fix: 🐞 updating poetry lock file version and fixing numpy,pandas,maplotlib and python versions compatibility

[onuralpszr]

Description

• I upgraded poetry to 1.8.4 and I got proper numpy version ( I make sure I delete poetry caches and everything and re-tried with warm cache and without cache and both works and I was able to import it)

Attempt to Fix #1688 and update CI tests for make sure we cover this test cases too

[onuralpszr]

@LinasKo I fixed poetry installations for all versions and fixes for matplotlib,pandas too. I also include extra tests into CI such as "import test" and full pytest for each poetry envs. it should be better for good.

[LinasKo]

I've tested it, and I couldn't find anything wrong.

As an extra check, I verified that a relaxed-constraint version of inference can be installed and run alongside this.

pip install git+https://github.com/roboflow/inference.git@linas/allow-latest-rc-supervision

Test code

import cv2
import numpy as np
from inference import get_model
import supervision as sv


!wget -O dog.jpeg https://media.roboflow.com/notebooks/examples/dog.jpeg -q
image = cv2.imread("dog.jpeg")
assert image is not None

black_image = np.zeros_like(image)


box_model = get_model("yolov8n-640")
result = box_model.infer(image)[0]
detections = sv.Detections.from_inference(result)
assert len(detections) > 0

result = box_model.infer(black_image)[0]
detections = sv.Detections.from_inference(result)
assert len(detections) == 0


seg_model = get_model("yolov8n-seg-640")
result = seg_model.infer(image)[0]
detections = sv.Detections.from_inference(result)
assert len(detections) > 0

result = seg_model.infer(black_image)[0]
detections = sv.Detections.from_inference(result)
assert len(detections) == 0


kp_model = get_model("yolov8n-pose-640")
result = kp_model.infer(image)[0]

keypoints = sv.KeyPoints.from_inference(result)
assert len(keypoints) > 0

result = kp_model.infer(black_image)[0]
keypoints = sv.KeyPoints.from_inference(result)
assert len(keypoints) == 0