Spaces:
Runtime error
Runtime error
Upload 3 files
Browse files- Dockerfile +13 -0
- app.py +108 -0
- requirements.txt +6 -0
Dockerfile
ADDED
@@ -0,0 +1,13 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
FROM python:3.12.5
|
2 |
+
|
3 |
+
RUN useradd -m -u 1000 user
|
4 |
+
USER user
|
5 |
+
ENV PATH="/home/user/.local/bin:$PATH"
|
6 |
+
|
7 |
+
WORKDIR /app
|
8 |
+
|
9 |
+
COPY --chown=user ./requirements.txt requirements.txt
|
10 |
+
RUN pip install --no-cache-dir --upgrade -r requirements.txt
|
11 |
+
|
12 |
+
COPY --chown=user . /app
|
13 |
+
CMD ["uvicorn", "app:app", "--host", "0.0.0.0", "--port", "7860"]
|
app.py
ADDED
@@ -0,0 +1,108 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import os
|
2 |
+
import time
|
3 |
+
from fastapi import FastAPI
|
4 |
+
from fastapi.responses import HTMLResponse
|
5 |
+
from fastapi.staticfiles import StaticFiles
|
6 |
+
from llama_index.core import StorageContext, load_index_from_storage, VectorStoreIndex, SimpleDirectoryReader, ChatPromptTemplate, Settings
|
7 |
+
from llama_index.llms.huggingface import HuggingFaceInferenceAPI
|
8 |
+
from llama_index.embeddings.huggingface import HuggingFaceEmbedding
|
9 |
+
from pydantic import BaseModel
|
10 |
+
import datetime
|
11 |
+
# Define Pydantic model for incoming request body
|
12 |
+
class MessageRequest(BaseModel):
|
13 |
+
message: str
|
14 |
+
|
15 |
+
|
16 |
+
os.environ["HF_TOKEN"] = ""
|
17 |
+
app = FastAPI()
|
18 |
+
|
19 |
+
app.mount("/static", StaticFiles(directory="D:\SRUNU (1)\content\SRUNU\static"), name="static")
|
20 |
+
|
21 |
+
|
22 |
+
# Configure Llama index settings
|
23 |
+
Settings.llm = HuggingFaceInferenceAPI(
|
24 |
+
model_name="meta-llama/Meta-Llama-3-8B-Instruct",
|
25 |
+
tokenizer_name="meta-llama/Meta-Llama-3-8B-Instruct",
|
26 |
+
context_window=3000,
|
27 |
+
token="",
|
28 |
+
max_new_tokens=512,
|
29 |
+
generate_kwargs={"temperature": 0.1},
|
30 |
+
)
|
31 |
+
Settings.embed_model = HuggingFaceEmbedding(
|
32 |
+
model_name="BAAI/bge-small-en-v1.5"
|
33 |
+
)
|
34 |
+
|
35 |
+
PERSIST_DIR = "db"
|
36 |
+
PDF_DIRECTORY = 'data'
|
37 |
+
|
38 |
+
# Ensure directories exist
|
39 |
+
os.makedirs(PDF_DIRECTORY, exist_ok=True)
|
40 |
+
os.makedirs(PERSIST_DIR, exist_ok=True)
|
41 |
+
chat_history = []
|
42 |
+
current_chat_history = []
|
43 |
+
def data_ingestion_from_directory():
|
44 |
+
documents = SimpleDirectoryReader(PDF_DIRECTORY).load_data()
|
45 |
+
storage_context = StorageContext.from_defaults()
|
46 |
+
index = VectorStoreIndex.from_documents(documents)
|
47 |
+
index.storage_context.persist(persist_dir=PERSIST_DIR)
|
48 |
+
|
49 |
+
def initialize():
|
50 |
+
start_time = time.time()
|
51 |
+
data_ingestion_from_directory() # Process PDF ingestion at startup
|
52 |
+
print(f"Data ingestion time: {time.time() - start_time} seconds")
|
53 |
+
|
54 |
+
initialize() # Run initialization tasks
|
55 |
+
|
56 |
+
|
57 |
+
def handle_query(query):
|
58 |
+
chat_text_qa_msgs = [
|
59 |
+
(
|
60 |
+
"user",
|
61 |
+
"""
|
62 |
+
You are the Clara Redfernstech chatbot. Your goal is to provide accurate, professional, and helpful answers to user queries based on the company's data. Always ensure your responses are clear and concise. Give response within 10-15 words only
|
63 |
+
{context_str}
|
64 |
+
Question:
|
65 |
+
{query_str}
|
66 |
+
"""
|
67 |
+
)
|
68 |
+
]
|
69 |
+
text_qa_template = ChatPromptTemplate.from_messages(chat_text_qa_msgs)
|
70 |
+
|
71 |
+
storage_context = StorageContext.from_defaults(persist_dir=PERSIST_DIR)
|
72 |
+
index = load_index_from_storage(storage_context)
|
73 |
+
context_str = ""
|
74 |
+
for past_query, response in reversed(current_chat_history):
|
75 |
+
if past_query.strip():
|
76 |
+
context_str += f"User asked: '{past_query}'\nBot answered: '{response}'\n"
|
77 |
+
|
78 |
+
|
79 |
+
query_engine = index.as_query_engine(text_qa_template=text_qa_template, context_str=context_str)
|
80 |
+
answer = query_engine.query(query)
|
81 |
+
|
82 |
+
if hasattr(answer, 'response'):
|
83 |
+
response=answer.response
|
84 |
+
elif isinstance(answer, dict) and 'response' in answer:
|
85 |
+
response =answer['response']
|
86 |
+
else:
|
87 |
+
response ="Sorry, I couldn't find an answer."
|
88 |
+
current_chat_history.append((query, response))
|
89 |
+
return response
|
90 |
+
@app.get("/", response_class=HTMLResponse)
|
91 |
+
async def read_root():
|
92 |
+
with open("static/index.html") as f:
|
93 |
+
return f.read()
|
94 |
+
|
95 |
+
@app.post("/chat/")
|
96 |
+
async def chat(request: MessageRequest):
|
97 |
+
message = request.message # Access the message from the request body
|
98 |
+
response = handle_query(message) # Process the message
|
99 |
+
message_data = {
|
100 |
+
"sender": "User",
|
101 |
+
"message": message,
|
102 |
+
"response": response,
|
103 |
+
"timestamp": datetime.datetime.now().isoformat()
|
104 |
+
}
|
105 |
+
chat_history.append(message_data)
|
106 |
+
return {"response": response}
|
107 |
+
|
108 |
+
|
requirements.txt
ADDED
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
python-dotenv
|
2 |
+
llama-index
|
3 |
+
llama-index-embeddings-huggingface
|
4 |
+
llama-index-llms-huggingface
|
5 |
+
uvicorn
|
6 |
+
fastapi
|