Update with more models

[simonw]

https://app.endpoints.anyscale.com/explorer now lists:

I only have these currently:

https://github.com/simonw/llm-anyscale-endpoints/blob/86f395560530374da237092a500fa06b2022faeb/llm_anyscale_endpoints.py#L4-L10

Would be better if the plugin could retrieve the list from somewhere.

[simonw]

Full list of models: https://docs.endpoints.anyscale.com/category/supported-models

[simonw]

I tried Meta-Llama/Llama-Guard-7b but got this error:

endpoints llm -m Meta-Llama/Llama-Guard-7b hello Error: rayllm.common.models.PromptFormatDisabledError: This model doesn't support chat completions. Please use the completions endpoint instead. (Request ID: TImvQaY61Ge0wkfMmPV56zDPFJh1_pUHjBJI9_0UYOU)

[simonw]

Mixtral really is very good:

llm -m mistralai/Mixtral-8x7B-Instruct-v0.1 \
  'Write a Python function that accepts a URL to a CSV file, downloads it and loads it into a SQLite database, creating a table with the correct columns'

Here's a Python function that accomplishes the task you described, using the requests, csv, sqlite3, and io libraries:

import requests
import csv
import sqlite3
from io import StringIO

def download_csv_to_sqlite(url):
    # Download the CSV file from the URL
    response = requests.get(url)
    response.raise_for_status()

    # Load the CSV data into a StringIO buffer
    csv_buffer = StringIO(response.text)

    # Connect to a new SQLite database in memory
    conn = sqlite3.connect(":memory:")

    # Create a new cursor for executing SQL commands
    cursor = conn.cursor()

    # Read the first row of the CSV file to determine the column names
    csv_reader = csv.reader(csv_buffer)
    columns = next(csv_reader)

    # Create a new table in the SQLite database with the correct columns
    column_defs = ", ".join([f"{col} TEXT" for col in columns])
    cursor.execute(f"CREATE TABLE csv_table ({column_defs});")

    # Insert the remaining rows of the CSV file into the table
    for row in csv_reader:
        value_defs = ", ".join(["?" for _ in row])
        cursor.execute(f"INSERT INTO csv_table VALUES ({value_defs});", row)

    # Commit the changes and close the connection
    conn.commit()
    conn.close()

# Example usage
url = "https://example.com/data.csv"
download_csv_to_sqlite(url)

This function downloads the CSV file from the given URL, loads it into a SQLite database in memory, creates a new table with the correct columns, and inserts all the rows from the CSV file into the table.

Note that this function uses an in-memory SQLite database for simplicity, but you can modify it to use a persistent database file if needed. Also, the function assumes that the CSV file contains text data, but you can modify it to handle other data types as necessary.