charlesdedampierre / bunkatopics Goto Github PK

Hello there,
I successfully embed and create topics, but when I get to bunka.get_topics to extract them I get the following error message which I cannot solve nor understand. What is the problem? Tahnk you!

2024-03-07 17:44:14 - Bunka - INFO - Computing the topics
INFO:Bunka:Computing the topics

TypeError Traceback (most recent call last)
in <cell line: 2>()
1 # Get the list of topics
----> 2 df_topics = bunka.get_topics(n_clusters = 15, name_length=5, min_count_terms = 2)
3 df_topics = df_topics[['topic_id', 'name', 'size']].copy()

4 frames
/usr/local/lib/python3.10/dist-packages/pandas/core/reshape/reshape.py in unstack(obj, level, fill_value, sort)
504 return _unstack_frame(obj, level, fill_value=fill_value, sort=sort)
505 else:
--> 506 return obj.T.stack(future_stack=True)
507 elif not isinstance(obj.index, MultiIndex):
508 # GH 36113

TypeError: DataFrame.stack() got an unexpected keyword argument 'future_stack'

@charlesdedampierre First and foremost, thank you for creating such an excellent library.

Description

I have successfully trained a topic model using your library on our dataset, which has been instrumental in uncovering various topics across the corpus. The model training and initial analysis have provided valuable insights into the dominant themes within my data.

However, we've identified a gap in documentation and usage guidelines, specifically regarding the application of the trained model to predict the cluster (or topic distribution) for new, unseen documents.

Is it possible to add a reference document on how to predict cluster membership with this library?

Bonjour Charles,

J'obtiens cette erreur :

Creating new labels for clusters: 0%| | 0/10 [00:00<?, ?it/s]

ValueError Traceback (most recent call last)
in <cell line: 13>()
11 }
12
---> 13 bourdieu_fig = bunka.visualize_bourdieu(
14 generative_model=llm,
15 x_left_words=["this is a positive content"],

12 frames

/usr/local/lib/python3.10/dist-packages/langchain/llms/huggingface_hub.py in _call(self, prompt, stop, run_manager, **kwargs)
112 response = self.client(inputs=prompt, params=params)
113 if "error" in response:
--> 114 raise ValueError(f"Error raised by inference API: {response['error']}")
115 if self.client.task == "text-generation":
116 # Text generation return includes the starter text.

ValueError: Error raised by inference API: Internal Server Error

Just installed bunkatopics and get this message error. I guess you may have forgotten the ipywidget in the dependency list