How to mock multipart/form-data with pytest?

Question

I have a POST endpoint with fastapi that goes a little like this. This takes in an excel file and turns it into a pandas dataframe.

@router.post(
"/endpoint/upload_excel",
status_code=status.HTTP_200_OK
)
async def upload_excel(user_excel_file: UploadFile = File(...)):
     """
     read user Excel file from multipart/form-data into pd dataframe
     """

user_excel_bytes = await user_excel_file.read()
user_excel_bytes_io = io.BytesIO(user_excel_bytes)
user_lineage = pd.read_excel(user_excel_bytes_io)

This function continues to manipulate the pandas dataframe, such as sorting columns, checking for nulls, and checking data types.

How could I test this function with pytest in python? I believe I need to mock the the multipart/form-data (user_excel_file: UploadFile = File(...)). I want to test for the edge cases like catching nulls or having ints where strings should be, but I don't want to create a separate test excel file for each case.

If the input to this function was a pandas dataframe, I could easily mock that! But how can I mock the input excel?

If I'm understanding you correctly (please say if this doesn't answer your question). You can avoid creating real Excel files by creating an in-memory Excel file using pandas and io.BytesIO, then wrapping it in a FastAPI UploadFile mock. This lets you mock any Excel content dynamically without creating real files. You just build a DataFrame for each edge case and convert it to a mock UploadFile. — ninadepina
– ninadepina, Commented Oct 23 at 20:59
@ninadepina That sounds like what I want! Do you know any resources on how to create an in-memory Excel file with pandas and io.BytesIO? — nijj67
– nijj67, Commented Oct 23 at 21:20
I can't quickly find one source that explains it all.. You should puzzle some sources together. The links I found in the few minutes of googling: Some Medium source about simplifying file exports with bytesio and pandas and Another stack overflow issue — ninadepina
– ninadepina, Commented Oct 23 at 21:39
I'd be most tempted to keep a collection of Excel files and iterate through 'em, even if you dynamically generate some, especially because this enables collecting test cases — ti7
– ti7, Commented Oct 23 at 22:11
The easiest solution is to stick excel-files that are examples of what you want as fixtures somewhere, then use the built-in testclient in FastAPI to submit them to your endpoint. There is no need to mock or generate anything; use real dependencies and examples that show what you want to test. — MatsLindh
– MatsLindh, Commented Oct 24 at 11:57

Dmitri Galkin · Accepted Answer · 2025-12-12 08:43:49Z

0

Create DataFrames in tests and convert to excel bytes without like real file:

import io
import pandas as pd

def df_to_excel_bytes(df):
    buf = io.BytesIO()
    df.to_excel(buf, index=False)
    buf.seek(0)
    return buf.read()

def test_nulls():
    df = pd.DataFrame({"name": ["Alice", None]})
    response = client.post(
        "/endpoint/upload_excel",
        files={"user_excel_file": ("test.xlsx", df_to_excel_bytes(df))}
    )

For fake data I use Faker.

answered Dec 12 at 8:43

Dmitri Galkin

6175 silver badges12 bronze badges

Sign up to request clarification or add additional context in comments.

Collectives™ on Stack Overflow

How to mock multipart/form-data with pytest?

1 Answer 1

Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related