrent Phonemes: {}\n".format(len(current))
+li = list(current)
+li.sort()
+print li
+
+missing = phonemes - current
+
+print "\nMissing Phonemes: {}\n".format(len(missing))
+li = list(missing)
+li.sort()
+print li
+