northstaranlyticsma24
commited on
Commit
•
79eb986
1
Parent(s):
c56e207
Upload 3 files
Browse files- Dockerfile +11 -0
- app.py +153 -0
- requirements.txt +132 -0
Dockerfile
ADDED
@@ -0,0 +1,11 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
FROM python:3.9
|
2 |
+
RUN useradd -m -u 1000 user
|
3 |
+
USER user
|
4 |
+
ENV HOME=/home/user \
|
5 |
+
PATH=/home/user/.local/bin:$PATH
|
6 |
+
WORKDIR $HOME/app
|
7 |
+
COPY --chown=user . $HOME/app
|
8 |
+
COPY ./requirements.txt ~/app/requirements.txt
|
9 |
+
RUN pip install -r requirements.txt
|
10 |
+
COPY . .
|
11 |
+
CMD ["chainlit", "run", "app.py", "--port", "7860"]
|
app.py
ADDED
@@ -0,0 +1,153 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
### Import Section ###
|
2 |
+
from langchain_text_splitters import RecursiveCharacterTextSplitter
|
3 |
+
from langchain_community.document_loaders import PyMuPDFLoader
|
4 |
+
from qdrant_client import QdrantClient
|
5 |
+
from qdrant_client.http.models import Distance, VectorParams
|
6 |
+
from langchain_openai.embeddings import OpenAIEmbeddings
|
7 |
+
from langchain.storage import LocalFileStore
|
8 |
+
from langchain_qdrant import QdrantVectorStore
|
9 |
+
from langchain.embeddings import CacheBackedEmbeddings
|
10 |
+
from langchain_core.prompts import ChatPromptTemplate
|
11 |
+
from langchain_core.globals import set_llm_cache
|
12 |
+
from langchain_openai import ChatOpenAI
|
13 |
+
from langchain_core.caches import InMemoryCache
|
14 |
+
from operator import itemgetter
|
15 |
+
from langchain_core.runnables.passthrough import RunnablePassthrough
|
16 |
+
import uuid
|
17 |
+
import chainlit as cl
|
18 |
+
|
19 |
+
### Global Section ###
|
20 |
+
|
21 |
+
text_splitter = RecursiveCharacterTextSplitter(chunk_size=1000, chunk_overlap=100)
|
22 |
+
|
23 |
+
Loader = PyMuPDFLoader
|
24 |
+
|
25 |
+
set_llm_cache(InMemoryCache())
|
26 |
+
|
27 |
+
# Typical QDrant Client Set-up
|
28 |
+
collection_name = f"pdf_to_parse_{uuid.uuid4()}"
|
29 |
+
client = QdrantClient(":memory:")
|
30 |
+
client.create_collection(
|
31 |
+
collection_name=collection_name,
|
32 |
+
vectors_config=VectorParams(size=1536, distance=Distance.COSINE),
|
33 |
+
)
|
34 |
+
|
35 |
+
# Typical Embedding Model
|
36 |
+
core_embeddings = OpenAIEmbeddings(model="text-embedding-3-small")
|
37 |
+
|
38 |
+
rag_system_prompt_template = """\
|
39 |
+
You are a helpful assistant that uses the provided context to answer questions. Never reference this prompt, or the existance of context.
