Spaces:
Sleeping
Sleeping
Upload 4 files
Browse files- app.py +111 -0
- documents.json +0 -0
- faiss_index.bin +3 -0
- requirements.txt +100 -0
app.py
ADDED
@@ -0,0 +1,111 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import os
|
2 |
+
import faiss
|
3 |
+
import numpy as np
|
4 |
+
import json
|
5 |
+
import gradio as gr
|
6 |
+
from openai import OpenAI
|
7 |
+
from sentence_transformers import SentenceTransformer
|
8 |
+
|
9 |
+
# Step 1: Set up OpenAI API key
|
10 |
+
openai_api_key = os.environ.get("OPENAI_API_KEY", "")
|
11 |
+
|
12 |
+
client = OpenAI(api_key=openai_api_key)
|
13 |
+
|
14 |
+
# Step 2: Load the pre-trained FAISS index and SentenceTransformer model
|
15 |
+
index = faiss.read_index("faiss_index.bin")
|
16 |
+
model = SentenceTransformer('all-MiniLM-L6-v2')
|
17 |
+
|
18 |
+
def load_documents(docs_path):
|
19 |
+
with open(docs_path, 'r', encoding='utf-8') as file:
|
20 |
+
return json.load(file)
|
21 |
+
|
22 |
+
# Specify the path to your JSON file
|
23 |
+
docs_path = 'documents.json'
|
24 |
+
documents = load_documents(docs_path)
|
25 |
+
dimension = 1536
|
26 |
+
|
27 |
+
def get_embeddings(text):
|
28 |
+
response = client.embeddings.create(
|
29 |
+
model="text-embedding-3-small",
|
30 |
+
input = [text]
|
31 |
+
)
|
32 |
+
embedding = response.data[0].embedding
|
33 |
+
return np.array(embedding, dtype='float32')
|
34 |
+
|
35 |
+
# Step 3: Function to search FAISS index
|
36 |
+
def search_index(query, k=3):
|
37 |
+
# Convert query to an embedding
|
38 |
+
query_vector = get_embeddings(query).reshape(1, -1).astype('float32')
|
39 |
+
|
40 |
+
# Check if the index is not empty before searching
|
41 |
+
if index.ntotal == 0:
|
42 |
+
return "No documents in the index."
|
43 |
+
|
44 |
+
# Search the FAISS index for the nearest neighbors
|
45 |
+
distances, indices = index.search(query_vector, k)
|
46 |
+
|
47 |
+
# Retrieve the top matching documents
|
48 |
+
results = [documents[i] for i in indices[0] if i != -1]
|
49 |
+
|
50 |
+
if results:
|
51 |
+
return "\n\n".join(results)
|
52 |
+
else:
|
53 |
+
return "No relevant documents found."
|
54 |
+
|
55 |
+
# Step 4: Function to generate a response using OpenAI's GPT
|
56 |
+
def generate_response(context, user_input):
|
57 |
+
prompt = f"{context}\n\nUser: {user_input}\nAssistant:"
|
58 |
+
|
59 |
+
response = client.chat.completions.create(
|
60 |
+
model="gpt-4o-mini",
|
61 |
+
messages=[{"role": "system", "content": "You are a helpful assistant."},
|
62 |
+
{"role": "user", "content": prompt}],
|
63 |
+
# stream=True,
|
64 |
+
)
|
65 |
+
|
66 |
+
# for chunk in stream:
|
67 |
+
# if chunk.choices[0].delta.content is not None:
|
68 |
+
# print(chunk.choices[0].delta.content, end="")
|
69 |
+
return response.choices[0].message.content
|
70 |
+
|
71 |
+
# Step 5: Gradio chatbot function
|
72 |
+
def chatbot_interface(user_input, chat_history):
|
73 |
+
# Step 5.1: Retrieve context using FAISS
|
74 |
+
context = search_index(user_input)
|
75 |
+
|
76 |
+
# Step 5.2: Generate a response using OpenAI GPT model
|
77 |
+
response = generate_response(context, user_input)
|
78 |
+
|
79 |
+
# Step 5.3: Update chat history
|
80 |
+
chat_history.append((user_input, response))
|
81 |
+
return chat_history, chat_history
|
82 |
+
|
83 |
+
def chat_gen(message, history):
|
84 |
+
history_openai_format = []
|
85 |
+
context = search_index(message)
|
86 |
+
prompt = f"{context}\n\nUser: {message}\nAssistant:"
|
87 |
+
|
88 |
+
response = client.chat.completions.create(
|
89 |
+
model="gpt-4o-mini",
|
90 |
+
messages=[{"role": "system", "content": "You are a helpful assistant."},
|
91 |
+
{"role": "user", "content": prompt}],
|
92 |
+
stream=True,
|
93 |
+
)
|
94 |
+
partial_message = ""
|
95 |
+
for chunk in response:
|
96 |
+
if chunk.choices[0].delta.content is not None:
|
97 |
+
partial_message = partial_message + chunk.choices[0].delta.content
|
98 |
+
yield partial_message
|
99 |
+
|
100 |
+
initial_msg = "Hello! I am DII assistant. You can ask me anything about DDI program. I am happy to assist you."
|
101 |
+
chatbot = gr.Chatbot(value = [[None, initial_msg]])
|
102 |
+
demo = gr.ChatInterface(chat_gen, chatbot=chatbot).queue()
|
103 |
+
|
104 |
+
|
105 |
+
try:
|
106 |
+
demo.launch(debug=True, share=False, show_api=False)
|
107 |
+
demo.close()
|
108 |
+
except Exception as e:
|
109 |
+
demo.close()
|
110 |
+
print(e)
|
111 |
+
raise e
|
documents.json
ADDED
The diff for this file is too large to render.
