ChatBot for EvaDB

convert_instructions_to_pdf.py

@@ -0,0 +1,49 @@
+import os
+import subprocess
+from pathlib import Path
+
+DOC_ROOT_DIR = "docs/source/"
+
+def convert_rst_to_pdf(rst_file_lst, output_path):
+
+    for i,rst_file in enumerate(rst_file_lst):
+
+        try:
+            # Run rst2pdf command using subprocess
+            file_name = rst_file.split("/")[-1]
+            file_name = file_name[:-4]
+
+            if not Path(output_path+file_name+".pdf").is_file():
+                output_file = output_path+file_name+".pdf"
+
+                subprocess.run(['rst2pdf', rst_file, '-o', output_file], check=True)
+                print(f"Conversion successful. PDF saved to {output_path+file_name}")
+
+        except subprocess.CalledProcessError as e:
+            print(f"Conversion failed. Error: {e}")
+
+        if i%10 == 0:
+            print(f"Completed {i} files")
+
+
+
+
+
+def get_all_files_recursively(directory):
+    directory_path = Path(directory)
+    all_files = list(directory_path.rglob('*'))
+    return all_files
+
+
+def convert_instr_to_pdf():
+
+    all_files_in_directory = get_all_files_recursively(DOC_ROOT_DIR)
+    all_files_in_directory = [str(f) for f in all_files_in_directory]
+    all_files_in_directory = [f for f in all_files_in_directory if f.endswith(".rst")]
+
+    print(f"TOtal nos of files: {len(all_files_in_directory)}")
+
+    os.mkdir("doc_pdf/")
+    convert_rst_to_pdf(all_files_in_directory, 'doc_pdf/')
+
+

run_chatbot.py

@@ -0,0 +1,51 @@
+import argparse
+from convert_instructions_to_pdf import *
+import os
+from llama_index import VectorStoreIndex, SimpleDirectoryReader
+
+
+parser = argparse.ArgumentParser(description='A chatbot for EvaDB')
+
+parser.add_argument("--reload_docs", help="Reload all the RST files and rebuild the index")
+parser.add_argument("--open_ai_key", help="OpenAI key")
+
+args = parser.parse_args()
+
+reload_docs_flag = args.reload_docs
+open_ai_key = args.open_ai_key
+
+os.environ["OPENAI_API_KEY"] = open_ai_key
+
+
+if reload_docs_flag == "True":
+    print("Started reloading the docs")
+    convert_instr_to_pdf()
+    print("Finished reloading docs")
+
+#llama
+print("Creating the Llama index.....")
+documents = SimpleDirectoryReader("doc_pdf/").load_data()
+index = VectorStoreIndex.from_documents(documents)
+
+query_engine = index.as_query_engine()
+
+print("Finished Creating the Llama index....")
+
+
+while(True):
+    print("Welcome to EvaDB ")
+    print("Please enter your query. Enter 1 to exit the chatbot")
+
+    user_query = input()
+
+    if user_query == "1":
+        break
+
+    else:
+        response = query_engine.query(user_query)
+        print(response)
+
+
+print("Terminating the chatbot.. See you next time!")
+
+