clementsan
commited on
Commit
•
32a58be
1
Parent(s):
3bf3580
Remove potential spaces in collection_name
Browse files
app.py
CHANGED
@@ -152,16 +152,19 @@ def initialize_llmchain(llm_model, temperature, max_tokens, top_k, vector_db, pr
|
|
152 |
# Initialize database
|
153 |
def initialize_database(list_file_obj, chunk_size, chunk_overlap, progress=gr.Progress()):
|
154 |
# Create list of documents (when valid)
|
155 |
-
#file_path = file_obj.name
|
156 |
list_file_path = [x.name for x in list_file_obj if x is not None]
|
|
|
|
|
157 |
collection_name = Path(list_file_path[0]).stem
|
|
|
|
|
158 |
collection_name = collection_name[:50]
|
159 |
# print('list_file_path: ', list_file_path)
|
160 |
-
|
161 |
progress(0.25, desc="Loading document...")
|
162 |
# Load document and create splits
|
163 |
doc_splits = load_doc(list_file_path, chunk_size, chunk_overlap)
|
164 |
-
# Create or load
|
165 |
progress(0.5, desc="Generating vector database...")
|
166 |
# global vector_db
|
167 |
vector_db = create_db(doc_splits, collection_name)
|
|
|
152 |
# Initialize database
|
153 |
def initialize_database(list_file_obj, chunk_size, chunk_overlap, progress=gr.Progress()):
|
154 |
# Create list of documents (when valid)
|
|
|
155 |
list_file_path = [x.name for x in list_file_obj if x is not None]
|
156 |
+
# Create collection_name for vector database
|
157 |
+
progress(0.1, desc="Creating collection name...")
|
158 |
collection_name = Path(list_file_path[0]).stem
|
159 |
+
# Fix potential issues from naming convention
|
160 |
+
collection_name = collection_name.replace(" ","-")
|
161 |
collection_name = collection_name[:50]
|
162 |
# print('list_file_path: ', list_file_path)
|
163 |
+
print('Collection name: ', collection_name)
|
164 |
progress(0.25, desc="Loading document...")
|
165 |
# Load document and create splits
|
166 |
doc_splits = load_doc(list_file_path, chunk_size, chunk_overlap)
|
167 |
+
# Create or load vector database
|
168 |
progress(0.5, desc="Generating vector database...")
|
169 |
# global vector_db
|
170 |
vector_db = create_db(doc_splits, collection_name)
|