OnlyTheTruth03 commited on
Commit
2a71f29
·
1 Parent(s): 0b4e744

config fix

Browse files
Files changed (1) hide show
  1. src/config.py +14 -3
src/config.py CHANGED
@@ -1,11 +1,22 @@
1
  # src/config.py
2
  from pathlib import Path
3
 
4
- BASE_DIR = Path(__file__).parent
 
 
 
 
 
 
 
 
5
 
6
  INDEX_DIR = BASE_DIR / "index"
7
  FAISS_INDEX_PATH = INDEX_DIR / "faiss.index"
8
  DOCS_PATH = INDEX_DIR / "documents.pkl"
9
 
10
- DATASET_NAME = "OnlyTheTruth03/ott" # 👈 your HF dataset
11
- TOP_K = 4
 
 
 
 
1
  # src/config.py
2
  from pathlib import Path
3
 
4
+ # -------------------------------
5
+ # Hugging Face Dataset
6
+ # -------------------------------
7
+ DATASET_NAME = "OnlyTheTruth03/OTT_PDFs" # change ONLY if dataset name changes
8
+
9
+ # -------------------------------
10
+ # FAISS storage paths
11
+ # -------------------------------
12
+ BASE_DIR = Path(__file__).resolve().parent
13
 
14
  INDEX_DIR = BASE_DIR / "index"
15
  FAISS_INDEX_PATH = INDEX_DIR / "faiss.index"
16
  DOCS_PATH = INDEX_DIR / "documents.pkl"
17
 
18
+ # -------------------------------
19
+ # Chunking config
20
+ # -------------------------------
21
+ CHUNK_SIZE = 500
22
+ CHUNK_OVERLAP = 50