viktor/wrongmove
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

detect floorplan using asyncio

Browse source
This commit is contained in:
Viktor Barzin 2025-05-17 22:58:35 +00:00
parent 68cc70bd11
commit b1e0ed170b
No known key found for this signature in database
GPG key ID: 4056458DBDBF8863
3 changed files with 22 additions and 9 deletions
Download patch file Download diff file Expand all files Collapse all files

22
crawler/4_detect_floorplan.py
View file

@ -1,15 +1,25 @@
import asyncio
import pathlib
from data_access import Listing
from tqdm import tqdm
from tqdm.asyncio import tqdm
import multiprocessing
def detect_floorplan(listing_paths: list[str]):
async def detect_floorplan(listing_paths: list[str]):
listings = Listing.get_all_listings(listing_paths)
cpu_count = multiprocessing.cpu_count() / 4
semaphore = asyncio.Semaphore(cpu_count)
for listing in tqdm(listings):
tqdm.write(str(listing.identifier))
# listing.calculate_sqm_model() # using google/deplot model. Too slow, rather use tesseract
listing.calculate_sqm_ocr(recalculate=False)
await tqdm.gather(
*[_detect_floorplan_with_semaphore(listing, semaphore) for listing in listings]
)
async def _detect_floorplan_with_semaphore(
listing: Listing, semaphore: asyncio.Semaphore
):
async with semaphore:
return await listing.calculate_sqm_ocr(recalculate=False)
def main():