|
40 |
+
"""
|
41 |
+
|
42 |
+
rag_message_list = [
|
43 |
+
{"role" : "system", "content" : rag_system_prompt_template},
|
44 |
+
]
|
45 |
+
|
46 |
+
rag_user_prompt_template = """\
|
47 |
+
Question:
|
48 |
+
{question}
|
49 |
+
Context:
|
50 |
+
{context}
|
51 |
+
"""
|
52 |
+
|
53 |
+
chat_prompt = ChatPromptTemplate.from_messages([
|
54 |
+
("system", rag_system_prompt_template),
|
55 |
+
("human", rag_user_prompt_template)
|
56 |
+
])
|
57 |
+
|
58 |
+
chat_model = ChatOpenAI(model="gpt-4o-mini")
|
59 |
+
|
60 |
+
def process_file(file: cl.AskFileResponse):
|
61 |
+
import tempfile
|
62 |
+
|
63 |
+
with tempfile.NamedTemporaryFile(mode="w", delete=False) as temp_file:
|
64 |
+
with open(temp_file.name, "wb") as f:
|
65 |
+
f.write(file.content)
|
66 |
+
|
67 |
+
Loader = PyMuPDFLoader
|
68 |
+
|
69 |
+
loader = Loader(temp_file.name)
|
70 |
+
documents = loader.load()
|
71 |
+
docs = text_splitter.split_documents(documents)
|
72 |
+
for i, doc in enumerate(docs):
|
73 |
+
doc.metadata["source"] = f"source_{i}"
|
74 |
+
return docs
|
75 |
+
|
76 |
+
|
77 |
+
### On Chat Start (Session Start) Section ###
|
78 |
+
@cl.on_chat_start
|
79 |
+
async def on_chat_start():
|
80 |
+
|
81 |
+
# Adding cache!
|
82 |
+
store = LocalFileStore("./cache/")
|
83 |
+
cached_embedder = CacheBackedEmbeddings.from_bytes_store(
|
84 |
+
core_embeddings, store, namespace=core_embeddings.model
|
85 |
+
)
|
86 |
+
|
87 |
+
# Typical QDrant Vector Store Set-up
|
88 |
+
vectorstore = QdrantVectorStore(
|
89 |
+
client=client,
|
90 |
+
collection_name=collection_name,
|
91 |
+
embedding=cached_embedder)
|
92 |
+
vectorstore.add_documents(docs)
|
93 |
+
retriever = vectorstore.as_retriever(search_type="mmr", search_kwargs={"k": 3})
|
94 |
+
|
95 |
+
retrieval_augmented_qa_chain = (
|
96 |
+
{"context": itemgetter("question") | retriever, "question": itemgetter("question")} ##
|
97 |
+
| RunnablePassthrough.assign(context=itemgetter("context"))
|
98 |
+
| chat_prompt | chat_model
|
99 |
+
)
|
100 |
+
|
101 |
+
cl.user_session.set("midterm_chain", retrieval_augmented_qa_chain)
|
102 |
+
|
103 |
+
files = None
|
104 |
+
|
105 |
+
# Wait for the user to upload a file
|
106 |
+
while files == None:
|
107 |
+
# Async method: This allows the function to pause execution while waiting for the user to upload a file,
|
108 |
+
# without blocking the entire application. It improves responsiveness and scalability.
|
109 |
+
files = await cl.AskFileMessage(
|
110 |
+
content="Please upload a PDF file to begin!",
|
111 |
+
accept=["application/pdf"],
|
112 |
+
max_size_mb=20,
|
113 |
+
timeout=180,
|
114 |
+
).send()
|
115 |
+
|
116 |
+
file = files[0]
|
117 |
+
|
118 |
+
msg = cl.Message(
