[TypeError] `guess_type` of media raising `Exception`

[Pk13055]

First Check

• I added a very descriptive title to this issue.
• I used the GitHub search to find a similar issue and didn't find it.
• I searched the FastAPI documentation, with the integrated search.
• I already searched in Google "How to X in FastAPI" and didn't find any information.
• I already read and followed all the tutorial in the docs and didn't find an answer.
• I already checked if it is not related to FastAPI but to Pydantic.
• I already checked if it is not related to FastAPI but to Swagger UI.
• I already checked if it is not related to FastAPI but to ReDoc.

Commit to Help

• I commit to help with one of those options 👆

Example Code

from fastapi import FastAPI, File, UploadFile
from fastapi.responses import FileResponse
import pandas as pd

from config import failure_path


app = FastAPI()

@app.post("/infer", tags=["inference"], response_class=FileResponse)
async def infer(input: UploadFile = File(...)) -> Union[FileResponse, dict]:
    data = None

    # Convert from CSV to pandas
    if input.content_type == "text/csv":
        data = await input.read()
        temp_io = io.StringIO(data.decode("utf-8"))
        data = pd.read_csv(temp_io)
    else:
        return {
            "success": False,
            "message": f"Content type {input.content_type} not supported (only CSV, 'text/csv' data allowed)",
        }

    print(f"Invoked with {data.shape[0]} records")

    # Do the prediction
    try:
        predictions = model_server.predict(data, data_schema)
        # Convert from dataframe to CSV
        with NamedTemporaryFile() as temp:
            predictions.to_csv(temp.name, index=False)
            temp.seek(0)
            return FileResponse(temp.name, media_type="text/csv")
    except Exception as err:
        # Write out an error file. This will be returned as the failureReason to the client.
        trc = traceback.format_exc()
        with open(failure_path, "w") as s:
            s.write("Exception during inference: " + str(err) + "\n" + trc)
        # Printing this causes the exception to be in the training job logs, as well.
        print("Exception during inference: " + str(err) + "\n" + trc, file=sys.stderr)
        # A non-zero exit code causes the training job to be marked as Failed.
        return {
            "success": False,
            "message": f"Exception during inference: {str(err)} (check failure file for more details)",
        }

Description

• run the server (using docker)
• Open postman and make call to /infer with the appropriate csv as input.
• Returns output as Internal Server Error

(PS - in order to test it, you can just use any random csv data, since the failing line should be caught by the exception anyways)

Operating System

Linux, macOS

Operating System Details

No response

FastAPI Version

0.86.0

Python Version

Python 3.10.8

Additional Context

Essentially, when I try to run this with some sample data, the line:

predictions = model_server.predict(data, data_schema)

throws an Exception, which should be caught by except block and subsequently return a JSON with the error message. However, I run into an Internal Server Error, with the following traceback:

172.17.0.1 - - [12/Nov/2022:15:14:11 +0000] "POST /infer HTTP/1.1" 500 21 "-" "curl/7.84.0"
Invoked with 17 records
Exception during inference: "['nan'] not in index"
Traceback (most recent call last):
  File "/opt/app/./inference_app.py", line 74, in infer
    predictions = model_server.predict(data, data_schema)
  File "/opt/app/./algorithm/model_server.py", line 76, in predict
    preds_df["prediction"] = pd.DataFrame(preds_df[class_names], columns = class_names).idxmax(axis=1)
  File "/usr/local/lib/python3.10/site-packages/pandas/core/frame.py", line 3810, in __getitem__
    indexer = self.columns._get_indexer_strict(key, "columns")[1]
  File "/usr/local/lib/python3.10/site-packages/pandas/core/indexes/base.py", line 6111, in _get_indexer_strict
    self._raise_if_missing(keyarr, indexer, axis_name)
  File "/usr/local/lib/python3.10/site-packages/pandas/core/indexes/base.py", line 6174, in _raise_if_missing
    raise KeyError(f"{not_found} not in index")
KeyError: "['nan'] not in index"

