refactor detect floorplan to use model listings

crawler/4_detect_floorplan.py

@@ -1,21 +1,43 @@
 import asyncio
-from data_access import Listing
+from models import Listing
+from rec import floorplan
+from repositories.listing_repository import ListingRepository
 from tqdm.asyncio import tqdm
 import multiprocessing
 
 
-async def detect_floorplan(listing_paths: list[str]):
-    listings = Listing.get_all_listings(listing_paths)
+async def detect_floorplan(repository: ListingRepository):
+    listings = await repository.get_listings()
     cpu_count = multiprocessing.cpu_count() // 4
     semaphore = asyncio.Semaphore(cpu_count)
 
-    await tqdm.gather(
-        *[_detect_floorplan_with_semaphore(listing, semaphore) for listing in listings]
-    )
+    updated_listings = [
+        listing
+        for listing in await tqdm.gather(
+            *[_calculate_sqm_ocr(listing, semaphore) for listing in listings]
+        )
+        if listing is not None
+    ]
+    await repository.upsert_listings(updated_listings)
 
 
-async def _detect_floorplan_with_semaphore(
+async def _calculate_sqm_ocr(
     listing: Listing, semaphore: asyncio.Semaphore
-):
-    async with semaphore:
-        return await listing.calculate_sqm_ocr(recalculate=False)
+) -> Listing | None:
+    if listing.square_meters is not None:
+        return None
+    sqms = []
+    for floorplan_path in listing.floorplan_image_paths:
+        async with semaphore:
+            estimated_sqm, _ = await asyncio.to_thread(
+                floorplan.calculate_ocr, floorplan_path
+            )
+            if estimated_sqm is not None:
+                sqms.append(estimated_sqm)
+    max_sqm = max(sqms, default=None)
+    if max_sqm is not None:
+        listing.square_meters = max_sqm
+        return listing
+    else:
+        listing.square_meters = None
+        return None

crawler/main.py

@@ -165,7 +165,7 @@ def dump_listings(
 @click.pass_context
 def dump_images(ctx: click.core.Context):
     data_dir = ctx.obj["data_dir"]
-    click.echo(f"Running dump_images stored in {data_dir}")
+    click.echo(f"Running dump_images for listings stored in {engine.url}")
     repository = ListingRepository(engine=engine)
     asyncio.run(dump_images_module.dump_images(repository, image_base_path=data_dir))
 
@@ -174,9 +174,9 @@ def dump_images(ctx: click.core.Context):
 @click.pass_context
 def detect_floorplan(ctx: click.core.Context):
     data_dir = ctx.obj["data_dir"]
-    click.echo(f"Running detect_floorplan in {data_dir}")
-    listing_paths = sorted(list(pathlib.Path(data_dir).glob("*/listing.json")))
-    asyncio.run(detect_floorplan_module.detect_floorplan(listing_paths))
+    click.echo(f"Running detect_floorplan for listings stored in {engine.url}")
+    repository = ListingRepository(engine=engine)
+    asyncio.run(detect_floorplan_module.detect_floorplan(repository))
 
 
 @cli.command()

crawler/rec/floorplan.py

@@ -44,7 +44,7 @@ def improve_img_for_ocr(img: Image):
     return Image.fromarray(thresh)
 
 
-def calculate_ocr(image_path):
+def calculate_ocr(image_path) -> tuple[float | None, str]:
     import pytesseract
 
     img = Image.open(image_path)