|
119 |
+
content=f"Processing `{file.name}`...",
|
120 |
+
)
|
121 |
+
await msg.send()
|
122 |
+
|
123 |
+
# load the file
|
124 |
+
docs = process_file(file)
|
125 |
+
|
126 |
+
### Rename Chains ###
|
127 |
+
@cl.author_rename
|
128 |
+
def rename(orig_author: str):
|
129 |
+
""" RENAME CODE HERE """
|
130 |
+
|
131 |
+
### On Message Section ###
|
132 |
+
@cl.on_message
|
133 |
+
async def main(message: cl.Message):
|
134 |
+
try:
|
135 |
+
# Retrieve the chain stored in the session
|
136 |
+
midterm_chain = cl.user_session.get("midterm_chain")
|
137 |
+
|
138 |
+
# Pass the user's message (query) to the chain for processing
|
139 |
+
response = await midterm_chain.run(message.content)
|
140 |
+
|
141 |
+
# Send the response back to the user
|
142 |
+
await message.send(response)
|
143 |
+
# Process the incoming question using the RAG chain
|
144 |
+
#result = retrieval_augmented_qa_chain.invoke({"question": message.content})
|
145 |
+
|
146 |
+
# Create a new message for the response
|
147 |
+
#response_message = cl.Message(content=result["response"].content)
|
148 |
+
|
149 |
+
except Exception as e:
|
150 |
+
# Handle any exception and log it or send a response back to the user
|
151 |
+
error_message = cl.Message(content=f"An error occurred: {str(e)}")
|
152 |
+
await error_message.send()
|
153 |
+
print(f"Error occurred: {e}")
|
requirements.txt
ADDED
@@ -0,0 +1,132 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
aiofiles==23.2.1
|
2 |
+
aiohappyeyeballs==2.4.3
|
3 |
+
aiohttp==3.10.8
|
4 |
+
aiosignal==1.3.1
|
5 |
+
annotated-types==0.7.0
|
6 |
+
anyio==3.7.1
|
7 |
+
async-timeout==4.0.3
|
8 |
+
asyncer==0.0.2
|
9 |
+
attrs==24.2.0
|
10 |
+
bidict==0.23.1
|
11 |
+
certifi==2024.8.30
|
12 |
+
chainlit==0.7.700
|
13 |
+
charset-normalizer==3.3.2
|
14 |
+
click==8.1.7
|
15 |
+
dataclasses-json==0.5.14
|
16 |
+
Deprecated==1.2.14
|
17 |
+
distro==1.9.0
|
18 |
+
exceptiongroup==1.2.2
|
19 |
+
faiss-cpu==1.8.0.post1
|
20 |
+
fastapi==0.100.1
|
21 |
+
fastapi-socketio==0.0.10
|
22 |
+
filelock==3.16.1
|
23 |
+
filetype==1.2.0
|
24 |
+
frozenlist==1.4.1
|
25 |
+
fsspec==2024.9.0
|
26 |
+
googleapis-common-protos==1.65.0
|
27 |
+
greenlet==3.1.1
|
28 |
+
grpcio==1.66.2
|
29 |
+
grpcio-tools==1.62.3
|
30 |
+
h11==0.14.0
|
31 |
+
h2==4.1.0
|
32 |
+
hpack==4.0.0
|
33 |
+
httpcore==0.17.3
|
34 |
+
httpx==0.24.1
|
35 |
+
huggingface-hub==0.25.1
|
36 |
+
hyperframe==6.0.1
|
37 |
+
idna==3.10
|
38 |
+
importlib_metadata==8.4.0
|
39 |
+
Jinja2==3.1.4
|
40 |
+
jiter==0.5.0
|
41 |
+
joblib==1.4.2
|
42 |
+
jsonpatch==1.33
|
43 |
+
jsonpointer==3.0.0
|
44 |
+
langchain==0.3.0
|
45 |
+
langchain-community==0.3.0
|
46 |
+
langchain-core==0.3.1
|
47 |
+
