Spaces:
Running
Running
Update app.py
Browse files
app.py
CHANGED
@@ -38,7 +38,6 @@ class PDFDocument(Document):
|
|
38 |
|
39 |
def initialize_database(document, chunk_size, chunk_overlap, progress=gr.Progress()):
|
40 |
logger.info("Initializing database...")
|
41 |
-
embedding_function = Chroma.from_pretrained("chroma-rt")
|
42 |
documents = []
|
43 |
for file in document:
|
44 |
loader = UnstructuredPDFLoader(file.name)
|
@@ -48,11 +47,12 @@ def initialize_database(document, chunk_size, chunk_overlap, progress=gr.Progres
|
|
48 |
pages = splitter.split_document(doc)
|
49 |
for page in pages:
|
50 |
documents.append(PDFDocument(page_content=page.page_content, metadata={"filename": file.name}))
|
51 |
-
|
|
|
52 |
progress.update(0.5)
|
53 |
logger.info("Database initialized successfully.")
|
54 |
return vectorstore, "Initialized"
|
55 |
-
|
56 |
def initialize_LLM(llm_option, llm_temperature, max_tokens, top_k, vector_db, progress=gr.Progress(), language="italian"):
|
57 |
logger.info("Initializing LLM chain...")
|
58 |
llm_name = list_llm[llm_option]
|
|
|
38 |
|
39 |
def initialize_database(document, chunk_size, chunk_overlap, progress=gr.Progress()):
|
40 |
logger.info("Initializing database...")
|
|
|
41 |
documents = []
|
42 |
for file in document:
|
43 |
loader = UnstructuredPDFLoader(file.name)
|
|
|
47 |
pages = splitter.split_document(doc)
|
48 |
for page in pages:
|
49 |
documents.append(PDFDocument(page_content=page.page_content, metadata={"filename": file.name}))
|
50 |
+
|
51 |
+
vectorstore = Chroma.from_documents(documents, embedding_function)
|
52 |
progress.update(0.5)
|
53 |
logger.info("Database initialized successfully.")
|
54 |
return vectorstore, "Initialized"
|
55 |
+
|
56 |
def initialize_LLM(llm_option, llm_temperature, max_tokens, top_k, vector_db, progress=gr.Progress(), language="italian"):
|
57 |
logger.info("Initializing LLM chain...")
|
58 |
llm_name = list_llm[llm_option]
|