updated

app.py

@@ -1,5 +1,6 @@
-# for setting/extracting environment variables such as API keysimport os
+# for setting/extracting environment variables such as API keys 
 import os
+
 ### 1. For Web Scraping
 # for querying Financial Modelling Prep API
 from urllib.request import urlopen
@@ -7,7 +8,6 @@ import json
 
 ### 2. For Converting Scraped Text Into a Vector Store of Chunked Documents
 # for tokenizing texts and splitting them into chunks of documents
-from transformers import GPT2TokenizerFast
 from langchain.text_splitter import RecursiveCharacterTextSplitter
 # for turning documents into embeddings before putting them in vector store
 from langchain.embeddings import HuggingFaceEmbeddings
@@ -25,17 +25,16 @@ import gradio as gr
 
 fmp_api_key = os.environ['FMP_API_KEY']
 
+# initialize the default model for embedding the tokenized texts, the articles are stored in this embedded form in the vector database
+hf_embeddings = HuggingFaceEmbeddings()
+
 if os.path.exists("chromadb_earnings_transcripts_extracted"):
  os.system("rm -r chromadb_earnings_transcripts_extracted")
 if os.path.exists("earnings_transcripts_chromadb.zip"):
  os.system("rm earnings_transcripts_chromadb.zip")
-
 os.system("wget https://github.com/damianboh/test_earnings_calls/raw/main/earnings_transcripts_chromadb.zip")
 os.system("unzip earnings_transcripts_chromadb.zip -d chromadb_earnings_transcripts_extracted")
 
-# initialize the default model for embedding the tokenized texts, the articles are stored in this embedded form in the vector database
-hf_embeddings = HuggingFaceEmbeddings()
-
 chroma_db = Chroma(persist_directory='chromadb_earnings_transcripts_extracted/chromadb_earnings_transcripts',embedding_function=hf_embeddings)
 
 # Load the huggingface inference endpoint of an LLM model
@@ -45,7 +44,8 @@ model = "mistralai/Mistral-7B-Instruct-v0.1"
 # This is an inference endpoint API from huggingface, the model is not run locally, it is run on huggingface
 hf_llm = HuggingFaceHub(repo_id=model,model_kwargs={'temperature':0.5,"max_new_tokens":300})
 
-def source_question_answer(query,vectorstore:Chroma=chroma_db,llm:HuggingFaceHub=hf_llm):
+
+def source_question_answer(query:str,vectorstore:Chroma=chroma_db,llm:HuggingFaceHub=hf_llm):
  """
  Return answer to the query
  """
@@ -64,14 +64,16 @@ def source_question_answer(query,vectorstore:Chroma=chroma_db,llm:HuggingFaceHub
  source_title_3 = input_docs[2].metadata['title']
  source_title_4 = input_docs[3].metadata['title']
 
- return response, source_docs_1, source_docs_2, source_docs_3, source_docs_4, source_title_1, source_title_2, source_title_3, source_title_4
+ return response,source_docs_1 ,source_docs_2,source_docs_3,source_docs_4, source_title_1, source_title_2, source_title_3, source_title_4
+
 
 with gr.Blocks() as app:
+
  with gr.Row():
  gr.HTML("<h1>Chat with Tesla 2023 Earnings Calls Transcripts</h1>")
 
  with gr.Row():
- query = gr.Textbox("Is Elon happy about Tesla?", placeholder="Enter question here...", label="Enter question")
+ query = gr.Textbox("How is Tesla planning to expand?", placeholder="Enter question here...", label="Enter question")
  btn = gr.Button("Ask Question")
 
  with gr.Row():
@@ -98,5 +100,11 @@ with gr.Blocks() as app:
  with gr.Column():
  source_title_4 = gr.Markdown()
  source4 = gr.Textbox(label="Source Text 4")
+ 
+ query.submit(fn=source_question_answer, inputs=[query],
+ outputs=[answer, source1, source2, source3, source4, source_title_1, source_title_2, source_title_3, source_title_4])
+ 
+ btn.click(fn=source_question_answer, inputs=[query],
+ outputs=[answer, source1, source2, source3, source4, source_title_1, source_title_2, source_title_3, source_title_4])
 
 app.launch()