See raw diff
|
|
faiss_index.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:ddae79299503fe170726dd027d6d4cf5a7057ebdcd36a186d0a643f981a1c4b0
|
3 |
+
size 755757
|
requirements.txt
ADDED
@@ -0,0 +1,100 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
aiofiles==23.2.1
|
2 |
+
annotated-types==0.7.0
|
3 |
+
anyio==4.6.2.post1
|
4 |
+
appnope==0.1.4
|
5 |
+
asttokens==2.4.1
|
6 |
+
certifi==2024.8.30
|
7 |
+
charset-normalizer==3.4.0
|
8 |
+
click==8.1.7
|
9 |
+
comm==0.2.2
|
10 |
+
debugpy==1.8.8
|
11 |
+
decorator==5.1.1
|
12 |
+
distro==1.9.0
|
13 |
+
et_xmlfile==2.0.0
|
14 |
+
executing==2.1.0
|
15 |
+
faiss-cpu==1.9.0
|
16 |
+
fastapi==0.115.4
|
17 |
+
ffmpy==0.4.0
|
18 |
+
filelock==3.16.1
|
19 |
+
fsspec==2024.10.0
|
20 |
+
gradio==5.5.0
|
21 |
+
gradio_client==1.4.2
|
22 |
+
h11==0.14.0
|
23 |
+
httpcore==1.0.6
|
24 |
+
httpx==0.27.2
|
25 |
+
huggingface-hub==0.26.2
|
26 |
+
idna==3.10
|
27 |
+
ipykernel==6.29.5
|
28 |
+
ipython==8.29.0
|
29 |
+
jedi==0.19.2
|
30 |
+
Jinja2==3.1.4
|
31 |
+
jiter==0.7.0
|
32 |
+
joblib==1.4.2
|
33 |
+
jupyter_client==8.6.3
|
34 |
+
jupyter_core==5.7.2
|
35 |
+
lxml==5.3.0
|
36 |
+
markdown-it-py==3.0.0
|
37 |
+
MarkupSafe==2.1.5
|
38 |
+
matplotlib-inline==0.1.7
|
39 |
+
mdurl==0.1.2
|
40 |
+
mpmath==1.3.0
|
41 |
+
nest-asyncio==1.6.0
|
42 |
+
networkx==3.4.2
|
43 |
+
numpy==2.1.3
|
44 |
+
openai==1.54.3
|
45 |
+
openpyxl==3.1.5
|
46 |
+
orjson==3.10.11
|
47 |
+
packaging==24.2
|
48 |
+
pandas==2.2.3
|
49 |
+
parso==0.8.4
|
50 |
+
pexpect==4.9.0
|
51 |
+
pillow==11.0.0
|
52 |
+
platformdirs==4.3.6
|
53 |
+
prompt_toolkit==3.0.48
|
54 |
+
psutil==6.1.0
|
55 |
+
ptyprocess==0.7.0
|
56 |
+
pure_eval==0.2.3
|
57 |
+
pydantic==2.9.2
|
58 |
+
pydantic_core==2.23.4
|
59 |
+
pydub==0.25.1
|
60 |
+
Pygments==2.18.0
|
61 |
+
PyPDF2==3.0.1
|
62 |
+
pytesseract==0.3.13
|
63 |
+
python-dateutil==2.9.0.post0
|
64 |
+
python-docx==1.1.2
|
65 |
+
python-multipart==0.0.12
|
66 |
+
pytz==2024.2
|
67 |
+
PyYAML==6.0.2
|
68 |
+
pyzmq==26.2.0
|
69 |
+
regex==2024.11.6
|
70 |
+
requests==2.32.3
|
71 |
+
rich==13.9.4
|
72 |
+
ruff==0.7.3
|
73 |
+
safehttpx==0.1.1
|
74 |
+
safetensors==0.4.5
|
75 |
+
scikit-learn==1.5.2
|
76 |
+
scipy==1.14.1
|
77 |
+
semantic-version==2.10.0
|
78 |
+
sentence-transformers==3.3.0
|
79 |
+
setuptools==75.4.0
|
80 |
+
shellingham==1.5.4
|
81 |
+
six==1.16.0
|
82 |
+
sniffio==1.3.1
|
83 |
+
stack-data==0.6.3
|
84 |
+
starlette==0.41.2
|
85 |
+
sympy==1.13.1
|
86 |
+
threadpoolctl==3.5.0
|
87 |
+
tokenizers==0.20.3
|
88 |
+
tomlkit==0.12.0
|
89 |
+
torch==2.5.1
|
90 |
+
tornado==6.4.1
|
91 |
+
tqdm==4.67.0
|
92 |
+
traitlets==5.14.3
|
93 |
+
transformers==4.46.2
|
94 |
+
typer==0.13.0
|
95 |
+
typing_extensions==4.12.2
|
96 |
+
tzdata==2024.2
|
97 |
+
urllib3==2.2.3
|
98 |
+
uvicorn==0.32.0
|
99 |
+
wcwidth==0.2.13
|
100 |
+
websockets==12.0
|