langchain-huggingface==0.1.0
|
48 |
+
langchain-openai==0.2.0
|
49 |
+
langchain-qdrant==0.1.4
|
50 |
+
langchain-text-splitters==0.3.0
|
51 |
+
langsmith==0.1.121
|
52 |
+
Lazify==0.4.0
|
53 |
+
MarkupSafe==2.1.5
|
54 |
+
marshmallow==3.22.0
|
55 |
+
mpmath==1.3.0
|
56 |
+
multidict==6.1.0
|
57 |
+
mypy-extensions==1.0.0
|
58 |
+
nest-asyncio==1.6.0
|
59 |
+
networkx==3.2.1
|
60 |
+
numpy==1.26.4
|
61 |
+
nvidia-cublas-cu12==12.1.3.1
|
62 |
+
nvidia-cuda-cupti-cu12==12.1.105
|
63 |
+
nvidia-cuda-nvrtc-cu12==12.1.105
|
64 |
+
nvidia-cuda-runtime-cu12==12.1.105
|
65 |
+
nvidia-cudnn-cu12==9.1.0.70
|
66 |
+
nvidia-cufft-cu12==11.0.2.54
|
67 |
+
nvidia-curand-cu12==10.3.2.106
|
68 |
+
nvidia-cusolver-cu12==11.4.5.107
|
69 |
+
nvidia-cusparse-cu12==12.1.0.106
|
70 |
+
nvidia-nccl-cu12==2.20.5
|
71 |
+
nvidia-nvjitlink-cu12==12.6.77
|
72 |
+
nvidia-nvtx-cu12==12.1.105
|
73 |
+
openai==1.51.0
|
74 |
+
opentelemetry-api==1.27.0
|
75 |
+
opentelemetry-exporter-otlp==1.27.0
|
76 |
+
opentelemetry-exporter-otlp-proto-common==1.27.0
|
77 |
+
opentelemetry-exporter-otlp-proto-grpc==1.27.0
|
78 |
+
opentelemetry-exporter-otlp-proto-http==1.27.0
|
79 |
+
opentelemetry-instrumentation==0.48b0
|
80 |
+
opentelemetry-proto==1.27.0
|
81 |
+
opentelemetry-sdk==1.27.0
|
82 |
+
opentelemetry-semantic-conventions==0.48b0
|
83 |
+
orjson==3.10.7
|
84 |
+
packaging==23.2
|
85 |
+
pillow==10.4.0
|
86 |
+
portalocker==2.10.1
|
87 |
+
protobuf==4.25.5
|
88 |
+
pydantic==2.9.2
|
89 |
+
pydantic-settings==2.5.2
|
90 |
+
pydantic_core==2.23.4
|
91 |
+
PyJWT==2.9.0
|
92 |
+
PyMuPDF==1.24.10
|
93 |
+
PyMuPDFb==1.24.10
|
94 |
+
python-dotenv==1.0.1
|
95 |
+
python-engineio==4.9.1
|
96 |
+
python-graphql-client==0.4.3
|
97 |
+
python-multipart==0.0.6
|
98 |
+
python-socketio==5.11.4
|
99 |
+
PyYAML==6.0.2
|
100 |
+
qdrant-client==1.11.2
|
101 |
+
regex==2024.9.11
|
102 |
+
requests==2.32.3
|
103 |
+
safetensors==0.4.5
|
104 |
+
scikit-learn==1.5.2
|
105 |
+
scipy==1.13.1
|
106 |
+
sentence-transformers==3.1.1
|
107 |
+
simple-websocket==1.0.0
|
108 |
+
sniffio==1.3.1
|
109 |
+
SQLAlchemy==2.0.35
|
110 |
+
starlette==0.27.0
|
111 |
+
sympy==1.13.3
|
112 |
+
syncer==2.0.3
|
113 |
+
tenacity==8.5.0
|
114 |
+
threadpoolctl==3.5.0
|
115 |
+
tiktoken==0.7.0
|
116 |
+
tokenizers==0.20.0
|
117 |
+
tomli==2.0.1
|
118 |
+
torch==2.4.1
|
119 |
+
tqdm==4.66.5
|
120 |
+
transformers==4.45.1
|
121 |
+
triton==3.0.0
|
122 |
+
typing-inspect==0.9.0
|
123 |
+
typing_extensions==4.12.2
|
124 |
+
uptrace==1.26.0
|
125 |
+
urllib3==2.2.3
|
126 |
+
uvicorn==0.23.2
|
127 |
+
watchfiles==0.20.0
|
128 |
+
websockets==13.1
|
129 |
+
wrapt==1.16.0
|
130 |
+
wsproto==1.2.0
|
131 |
+
yarl==1.13.1
|
132 |
+
zipp==3.20.2
|