Convert Ollama's Structured Output from JSON to Python DF

penguin_pam

I'm having a devil of a time converting structured output from Ollama's new API to a df. The following is JSON output from the Data Extraction example on the Ollama page on structured outputs ( https://ollama.com/blog/structured-outputs ): pets=[ Pet(name='Luna', animal='cat', age=5, color='grey', favorite_toy='yarn'), Pet(name='Loki', animal='cat', age=2, color='black', favorite_toy='tennis balls') ] After trying several different python-based...

Originally asked by BBJonz on Stack Exchange (stackoverflow) (CC BY-SA 4.0).

uphilltech

Great question — answering as the UphillTech team:

Well, it took a while but here's what worked. I had to spend some quality time learning more about JSON nested objects and how to process them. I found this website particularly helpful. The code is: temp = json.loads(response.message.content) df = pd.json_normalize(temp['pets']) Here's the output as a pandas df: name animal age color favorite_toy 0 Luna cat 5 grey yarn 1 Loki cat 2 black tennis balls

Full answer by BBJonz: https://stackoverflow.stackexchange.com/a/79310505 (CC BY-SA 4.0).

Need it done for your business? UphillTech.