How can I extract data from documents in JSON Output Format?

Hello,

I cannot find how can I extract data from documents in a JSON Output Format?

Just like https://hackernoon.com/unlocking-structured-json-data-with-langchain-and-gpt-a-step-by-step-tutorial for OpenAI.

I have tried to prompt through the following lines:

```
document_query = "Crear un esquema basado en este texto: " + document[0].page_content
_input = prompt.format_prompt(question=document_query)
output = chat_model(_input.to_string())
print(output)
parsed = parser.parse(output)
texts_json.append(json.dumps(parsed.dict()).encode('utf-8').decode('unicode-escape'))
```

However I cannot get just the JSON output. I get long unparsed conversations (for example, human-assistant dialogues if I use vicuna ggml-vicuna-13b-4bit.bin).

Could somebody tell me if Llamacpp can already be used for those goals?

Thanks in advance.