Update document_processor.py
Browse files- document_processor.py +3 -3
document_processor.py
CHANGED
|
@@ -7,7 +7,7 @@ from utils import clean_text, setup_logger
|
|
| 7 |
|
| 8 |
logger = setup_logger('document_processor')
|
| 9 |
|
| 10 |
-
def split_into_chunks(text, chunk_size=
|
| 11 |
"""
|
| 12 |
Split text into overlapping chunks
|
| 13 |
|
|
@@ -48,7 +48,7 @@ def split_into_chunks(text, chunk_size=400, overlap=75):
|
|
| 48 |
|
| 49 |
return chunks
|
| 50 |
|
| 51 |
-
def load_documents(folder_path, chunk_size=
|
| 52 |
"""
|
| 53 |
Load all .txt documents from folder and split them into chunks
|
| 54 |
|
|
@@ -102,7 +102,7 @@ def load_documents(folder_path, chunk_size=400, overlap=75):
|
|
| 102 |
|
| 103 |
return df
|
| 104 |
|
| 105 |
-
def load_single_document(file_path, chunk_size=
|
| 106 |
"""
|
| 107 |
Load a single document and split it into chunks
|
| 108 |
|
|
|
|
| 7 |
|
| 8 |
logger = setup_logger('document_processor')
|
| 9 |
|
| 10 |
+
def split_into_chunks(text, chunk_size=300, overlap=100):
|
| 11 |
"""
|
| 12 |
Split text into overlapping chunks
|
| 13 |
|
|
|
|
| 48 |
|
| 49 |
return chunks
|
| 50 |
|
| 51 |
+
def load_documents(folder_path, chunk_size=300, overlap=100):
|
| 52 |
"""
|
| 53 |
Load all .txt documents from folder and split them into chunks
|
| 54 |
|
|
|
|
| 102 |
|
| 103 |
return df
|
| 104 |
|
| 105 |
+
def load_single_document(file_path, chunk_size=300, overlap=100):
|
| 106 |
"""
|
| 107 |
Load a single document and split it into chunks
|
| 108 |
|