beerds/server/main.py

174 lines
4.6 KiB
Python

from pathlib import Path
from PIL import Image
from starlite import (
Controller,
StaticFilesConfig,
get,
post,
Body,
MediaType,
RequestEncodingType,
Starlite,
UploadFile,
Template,
TemplateConfig,
)
from starlite.contrib.jinja import JinjaTemplateEngine
import io
import os
import json
import torch
from torchvision import transforms # type: ignore
import torch.nn.functional as F
os.environ["CUDA_VISIBLE_DEVICES"] = "-1"
def load_model(model_path, device="cpu"):
model = torch.load(model_path, map_location=device, weights_only=False)
model.eval()
return model
with open("server/meta/images.json", "r") as f:
IMAGES = json.loads(f.read())
DOG_MODEL = load_model("server/models/dogs_model.pth")
CAT_MODEL = load_model("server/models/cats_model.pth")
with open("server/meta/labels_dogs.json", "r") as f:
data_labels = f.read()
labels_dogs = json.loads(data_labels)
with open("server/meta/labels_cats.json", "r") as f:
data_labels = f.read()
labels_cats = json.loads(data_labels)
def predict_image(image, model, device="cuda") -> list[tuple]:
img_size = (180, 180)
preprocess = transforms.Compose(
[
transforms.Resize(img_size),
transforms.ToTensor(),
transforms.Normalize((0.5, 0.5, 0.5), (0.5, 0.5, 0.5)),
]
)
input_tensor = preprocess(image)
input_batch = input_tensor.unsqueeze(0).to(device) # Добавляем dimension для батча
with torch.no_grad():
output = model(input_batch)
probabilities = torch.nn.functional.softmax(output[0], dim=0)
k = 5
topk_probs, predicted_idx = torch.topk(probabilities, k)
data = []
for i in range(k):
data.append((predicted_idx[i].item(), float(topk_probs[i].item())))
return data
class BeerdsController(Controller):
path = "/beerds"
@post("/dogs")
async def beerds_dogs(
self, data: UploadFile = Body(media_type=RequestEncodingType.MULTI_PART)
) -> dict:
body = await data.read()
img_file = Image.open(io.BytesIO(body))
predicted_data = predict_image(img_file, DOG_MODEL, "cpu")
results = {}
images = []
for d in predicted_data:
predicted_idx, probabilities = d
predicted_label = labels_dogs[str(predicted_idx)]
name = predicted_label.replace("_", " ")
images.append({"name": name, "url": IMAGES[name]})
results[probabilities] = name
return {
"results": results,
"images": images,
}
@post("/cats")
async def beerds_cats(
self, data: UploadFile = Body(media_type=RequestEncodingType.MULTI_PART)
) -> dict:
body = await data.read()
img_file = Image.open(io.BytesIO(body))
predicted_data = predict_image(img_file, CAT_MODEL, "cpu")
results = {}
for d in predicted_data:
predicted_idx, probabilities = d
predicted_label = labels_cats[str(predicted_idx)]
results[probabilities] = predicted_label
return {
"results": results,
"images": [],
}
class BaseController(Controller):
path = "/"
@get("/")
async def main(self) -> Template:
return Template(name="index.html")
@get("/cats")
async def cats(self) -> Template:
return Template(name="cats.html")
@get("/sitemap.xml", media_type=MediaType.XML)
async def sitemaps(self) -> bytes:
return """<?xml version="1.0" encoding="UTF-8"?>
<urlset
xmlns="http://www.sitemaps.org/schemas/sitemap/0.9"
xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
xsi:schemaLocation="http://www.sitemaps.org/schemas/sitemap/0.9
http://www.sitemaps.org/schemas/sitemap/0.9/sitemap.xsd">
<!-- created with Free Online Sitemap Generator www.xml-sitemaps.com -->
<url>
<loc>https://xn-----6kcp3cadbabfh8a0a.xn--p1ai/</loc>
<lastmod>2025-04-21T19:01:03+00:00</lastmod>
</url>
<url>
<loc>https://xn-----6kcp3cadbabfh8a0a.xn--p1ai/cats</loc>
<lastmod>2025-04-21T19:01:03+00:00</lastmod>
</url>
</urlset>
""".encode()
@get("/robots.txt", media_type=MediaType.TEXT)
async def robots(self) -> str:
return """
User-agent: *
Allow: /
Sitemap: https://xn-----6kcp3cadbabfh8a0a.xn--p1ai/sitemap.xml
"""
app = Starlite(
debug=True,
route_handlers=[BeerdsController, BaseController],
static_files_config=[
StaticFilesConfig(directories=[Path("server/static")], path="/static"),
],
template_config=TemplateConfig(
directory=Path("server/templates"),
engine=JinjaTemplateEngine,
),
)