Other Text Summarization libraries

 Given the challenges with gensim and transformers in Azure, there are a few alternative libraries that can perform text summarization and may work more seamlessly:

1. Sumy: A straightforward library for extractive summarization. Sumy offers several algorithms like LSA (Latent Semantic Analysis) and LexRank, which might be suitable for basic summarization tasks.

   python
   from sumy.parsers.plaintext import PlaintextParser
   from sumy.nlp.tokenizers import Tokenizer
   from sumy.summarizers.lsa import LsaSummarizer
   
   def summarize_text(text):
       parser = PlaintextParser.from_string(text, Tokenizer("english"))
       summarizer = LsaSummarizer()
       summary = summarizer(parser.document, 3)  # Adjust the number of sentences
       return " ".join(str(sentence) for sentence in summary)
   

2. spaCy: While spaCy doesn’t have a built-in summarizer, you can use a combination of spaCy's text processing along with libraries like pytextrank to perform extractive summarization.

   python
   import spacy
   import pytextrank
   
   nlp = spacy.load("en_core_web_sm")
   nlp.add_pipe("textrank")
   
   def summarize_text(text):
       doc = nlp(text)
       return "\n".join([str(p) for p in doc._.phrases[:5]])  # Adjust for number of phrases
   

3. TextRank4ZH: Although this is a Chinese-focused library, it supports multi-language summarization based on the TextRank algorithm.

You might try these libraries in your Azure environment, as they tend to have lighter dependencies compared to gensim and transformers. Let me know if you'd like more guidance on setting up any specific library!