refactor(converter): extract logic into package

1 files changed, 59 insertions, 0 deletions

diff --git a/converter/chunking.py b/converter/chunking.py
new file mode 100644
index 0000000..f649310
--- /dev/null
+++ b/converter/chunking.py
@@ -0,0 +1,59 @@
+"""Split extracted book text into TTS-sized chunks."""
+
+import re
+from typing import List
+
+from . import config
+
+
+def split_into_chunks(text: str, max_words: int = config.CHUNK_SIZE_WORDS) -> List[str]:
+    """Split text into chunks of at most ``max_words`` words.
+
+    Splits on sentence boundaries. Sentences longer than the limit are split
+    further at clause punctuation (which is kept attached for TTS prosody).
+    A single sentence with no clause punctuation longer than the limit is
+    kept intact as one oversized chunk.
+    """
+    if not text.strip():
+        return []
+
+    sentences = re.split(r"(?<=[.!?])\s+", text)
+    chunks = []
+    current_chunk = ""
+    current_words = 0
+
+    for sentence in sentences:
+        sentence_words = len(sentence.split())
+
+        if sentence_words > max_words:
+            if current_chunk:
+                chunks.append(current_chunk.strip())
+                current_chunk = ""
+                current_words = 0
+
+            # Split long sentences at clause boundaries, keeping punctuation.
+            parts = re.split(r"(?<=[,;:])\s*", sentence)
+            for part in parts:
+                part_words = len(part.split())
+                if current_words + part_words <= max_words:
+                    current_chunk += part + " "
+                    current_words += part_words
+                else:
+                    if current_chunk:
+                        chunks.append(current_chunk.strip())
+                    current_chunk = part + " "
+                    current_words = part_words
+        else:
+            if current_words + sentence_words <= max_words:
+                current_chunk += sentence + " "
+                current_words += sentence_words
+            else:
+                if current_chunk:
+                    chunks.append(current_chunk.strip())
+                current_chunk = sentence + " "
+                current_words = sentence_words
+
+    if current_chunk.strip():
+        chunks.append(current_chunk.strip())
+
+    return [chunk for chunk in chunks if chunk.strip()]