Update ppt_chunker.py

ppt_chunker.py

@@ -2,7 +2,7 @@ from io import StringIO
 from typing import List
 
 from unstructured.partition.pptx import partition_pptx
-from unstructured.cleaners.core import clean_trailing_punctuation, clean_bullets
+from unstructured.cleaners.core import clean_trailing_punctuation, clean_bullets, clean
 
 from ordered_multimap import OrderedMultiIndexMapWeakRef
 
@@ -162,4 +162,23 @@ def ppt_chunk(file_like, nlp):
     return weakDict, tables
 
     raise NotImplementedError(
-        "file type not supported yet(pptx)")
+        "file type not supported yet(pptx)")
+
+def ppt_chunker(file_like, llm):
+    import time
+
+    s = time.time()
+    elements = partition_pptx(file=file_like)
+
+    chunks = []
+    current_chunk = []
+
+    for elem in elements:
+        if elem.category == 'PageBreak':
+            chunks.append(current_chunk)
+            current_chunk = []
+        current_chunk.append(clean(elem.text, extra_whitespace=True, dashes=True, bullets=True, lowercase=True, trailing_punctuation=True))
+
+    for chunk in chunk:
+        print(f' TEXT : {chunk}')
+    return chunks