import Agentlyimport requestsimport time
# Model Settingsagent_factory = Agently.AgentFactory()\.set_settings("model.OpenAI.auth", { "api_key": "" })\.set_settings("model.OpenAI.options", { "model": "gpt-3.5-turbo-16k" })# recommend using 16k or larger context model for this kind of tasks
# Download documentdocument_link = "https://raw.githubusercontent.com/Maplemx/Agently/main/README.md"document_content = ""response = requests.get(document_link)if response.status_code == 200:document_content = response.content.decode("utf-8")
# Work Settingspiece_length_control = 1000sleep_time = 5 # sleep for a while in case of reaching API request limit
# Chop documentchunks = document_content.split("\n\n")paragraphs = []paragraph_num = -1for chunk in chunks:if chunk.startswith("#"):paragraphs.append(chunk + "\n\n")paragraph_num += 1else:paragraphs[paragraph_num] += chunk + "\n\n"
text_pieces = []text_piece_num = 0for paragraph in paragraphs:if len(text_pieces) == 0:text_pieces.append(paragraph)else:if len(text_pieces[text_piece_num] + paragraph) > piece_length_control:text_pieces.append(paragraph)text_piece_num += 1else:text_pieces[text_piece_num] += paragraph
# Generate QA Pairsqa_pairs = []agent = agent_factory.create_agent()for text_piece in text_pieces:print("[Working on]: ", text_piece.split("\n")[0])result = agent\.input({"text": text_piece })\.instruct("Generate at least 5 question and answer pairs about {text}")\.output([{"question": ("String", "Question you may ask about {text}"),"answer": ("String", "Your answer to {question} according {text}"),}])\.start()qa_pairs.append({"origin_piece": text_piece,"qa_pairs": result,})print("[Done] Start next work in " + str(sleep_time) + " seconds.")time.sleep(sleep_time)print("[All Works Done]\n")
# Print QA Parisfor item in qa_pairs:print("[Origin Text Piece]: \n", item["origin_piece"], end="\n")for qa in item["qa_pairs"]:print("Question: ", qa["question"])print("Answer: ", qa["answer"], end="\n")print("------")

import PyPDF2
document_piece = []with open("./README.pdf", "rb") as file:reader = PyPDF2.PdfReader(file)for page in reader.pages:document_piece.append(page.extract_text())

import requests
document_link = "https://raw.githubusercontent.com/Maplemx/Agently/main/README.md"document_content = ""response = requests.get(document_link)if response.status_code == 200:document_content = response.content.decode("utf-8")

chunks = document_content.split("\n\n")paragraphs = []paragraph_num = -1for chunk in chunks:if chunk.startswith("#"):paragraphs.append(chunk + "\n\n")paragraph_num += 1else:paragraphs[paragraph_num] += chunk + "\n\n"

text_pieces = []text_piece_num = 0for paragraph in paragraphs:if len(text_pieces) == 0:text_pieces.append(paragraph)else:if len(text_pieces[text_piece_num] + paragraph) > piece_length_control:text_pieces.append(paragraph)text_piece_num += 1else:text_pieces[text_piece_num] += paragraph

[{"question": ("String", "Question you may ask about {text}"),"answer": ("String", "Your answer to {question} according {text}"),}]

qa_pairs = []agent = agent_factory.create_agent()for text_piece in text_pieces:print("[Working on]: ", text_piece.split("\n")[0])result = agent\.input({"text": text_piece })\.instruct("Generate at least 5 question and answer pairs about {text}")\.output([{"question": ("String", "Question you may ask about {text}"),"answer": ("String", "Your answer to {question} according {text}"),}])\.start()qa_pairs.append({"origin_piece": text_piece,"qa_pairs": result,})print("[Done] Start next work in " + str(sleep_time) + " seconds.")time.sleep(sleep_time)print("[All Works Done]\n")