Upload 4 files

Added files from my local machine

Dockerfile

@@ -0,0 +1,11 @@
+FROM python:3.9
+RUN useradd -m -u 1000 user
+USER user
+ENV HOME=/home/user \
+    PATH=/home/user/.local/bin:$PATH
+WORKDIR $HOME/app
+COPY --chown=user . $HOME/app
+COPY ./requirements.txt ~/app/requirements.txt
+RUN pip install -r requirements.txt
+COPY . .
+CMD ["chainlit", "run", "app.py", "--port", "7860"]

app.py

@@ -0,0 +1,144 @@
+# chainlit run app.py -w
+# You can find this code for Chainlit python streaming here (https://docs.chainlit.io/concepts/streaming/python)
+# OpenAI Chat completion
+from dotenv import load_dotenv
+load_dotenv()
+
+import os
+import sys
+import getpass
+import nest_asyncio
+# import pandas as pd
+import faiss
+import openai
+
+import chainlit as cl  # importing chainlit for our app
+# https://docs.chainlit.io/api-reference/step-class#update-a-step
+# DEPRICATED: from chainlit.prompt import Prompt, PromptMessage  # importing prompt tools
+
+import llama_index
+from llama_index.core import Settings
+from llama_index.core import VectorStoreIndex
+from llama_index.core import StorageContext
+from llama_index.vector_stores.faiss import FaissVectorStore
+from llama_index.core import set_global_handler
+from llama_index.core.node_parser import MarkdownElementNodeParser
+from llama_index.llms.openai import OpenAI
+from llama_index.embeddings.openai import OpenAIEmbedding
+from llama_index.postprocessor.flag_embedding_reranker import FlagEmbeddingReranker
+from llama_parse import LlamaParse
+
+from openai import AsyncOpenAI  # importing openai for API usage
+
+os.environ["CUDA_VISIBLE_DEVICES"] = ""
+# GET KEYS
+LLAMA_CLOUD_API_KEY= os.getenv('LLAMA_CLOUD_API_KEY')
+OPENAI_API_KEY=os.getenv("OPENAI_API_KEY")
+
+"""
+os.environ["LLAMA_CLOUD_API_KEY"] = getpass.getpass("LLamaParse API Key:")
+os.environ["OPENAI_API_KEY"] = getpass.getpass("OpenAI API Key:")
+# os.environ["WANDB_API_KEY"] = getpass.getpass("WandB API Key: ")
+"""
+
+nest_asyncio.apply()
+
+# PARSING the pdf file
+parser = LlamaParse(
+    result_type="markdown",
+    verbose=True,
+    language="en",
+    num_workers=2,
+)
+
+nvidia_docs = parser.load_data(["./nvidia_2tables.pdf"])
+# Note: nvidia_docs contains only one file (it could contain more). nvidia_docs[0] is the pdf we loaded.
+print(nvidia_docs[0].text[:1000])
+
+# Getting Settings out of llama_index.core which is a major part of their v0.10 update!
+Settings.llm = OpenAI(model="gpt-3.5-turbo")
+Settings.embed_model = OpenAIEmbedding(model="text-embedding-3-small")
+
+# Using MarkdownElementNodeParser to help make sense of our Markdown objects so we can leverage the potentially structured information in the parsed documents.
+
+node_parser = MarkdownElementNodeParser(llm=OpenAI(model="gpt-3.5-turbo"), num_workers=8)
+
+nodes = node_parser.get_nodes_from_documents(documents=[nvidia_docs[0]])
+# Let's see what's in the metadata of the nodes:
+for nd in nodes:
+  print(nd.metadata)
+  for k,v in nd:
+    if k=='table_df':
+      print(nd)
+# Now we extract our `base_nodes` and `objects` to create the `VectorStoreIndex`.
+base_nodes, objects = node_parser.get_nodes_and_objects(nodes)
+
+# We could use the VectorStoreIndex from llama_index.core
+# Or we can use the llama_index FAISS llama-index-vector-stores-faiss
+# Trying the faiss, and setting its vectors' dimension.
+faiss_dim = 1536
+faiss_index = faiss.IndexFlatL2(faiss_dim) # default param overwrite=False, so it will append new vector.
+# Parameter overwrite=True suppresses appending a vector. 
+
+# Creating the FaissVectorStore and its recursicve_index_faiss
+llama_faiss_vector_store = FaissVectorStore(faiss_index=faiss_index)
+storage_context = StorageContext.from_defaults(vector_store=llama_faiss_vector_store)
+recursive_index_faiss = VectorStoreIndex(nodes=base_nodes+objects, storage_context=storage_context)
+
+# Now we can build our Recursive Query Engine with reranking!
+
+# We'll need to do a couple steps:
+# 1. Initalize our reranker using `FlagEmbeddingReranker` powered by the `BAAI/bge-reranker-large`.
+# 2. Set up our recursive query engine!
+
+reranker = FlagEmbeddingReranker(
+    top_n=5,
+    model="BAAI/bge-reranker-large",
+)
+
+recursive_query_engine = recursive_index_faiss.as_query_engine(
+    similarity_top_k=15,
+    node_postprocessors=[reranker],
+    verbose=True
+)
+
+"""
+# Create pandas dataframe to store query+generated response+added truth
+columns=["Query", "Response", "Truth"]
+gen_df = pd.DataFrame(columns=columns,dtype='str')
+"""
+
+# ChatOpenAI Templates
+system_template = """Use the following pieces of context to answer the user's question.
+If you don't know the answer, say that you don't know, do not try to make up an answer. 
+ALWAYS return a "SOURCES" part in your answer.
+The "SOURCES" part should be a reference to the source inside the document from which you got your answer.
+You are a helpful assistant who always speaks in a pleasant tone! """
+
+user_template = """ Think through your response step by step."""
+
+#user_query = "Who are the E-VP, Operations - and how old are they?"
+
+#response = recursive_query_engine.query(system_template + user_query + user_template)
+
+#str_resp ="{}".format(response)
+
+
+def retriever_resp(prompt):
+    import time
+    response = "this is my response"
+    time.sleep(5)
+    return response
+
+@cl.on_message  # marks a function that should be run each time the chatbot receives a message from a user
+async def main(message: cl.Message):
+    settings = cl.user_session.get("settings")
+
+    user_query = message.content
+#    prompt = system_template+user_query+user_template
+    response = recursive_query_engine.query(system_template + user_query + user_template)
+#    response = retriever_resp(prompt)
+#    print("AAA",user_query)
+    str_resp ="{}".format(response)
+    msg = cl.Message(content= str_resp)
+    await msg.send()

requirements.txt

@@ -0,0 +1,14 @@
+chainlit==1.0.401
+cohere==5.0.0a10
+openai==1.14.1
+python-dotenv==1.0.1
+faiss-cpu==1.8.0
+FlagEmbedding==1.2.5
+llama-index==0.10.20         
+llama-index-vector-stores-faiss==0.1.2
+llama-index-llms-openai==0.1.12
+llama-index-embeddings-openai==0.1.6
+llama-index-postprocessor-flag-embedding-reranker==0.1.2
+llama-parse==0.3.9
+# tiktoken==0.5.1
+# nest-asyncio==1.6.0