Update util.py
Browse files
util.py
CHANGED
|
@@ -38,7 +38,16 @@ def read_docx(file_path):
|
|
| 38 |
|
| 39 |
for para in doc.paragraphs:
|
| 40 |
text.append(para.text)
|
| 41 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 42 |
return "\n".join(text)
|
| 43 |
|
| 44 |
def read_pptx(file_path):
|
|
|
|
| 38 |
|
| 39 |
for para in doc.paragraphs:
|
| 40 |
text.append(para.text)
|
| 41 |
+
|
| 42 |
+
for table in doc.tables:
|
| 43 |
+
for row in table.rows:
|
| 44 |
+
row_text = []
|
| 45 |
+
|
| 46 |
+
for cell in row.cells:
|
| 47 |
+
row_text.append(cell.text)
|
| 48 |
+
|
| 49 |
+
text.append(" | ".join(row_text))
|
| 50 |
+
|
| 51 |
return "\n".join(text)
|
| 52 |
|
| 53 |
def read_pptx(file_path):
|