Spaces:
Sleeping
Sleeping
Update scripts/chunk_and_embed.py
Browse files- scripts/chunk_and_embed.py +17 -14
scripts/chunk_and_embed.py
CHANGED
|
@@ -3,18 +3,21 @@ from pathlib import Path
|
|
| 3 |
import pickle
|
| 4 |
import os
|
| 5 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 6 |
|
| 7 |
-
|
| 8 |
-
|
| 9 |
-
CHUNKS_PATH = BASE_DIR / "output" / "chunks.pkl"
|
| 10 |
-
|
| 11 |
-
with open(DOCS_PATH, "rb") as f:
|
| 12 |
-
all_docs = pickle.load(f)
|
| 13 |
-
|
| 14 |
-
splitter = RecursiveCharacterTextSplitter(chunk_size=1000, chunk_overlap=100)
|
| 15 |
-
chunks = splitter.split_documents(all_docs)
|
| 16 |
-
print(f"✅ Split into {len(chunks)} chunks")
|
| 17 |
-
|
| 18 |
-
with open(CHUNKS_PATH, "wb") as f:
|
| 19 |
-
pickle.dump(chunks, f)
|
| 20 |
-
print(f"📦 Chunks saved to {CHUNKS_PATH}")
|
|
|
|
| 3 |
import pickle
|
| 4 |
import os
|
| 5 |
|
| 6 |
+
def main():
|
| 7 |
+
BASE_DIR = Path(__file__).resolve().parent.parent
|
| 8 |
+
DOCS_PATH = BASE_DIR / "output" / "all_docs.pkl"
|
| 9 |
+
CHUNKS_PATH = BASE_DIR / "output" / "chunks.pkl"
|
| 10 |
+
|
| 11 |
+
with open(DOCS_PATH, "rb") as f:
|
| 12 |
+
all_docs = pickle.load(f)
|
| 13 |
+
|
| 14 |
+
splitter = RecursiveCharacterTextSplitter(chunk_size=1000, chunk_overlap=100)
|
| 15 |
+
chunks = splitter.split_documents(all_docs)
|
| 16 |
+
print(f"✅ Split into {len(chunks)} chunks")
|
| 17 |
+
|
| 18 |
+
with open(CHUNKS_PATH, "wb") as f:
|
| 19 |
+
pickle.dump(chunks, f)
|
| 20 |
+
print(f"📦 Chunks saved to {CHUNKS_PATH}")
|
| 21 |
|
| 22 |
+
if __name__ == "__main__":
|
| 23 |
+
main()
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|