Add tests and update README

espeak_phonemizer/VERSION

@@ -1 +1 @@
-1.0
+1.0.0

espeak_phonemizer/__init__.py

@@ -1,10 +1,23 @@
+"""Uses ctypes and libespeak-ng to get IPA phonemes from text"""
 import ctypes
 import re
 import typing
+from pathlib import Path
+
+_DIR = Path(__file__).parent
+__version__ = (_DIR / "VERSION").read_text().strip()
+
+# -----------------------------------------------------------------------------
 
 
 class Phonemizer:
-    """Use ctypes and libespeak-ng to get IPA phonemes from text"""
+    """
+    Use ctypes and libespeak-ng to get IPA phonemes from text.
+    Not thread safe.
+
+    Requires libc.so.6
+    Tries to use libespeak-ng.so or libespeak-ng.so.1
+    """
 
     SEEK_SET = 0
 
@@ -44,7 +57,23 @@ class Phonemizer:
         keep_language_flags: bool = False,
         no_stress: bool = False,
     ) -> str:
-        """Return IPA string for text"""
+        """
+        Return IPA string for text.
+        Not thread safe.
+
+        Args:
+            text: Text to phonemize
+            voice: optional voice (uses self.default_voice if None)
+            keep_clause_breakers: True if punctuation symbols should be kept
+            phoneme_separator: Separator character between phonemes
+            word_separator: Separator string between words (default: space)
+            punctuation_separator: Separator string between before punctuation (keep_clause_breakers=True)
+            keep_language_flags: True if language switching flags should be kept
+            no_stress: True if stress characters should be removed
+
+        Returns:
+            ipa - string of IPA phonemes
+        """
         self._maybe_init()
 
         voice = voice or self.default_voice
@@ -97,13 +126,20 @@ class Phonemizer:
                     Phonemizer.LANG_SWITCH_FLAG.sub("", line) for line in phoneme_lines
                 ]
 
+            if word_separator != " ":
+                # Split/re-join words
+                for line_idx in range(len(phoneme_lines)):
+                    phoneme_lines[line_idx] = word_separator.join(
+                        phoneme_lines[line_idx].split()
+                    )
+
             # Re-insert clause breakers
             if missing_breakers:
                 # pylint: disable=consider-using-enumerate
                 for line_idx in range(len(phoneme_lines)):
                     if line_idx < len(missing_breakers):
                         phoneme_lines[line_idx] += (
-                            word_separator + missing_breakers[line_idx]
+                            punctuation_separator + missing_breakers[line_idx]
                         )
 
             phonemes_str = word_separator.join(line.strip() for line in phoneme_lines)

espeak_phonemizer/__main__.py

@@ -1,14 +1,20 @@
+"""Command-line interface to espeak_phonemizer"""
 import argparse
 import csv
+import logging
 import os
 import sys
 
 from . import Phonemizer
 
+_LOGGER = logging.getLogger("espeak_phonemizer")
+
+# -----------------------------------------------------------------------------
+
 
 def main():
     parser = argparse.ArgumentParser(prog="espeak_phonemizer")
-    parser.add_argument("-v", "--voice", required=True, help="eSpeak voice to use")
+    parser.add_argument("-v", "--voice", help="eSpeak voice to use")
     parser.add_argument(
         "-p", "--phoneme-separator", help="Separator character between phonemes"
     )
@@ -48,8 +54,30 @@ def main():
     parser.add_argument(
         "--csv-delimiter", default="|", help="Delimiter in CSV input and output"
     )
+    parser.add_argument("--version", action="store_true", help="Print version and exit")
+    parser.add_argument(
+        "--debug", action="store_true", help="Print DEBUG messages to the console"
+    )
     args = parser.parse_args()
 
+    if args.debug:
+        logging.basicConfig(level=logging.DEBUG)
+    else:
+        logging.basicConfig(level=logging.INFO)
+
+    # -------------------------------------------------------------------------
+
+    if args.version:
+        # Print version and exit
+        from . import __version__
+
+        print(__version__)
+        sys.exit(0)
+
+    # -------------------------------------------------------------------------
+
+    assert args.voice, "Missing -v/--voice"
+
     if args.word_separator:
         assert (
             args.phoneme_separator.strip()
@@ -82,7 +110,7 @@ def main():
             phoneme_separator=args.phoneme_separator,
             keep_language_flags=args.keep_language_flags,
             no_stress=args.no_stress,
-            punctuation_separator=args.phoneme_separator,
+            punctuation_separator=args.phoneme_separator or "",
         )
 
         if args.word_separator: