Python script to list all ollama models from their library site

ollama_models.py

#!/usr/bin/env python

from bs4 import BeautifulSoup
import requests
import json


def main():
    soup = BeautifulSoup(requests.get("https://ollama.com/library?sort=popular").text)

    popular_models = [h2.text.strip() for h2 in soup.find_all("h2")]
    popular_models = [m for m in popular_models if m]

    model_tags = []

    for i, model_name in enumerate(popular_models):
        response = requests.get(f"https://ollama.com/library/{model_name}/tags")
        response.raise_for_status()
        soup = BeautifulSoup(
            response.text,
        )
        for a in soup.find_all("a"):
            if not a["href"].startswith(f"/library/{model_name}:"):
                continue

            metadata_spans = a.parent.parent.find_all("span")
            metadata_str = list(metadata_spans[0].children)[-1].text.strip()
            _, size, uploaded = metadata_str.split("•")

            result = {
                "model": model_name,
                "tag": a.text.strip(),
                "popularity": i,
                "hash": metadata_spans[-1].text.strip(),
                "size": size.strip(),
                "uploaded": uploaded.strip(),
            }
            model_tags.append(result)

    with open("ollama_models.json", "w") as f:
        json.dump(model_tags, f)


if __name__ == "__main__":
    main()