Create app.py

app.py

@@ -0,0 +1,41 @@
+from PyPDF2 import PdfReader
+from langchain.embeddings.openai import OpenAIEmbeddings
+from langchain.text_splitter import CharacterTextSplitter
+from langchain.vectorstores import ElasticVectorSearch, Pinecone, Weaviate, FAISS
+from datasets import load_dataset
+import os
+import gradio as gr
+from transformers import pipeline
+from datasets import DatasetDict
+
+os.environ["OPENAI_API_KEY"] = "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"
+
+reader = load_dataset("orderlymirror/The_48_Laws_Of_Power")
+reader = PdfReader('/content/drive/MyDrive/llm/48lawsofpower.pdf')
+
+raw_text = ''
+for i, page in enumerate(reader.pages):
+    text = page.extract_text()
+    if text:
+        raw_text += text
+        
+        
+text_splitter = CharacterTextSplitter(
+    separator = "\n",
+    chunk_size = 1000,
+    chunk_overlap  = 200,
+    length_function = len,
+)
+texts = text_splitter.split_text(raw_text)
+
+embeddings = OpenAIEmbeddings()
+
+docsearch = FAISS.from_texts(texts, embeddings)
+
+from langchain.chains.question_answering import load_qa_chain
+from langchain.llms import OpenAI
+        
+        
+query = "“Can you give me an example from history where the enemy was crushed totallyfrom the book?"
+docs = docsearch.similarity_search(query)
+chain.run(input_documents=docs, question=query)