INFO:      - "POST /infer HTTP/1.0" 500 Internal Server Error
ERROR:    Exception in ASGI application
Traceback (most recent call last):
  File "/usr/local/lib/python3.10/site-packages/uvicorn/protocols/http/httptools_impl.py", line 419, in run_asgi
    result = await app(  # type: ignore[func-returns-value]
  File "/usr/local/lib/python3.10/site-packages/uvicorn/middleware/proxy_headers.py", line 78, in __call__
    return await self.app(scope, receive, send)
  File "/usr/local/lib/python3.10/site-packages/fastapi/applications.py", line 270, in __call__
    await super().__call__(scope, receive, send)
  File "/usr/local/lib/python3.10/site-packages/starlette/applications.py", line 124, in __call__
    await self.middleware_stack(scope, receive, send)
  File "/usr/local/lib/python3.10/site-packages/starlette/middleware/errors.py", line 184, in __call__
    raise exc
  File "/usr/local/lib/python3.10/site-packages/starlette/middleware/errors.py", line 162, in __call__
    await self.app(scope, receive, _send)
  File "/usr/local/lib/python3.10/site-packages/starlette/middleware/exceptions.py", line 75, in __call__
    raise exc
  File "/usr/local/lib/python3.10/site-packages/starlette/middleware/exceptions.py", line 64, in __call__
    await self.app(scope, receive, sender)
  File "/usr/local/lib/python3.10/site-packages/fastapi/middleware/asyncexitstack.py", line 21, in __call__
    raise e
  File "/usr/local/lib/python3.10/site-packages/fastapi/middleware/asyncexitstack.py", line 18, in __call__
    await self.app(scope, receive, send)
  File "/usr/local/lib/python3.10/site-packages/starlette/routing.py", line 680, in __call__
    await route.handle(scope, receive, send)
  File "/usr/local/lib/python3.10/site-packages/starlette/routing.py", line 275, in handle
    await self.app(scope, receive, send)
  File "/usr/local/lib/python3.10/site-packages/starlette/routing.py", line 65, in app
    response = await func(request)
  File "/usr/local/lib/python3.10/site-packages/fastapi/routing.py", line 264, in app
    response = actual_response_class(content, **response_args)
  File "/usr/local/lib/python3.10/site-packages/starlette/responses.py", line 299, in __init__
    media_type = guess_type(filename or path)[0] or "text/plain"
  File "/usr/local/lib/python3.10/site-packages/starlette/responses.py", line 35, in guess_type
    return mimetypes_guess_type(url, strict)
  File "/usr/local/lib/python3.10/mimetypes.py", line 307, in guess_type
    return _db.guess_type(url, strict)
  File "/usr/local/lib/python3.10/mimetypes.py", line 122, in guess_type
    url = os.fspath(url)
TypeError: expected str, bytes or os.PathLike object, not dict

Now, I can/will fix the issue with the predict function, but even then, this error shows up (ie, circumventing the function for the time being). I don't understand where exactly the "type" of the response is being used, and what "url" is in this context.

[JarroVGIT]

where does failure_path comes from? I suspect that is failing..

[tiangolo]

Good catch @JarroVGIT ! 🤓😎

Yep, it seems you have an error in your exception handling with the failure_path.

[Pk13055]

@JarroVGIT No that is not failing since failure path is just a global variable imported from another config. (I have edited the code sample to reflect this). Additionally, even if it had been that by chance, wouldn't that particular line show up in the traceback?

[JarroVGIT]

Fair point. Unfortunately, I am unable to reproduce it on my end, so I am at a loss. It might help to set a breakpoint on starlette/responses.py line 35, to see what the value of url is when this error occurs. Maybe that would give a pointer on why that is a dict rather than a URL?

[Pk13055]

@JarroVGIT Okay this is very odd, but fixing the issue with the predictions function seems to have solved the problem. Additionally, there was an issue with the NamedTemporaryFile being out of context when returned, so I had to inject the file as a dependency. But apart from that, the error seems to have magically disappeared.
(PS - below is the temp file injection, as a reference for anyone who stumbles across this)

async def gen_temp_file(ext: str = ".csv"):
    """Generate a temporary file with a given extension"""
    with NamedTemporaryFile(suffix=ext, delete=True) as temp_file:
        yield temp_file.name

@app.post("/infer", tags=["inference"], response_class=FileResponse)
async def infer(
    input: UploadFile = File(...), temp=Depends(gen_temp_file)
) -> Union[FileResponse, dict]:
    """Do an inference on a single batch of data. In this sample server, we take data as CSV, convert
    it to a pandas data frame for internal use and then convert the predictions back to CSV (which really
    just means one prediction per line, since there's a single column.
    """
    data = None
    # rest of the function remains the same
    ...

(Closing the issue)

[tiangolo]

Thanks for the help here @JarroVGIT ! 👏 🙇

Thanks for reporting back and closing the issue @Pk13055 👍