import osfrom zhipuai import ZhipuAI
os.environ["ZHIPUAI_API_KEY"] = "your api key"client = ZhipuAI()

lecture_transcript_path =

import spacy
nlp = spacy.load("en_core_web_sm")
def preprocess_text(text):doc = nlp(text)sentences = [sent.text for sent in doc.sents]return sentences
def chunk_text(sentences, max_chunk_size=WORD_COUNT/50):chunks = []current_chunk = []current_length = 0
for sentence in sentences:sentence_length = len(sentence.split())if current_length + sentence_length > max_chunk_size:chunks.append(" ".join(current_chunk))current_chunk = []current_length = 0current_chunk.append(sentence)current_length += sentence_length
if current_chunk:chunks.append(" ".join(current_chunk))
return chunks

def summarize_chunk(chunk, summary_ratio):response = client.chat.completions.create(model="glm-4-long",messages=[{"role": "system","content": f"You are an assistant that reads a long lecture transcript and summarizes it to a short and concise note-taking format. The summary should be around {summary_ratio*100}% of the original length."},{"role": "user","content": chunk},],top_p=0.7,temperature=0.9)summarized_text = response.choices[0].message.contentreturn summarized_text
def summarize_text(text, summary_ratio):sentences = preprocess_text(text)
max_chunk_size = int(WORD_COUNT / 50)chunks = chunk_text(sentences, max_chunk_size)
summarized_chunks = []for chunk in chunks:summarized_chunk = summarize_chunk(chunk, summary_ratio)if summarized_chunk:summarized_chunks.append(summarized_chunk)
summarized_text = " ".join(summarized_chunks)
return summarized_text
first_summary_ratio = 0.2first_summarized_text = summarize_text(lecture_text, first_summary_ratio)
second_summary_ratio = 0.2final_summarized_text = summarize_text(first_summarized_text, second_summary_ratio)

markdown_notes = client.chat.completions.create(model="glm-4-long",messages=[{"role": "system","content": "Convert the summary to markdown format. Organize information into headings and subheadings, with no big paragraphs and no more than 5 bullet points under a subheading.",},{"role": "user","content": final_summarized_text,}],top_p=0.7,temperature=0.9)
with open("data/summarized_notes.md", "w") as file:file.write(markdown_notes.choices[0].message.content)