import os import sys import pandas as pd from typing import List import pinecone import difflib import cohere from langchain.embeddings.cohere import CohereEmbeddings from langchain.llms import Cohere from langchain.prompts import PromptTemplate from langchain.vectorstores import Pinecone, Qdrant from langchain.chains.question_answering import load_qa_chain sys.path.append(os.path.abspath('..')) from src.constants import SUMMARIZATION_MODEL, EXAMPLES_FILE_PATH PINECONE_API_KEY = os.environ.get("PINECONE_API_KEY") PINECONE_ENV = os.environ.get("PINECONE_ENV") COHERE_API_KEY = os.environ.get("COHERE_API_KEY") def replace_text(text): if text.startswith("The answer is "): text = text.replace("The answer is ", "", 1) return text def summarize( document: str, summary_length: str, summary_format: str, extractiveness: str = "high", temperature: float = 0.6, ) -> str: """ Generates a summary for the input document using Cohere's summarize API. Args: document (`str`): The document given by the user for which summary must be generated. summary_length (`str`): A value such as 'short', 'medium', 'long' indicating the length of the summary. summary_format (`str`): This indicates whether the generated summary should be in 'paragraph' format or 'bullets'. extractiveness (`str`, *optional*, defaults to 'high'): A value such as 'low', 'medium', 'high' indicating how close the generated summary should be in meaning to the original text. temperature (`str`): This controls the randomness of the output. Lower values tend to generate more “predictable” output, while higher values tend to generate more “creative” output. Returns: generated_summary (`str`): The generated summary from the summarization model. """ summary_response = cohere.Client(COHERE_API_KEY).summarize( text=document, length=summary_length, format=summary_format, model=SUMMARIZATION_MODEL, extractiveness=extractiveness, temperature=temperature, ) generated_summary = summary_response.summary return generated_summary def question_answer(input_document: str, history: List) -> str: """ Generates an appropriate answer for the question asked by the user based on the input document. Args: input_document (`str`): The document given by the user for which summary must be generated. history (`List[List[str,str]]`): A list made up of pairs of input question asked by the user & corresponding generated answers. It is used to keep track of the history of the chat between the user and the model. Returns: answer (`str`): The generated answer corresponding to the input question and document received from the user. """ pinecone.init( api_key=PINECONE_API_KEY, # find at app.pinecone.io environment=PINECONE_ENV # next to api key in console ) context = input_document # The last element of the `history` list contains the most recent question asked by the user whose answer needs to be generated. question = history[-1][0] word_list = context.split() texts = [" ".join(word_list[k : k + 200]) for k in range(0, len(word_list), 200)] # print(texts) embeddings = CohereEmbeddings( model="multilingual-22-12", cohere_api_key=COHERE_API_KEY ) context_index = Pinecone.from_texts(texts, embeddings, index_name="wiki-embed") prompt_template = """Text: {context} Question: {question} Answer the question based on the text provided. If the text doesn't contain the answer, reply that the answer is not available.""" PROMPT = PromptTemplate( template=prompt_template, input_variables=["context", "question"] ) # Generate the answer given the context chain = load_qa_chain( Cohere( model="command-xlarge-nightly", temperature=0, cohere_api_key=COHERE_API_KEY ), chain_type="stuff", prompt=PROMPT, ) relevant_context = context_index.similarity_search(question) answer = chain.run(input_documents=relevant_context, question=question) answer = answer.replace("\n", "").replace("Answer:", "") answer = replace_text(answer) return answer def generate_questions(input_document: str) -> str: co = cohere.Client(COHERE_API_KEY) prompt = f"""Write five different questions to test the understanding of the following text. The questions should be short answer, with one or two words each, and vary in difficulty from easy to hard. Provide the correct answer for each question after the question. Now write your own questions for this text: Text: {input_document} Question 1: (question_1) Answer: (answer_1) Question 2: (question_2) Answer: (answer_2) Question 3: (question_3) Answer: (answer_3) Question 4: (question_4) Answer: (answer_4) Question 5: (question_5) Answer: (answer_5)""" response = co.generate(model='command', prompt=prompt, temperature=2, max_tokens=1000, ) answer = response.generations[0].text.strip() print(answer) questions = answer.split('\n\n') print(questions) result = {} for question in questions: q, a = question.split('\n') result[q] = a.split(': ')[1] return answer def load_science(): examples_df = pd.read_csv(EXAMPLES_FILE_PATH) science_doc = examples_df["doc"].iloc[0] sample_question = examples_df["question"].iloc[0] return science_doc, sample_question def load_history(): examples_df = pd.read_csv(EXAMPLES_FILE_PATH) history_doc = examples_df["doc"].iloc[1] sample_question = examples_df["question"].iloc[1] return history_doc, sample_question def show_diff_html(seqm): """Unify operations between two compared strings seqm is a difflib.SequenceMatcher instance whose a & b are strings """ output = [] for opcode, a0, a1, b0, b1 in seqm.get_opcodes(): if opcode == 'equal': output.append(seqm.b[b0:b1]) elif opcode == 'insert': output.append(f"{seqm.b[b0:b1]}") # elif opcode == 'delete': # output.append(f"{seqm.a[a0:a1]}") elif opcode == 'replace': # output.append(f"{seqm.a[a0:a1]}") output.append(f"{seqm.b[b0:b1]}") else: if opcode == 'delete' or opcode == 'replace': continue raise RuntimeError("unexpected opcode") return ''.join(output) # define a function to paraphrase text using Cohere API def paraphrase(text): # create a cohere client with your API key client = cohere.Client(api_key=COHERE_API_KEY) # set the prompt for paraphrasing prompt = f"Rephrase this sentence in a different way: {text}" # generate a response using the multilingual-22-12 model response = client.generate( model="command-nightly", prompt=prompt, max_tokens=1000, ) # get the generated text rephrased_text = response[0].text print(rephrased_text) # compare the original and rephrased texts using difflib sm = difflib.SequenceMatcher(None, text, rephrased_text) html = show_diff_html(sm) return html if __name__ == "__main__": with open('sample_text.txt', 'r') as file: text = file.read() # summary = summarize(text, summary_length="short", summary_format="bullets") # print(summary) # answer = question_answer(text, [["what is photosynthesis", None]]) # print(answer) question = question_answer(text, ["Whats photosynthesis"]) print(question)