Penality commited on
Commit
9aee54a
·
verified ·
1 Parent(s): 2a7ef32

Update app.py

Browse files

added deug statements to observe code execution

Files changed (1) hide show
  1. app.py +7 -3
app.py CHANGED
@@ -26,20 +26,21 @@ def store_document(text):
26
  print("storing document")
27
 
28
  embedding = embedding_model.encode([text])
 
29
  index.add(np.array(embedding, dtype=np.float32))
30
  documents.append(text)
31
 
32
- print(f"your document has been stored: \n{documents}")
33
 
34
  return "Document stored!"
35
 
36
  def retrieve_document(query):
37
- print(f"retrieving doc based on {query}")
38
 
39
  query_embedding = embedding_model.encode([query])
40
  _, closest_idx = index.search(np.array(query_embedding, dtype=np.float32), 1)
41
 
42
- print(f"retrieved: {documents[closest_idx[0][0]]}")
43
 
44
  return documents[closest_idx[0][0]]
45
 
@@ -87,15 +88,18 @@ def chatbot(pdf_file, user_question):
87
  return f"Error retrieving document relevant to the query: {user_question} \n{e}"
88
 
89
  if doc:
 
90
  # Split into smaller chunks
91
  chunks = split_text(doc)
92
 
93
  # Use only the first chunk (to optimize token usage)
94
  prompt = f"Based on this document, answer the question:\n\nDocument:\n{chunks[0]}\n\nQuestion: {user_question}"
 
95
  else:
96
  prompt=user_question
97
 
98
  try:
 
99
  response = together.Completion.create(
100
  model="mistralai/Mistral-7B-Instruct-v0.1",
101
  prompt=prompt,
 
26
  print("storing document")
27
 
28
  embedding = embedding_model.encode([text])
29
+ print(f"embedding: \n{embedding}")
30
  index.add(np.array(embedding, dtype=np.float32))
31
  documents.append(text)
32
 
33
+ print(f"your document has been stored")
34
 
35
  return "Document stored!"
36
 
37
  def retrieve_document(query):
38
+ print(f"retrieving doc based on: \n{query}")
39
 
40
  query_embedding = embedding_model.encode([query])
41
  _, closest_idx = index.search(np.array(query_embedding, dtype=np.float32), 1)
42
 
43
+ print(f"retrieved: \n{documents[closest_idx[0][0]]}")
44
 
45
  return documents[closest_idx[0][0]]
46
 
 
88
  return f"Error retrieving document relevant to the query: {user_question} \n{e}"
89
 
90
  if doc:
91
+ print("found doc")
92
  # Split into smaller chunks
93
  chunks = split_text(doc)
94
 
95
  # Use only the first chunk (to optimize token usage)
96
  prompt = f"Based on this document, answer the question:\n\nDocument:\n{chunks[0]}\n\nQuestion: {user_question}"
97
+ print(f"prompt: \n{prompt}")
98
  else:
99
  prompt=user_question
100
 
101
  try:
102
+ print("asking")
103
  response = together.Completion.create(
104
  model="mistralai/Mistral-7B-Instruct-v0.1",
105
  prompt=prompt,