add re-validation in python

opus_submitter/submissions/api.py

@@ -3,7 +3,6 @@ from ninja.files import UploadedFile
 from ninja.pagination import paginate
 from django.db import transaction
 from django.core.files.base import ContentFile
-from django.http import Http404
 from django.utils import timezone
 from django.shortcuts import get_object_or_404
 from typing import List
@@ -97,12 +96,9 @@ def create_submission(
             file_index = 0
             for response_data in data.responses:
                 # Get the puzzle
-                try:
-                    puzzle = SteamCollectionItem.objects.get(id=response_data.puzzle_id)
-                except SteamCollectionItem.DoesNotExist:
-                    return 400, {
-                        "detail": f"Puzzle with id {response_data.puzzle_id} not found"
-                    }
+                puzzle = get_object_or_404(
+                    SteamCollectionItem, id=response_data.puzzle_id
+                )
 
                 # Create the puzzle response
                 response = PuzzleResponse.objects.create(
@@ -121,6 +117,7 @@ def create_submission(
                 # Process files for this response
                 # For simplicity, we'll take one file per response
                 # In a real implementation, you'd need better file-to-response mapping
+                print("FI", file_index, files)
                 if file_index < len(files):
                     uploaded_file = files[file_index]
 

opus_submitter/submissions/management/commands/recheck_ocr_validation.py

@@ -0,0 +1,20 @@
+from django.core.management.base import BaseCommand
+from submissions.utils import verify_and_validate_ocr_date_for_submission
+from submissions.models import SubmissionFile
+
+
+class Command(BaseCommand):
+    def add_arguments(self, parser):
+        parser.add_argument("--force", action="store_true", help="Force redo the OCR")
+
+    def handle(self, *args, **options):
+        for file in SubmissionFile.objects.prefetch_related(
+            "response__puzzle",
+            "response__submission__user",
+        ):
+            if not file.response.needs_manual_validation and not options.get(
+                "force", False
+            ):
+                continue
+
+            verify_and_validate_ocr_date_for_submission(file)

opus_submitter/submissions/utils.py

@@ -4,12 +4,16 @@ Utilities for fetching Steam Workshop collection data using Steam Web API
 
 import re
 import requests
-from submissions.models import SteamCollection
+from submissions.models import SteamCollection, SubmissionFile
 from datetime import datetime
 from django.utils import timezone
 from django.conf import settings
 from typing import Dict, List, Optional, Tuple
 import logging
+from PIL import Image
+import cv2
+import pytesseract
+import os
 
 logger = logging.getLogger(__name__)
 
@@ -425,3 +429,88 @@ def create_or_update_collection(url: str) -> Tuple[SteamCollection, bool]:
         )
 
     return collection, created
+
+
+def verify_ocr_data_for_file(file: str) -> tuple[str, int, int, int]:
+    # Convert GIF to JPG
+    with Image.open(file) as img:
+        # width, height = img.size
+        img.seek(0)
+        rgb_img = img.convert("RGB")
+        rgb_img.save("temp.jpg", "JPEG")
+
+    # Read image from which text needs to be extracted
+    img = cv2.imread("temp.jpg")
+
+    gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
+
+    # Optional: resize for better OCR
+    gray = cv2.resize(gray, None, fx=1, fy=1, interpolation=cv2.INTER_CUBIC)
+
+    # Manually crop regions based on known layout (x, y, w, h)
+    regions = [
+        (15, 600, 330, 28),  # PUZZLE NAME
+        (412, 603, 65, 22),  # COST
+        (577, 603, 65, 22),  # CYCLES
+        (739, 603, 65, 22),  # AREA
+    ]
+
+    output = img.copy()
+
+    def find_text(dims, gray, output, content):
+        x, y, w, h = dims
+        roi = gray[y : y + h, x : x + w]
+        roi = cv2.bitwise_not(roi)
+        if content == "digits" or content == "digits_with_6":
+            config = "--oem 3 --psm 7 -c tessedit_char_whitelist=0123456789"
+
+        else:
+            config = "--oem 3 --psm 7"
+
+        text = pytesseract.image_to_string(roi, config=config).strip()
+
+        # Remove the extra 6 (actually the G for Gold) for cost value
+        if content == "digits_with_6":
+            text = text[:-1]
+
+        cv2.rectangle(output, (x, y), (x + w, y + h), (0, 255, 0), 2)
+        return text
+
+    puzzle = find_text(regions[0], gray, output, "letters")
+    cost = find_text(regions[1], gray, output, "digits_with_6")
+    cycles = find_text(regions[2], gray, output, "digits")
+    area = find_text(regions[3], gray, output, "digits")
+
+    # Save image with green rectangles around the considered zones, for debug purposes
+    # cv2.imwrite("output_debug.jpg", output)
+
+    os.remove("temp.jpg")
+    return puzzle, cost, cycles, area
+
+
+def verify_and_validate_ocr_date_for_submission(file: SubmissionFile):
+    ocr_data = verify_ocr_data_for_file(file.file.path)
+    r = file.response
+
+    print(
+        f"{r.submission.user}: ({r.cost: >4} {r.cycles: >4} {r.area: >4}) -> ({ocr_data[1]: >4} {ocr_data[2]: >4} {ocr_data[3]: >4})"
+    )
+
+    valid_count = 0
+    if r.cost == ocr_data[1]:
+        r.validated_cost = r.cost
+        valid_count += 1
+
+    if r.cycles == ocr_data[2]:
+        r.validated_cycles = r.cycles
+        valid_count += 1
+
+    if r.area == ocr_data[3]:
+        r.validated_area = r.area
+        valid_count += 1
+
+    if valid_count == 3:
+        r.needs_manual_validation = False
+
+    if valid_count:
+        r.save()