Python

Python

Getting started

Overview
Overview
Import data
Import data

Languages

Python
Python
Formulas
Formulas

Reference

Support
Support

Python

Quadratic uses pyodide under the hood to provide a Python scripting environment directly in the browser. No more headaches around setting up or sharing a development environment. If it works for you, it’ll work when you share it.

Cells

icon
Cell positioning within a spreadsheet is explained in the Overview.

Referencing individual cells

To reference an individual cell, use the global function cell (or c for short) which returns the cell value.

cell(2, 3) # Returns the value of the cell

c(2, 3) # Returns the value of the cell

The resulting value can be used directly in a Pythonic way.

c(0, 0) + c(0, 1) # Adds cell 0, 0 and cell 0, 1

c(0, 0) == c(0, 1) # Is cell 0, 0 equal to cell 0, 1 ?

When a cell depends on other cells and the other cells update, this dependent cell will also update — just like in Excel.

Referencing a range of cells

To reference a range of cells, use the global function cells which returns a Pandas DataFrame.

cells((0, 0), (2, 2)) # Returns a DataFrame with the cell values

If the first row of cells is a header, you can set first_row_header as an argument.

cells((2, 2), (7, 52), first_row_header=True)

As an example, this code references a table of expenses, filters it based on a user-specified column, and returns the resulting DataFrame to the spreadsheet.

# Pull the full expenses table in as a DataFrame
expenses_table = cells((2, 2), (7, 52), first_row_header=True)

# Take user input at a cell (Category = "Gas")
category = cell(10, 0)

# Filter the full expenses table to the "Gas" category, return the resulting DataFrame
expenses_table[expenses_table["Category"] == category]

Returning a Python list of values

You can return multiple values to the spreadsheet by returning a list of values in Python. For example:

# Loop over a list of values
result = []
for x in range(0, 5):
    result.append(x)

# Return it to the spreadsheet
result

By default arrays are expanded vertically. To have an array expand horizontally, wrap it in a second array, e.g. [result]

# result: [0, 1, 2, 3, 4]

# Return it to the spreadsheet **vertically**
result

# [0][ ][ ][ ][ ]
# [1][ ][ ][ ][ ]
# [2][ ][ ][ ][ ]
# [3][ ][ ][ ][ ]
# [4][ ][ ][ ][ ]

# Return it to the spreadsheet horizontally
[result]

# [0][1][2][3][4]
# [ ][ ][ ][ ][ ]
# [ ][ ][ ][ ][ ]
# [ ][ ][ ][ ][ ]
# [ ][ ][ ][ ][ ]
0
1
2
3
4
5

Returning a Pandas DataFrame

The following code creates a DataFrame of 15 rows, by 4 columns filled with random numbers between 0 and 100. It is the last expression, so it is returned to the the spreadsheet.

# pandas and numpy are preloaded by default!
import pandas as pd
import numpy as np

# using numpy's randint, return df to Grid
pd.DataFrame(np.random.randint(
    0,
    100,
    size=(15, 4),
))

Fetching data

You can make an external HTTP request directly from a Quadratic Cell using js.fetch. See pyodide’s docs for detailed information on using the fetch API in Python

icon
Note: requests are made client-side from your web browser and are subject to CORS limitations.

Example: making an API call to OpenAPI with a prompt

import json
import pyodide

api_key = c(0,3)
model = c(0, 5)
max_tokens = c(0, 7)
temperature = c(0, 9)
top_p = c(0, 11)
prompt = c(0, 13)

# Make API Request
response = await pyodide.http.pyfetch(
  'https://api.openai.com/v1/completions',
    method= "POST",
    headers = {
        "Content-Type": "application/json",
        "Authorization": "Bearer {}".format(api_key),
    },
    body = json.dumps({
        "model": str(model),
        "prompt": str(prompt),
        "max_tokens": int(max_tokens),
        "temperature":  float(temperature),
        "top_p": float(top_p)
    })
)

# debug
# print(await response.string())

if (response.status == 401):
    raise Exception("Check your OpenAI API key")
else:
    response_json = await response.json()
    result = response_json["choices"][0]["text"].strip()

Example: pulling stock data from the Polygon API

import json
import pyodide

api_key = c(0,3)
model = c(0, 5)
max_tokens = c(0, 7)
temperature = c(0, 9)
top_p = c(0, 11)
prompt = c(0, 13)

# Make API Request
response = await pyodide.http.pyfetch(
  'https://api.openai.com/v1/completions',
    method= "POST",
    headers = {
        "Content-Type": "application/json",
        "Authorization": "Bearer {}".format(api_key),
    },
    body = json.dumps({
        "model": str(model),
        "prompt": str(prompt),
        "max_tokens": int(max_tokens),
        "temperature":  float(temperature),
        "top_p": float(top_p)
    })
)

# debug
# print(await response.string())

if (response.status == 401):
    raise Exception("Check your OpenAI API key")
else:
    response_json = await response.json()
    result = response_json["choices"][0]["text"].strip()

Packages

The following libraries are included by default:

You can import them like any other native Python package.

import pandas as pd

Additionally, you can use Micropip to install additional Python packages (and their dependencies).

import micropip

# `await` is necessary to wait until the package is available
await micropip.install("faker")

# import installed package
from faker import Faker

# use the package!
fake = Faker()
fake.name()

This only works for packages that are either pure Python or for packages with C extensions that are built in Pyodide. If a pure Python package is not found in the Pyodide repository it will be loaded from PyPI. Learn more about how packages work in Pyodide.