Improve dungeon floor selection and start flow

Src/bot_core.py

@@ -94,6 +94,7 @@ def __init__(self, device=None):
             self.unit_series
         ) = self.merge_series = self.df_groups = self.info = self.combat_step = None
         self.logger = logging.getLogger("__main__")
+        logging.getLogger("ocr_utils").setLevel(logging.DEBUG)
         if device is None:
             device = port_scan.get_device()
         if not device:
@@ -702,7 +703,20 @@ def mana_level(self, cards, hero_power=False):
     # Start a dungeon floor from PvE page
     def play_dungeon(self, floor=5):
         self.logger.debug(f"Starting Dungeon floor {floor}")
-        chapter_num = int(np.ceil(floor / 3))
+        # Explicit mapping of floors to chapters
+        floor_map = {
+            1: range(1, 4),
+            2: range(4, 7),
+            3: range(7, 10),
+            4: range(10, 13),
+            5: [13],
+            6: [14],
+        }
+        chapter_num = 1
+        for chap, floors in floor_map.items():
+            if floor in floors:
+                chapter_num = chap
+                break
         self.logger.debug(f"Looking for chapter {chapter_num}")
         pos = np.array([0, 0])
         avail_buttons = self.get_current_icons(available=True)
@@ -711,17 +725,26 @@ def play_dungeon(self, floor=5):
             # Swipe to the top
             [self.swipe([0, 0], [2, 0]) for _ in range(14)]
             self.click(30, 600, 5)  # stop scroll and scan screen for buttons
-            expanded = 0
+            self.getScreen()
+            # Log visible floors for debugging
+            visible = ocr_utils.find_chapter_headers(self.screenRGB)
+            for chap, xy in visible.items():
+                floors_here = ocr_utils.read_floor_from_chapter(self.screenRGB, xy)
+                self.logger.debug(f"Visible chapter {chap} floors: {floors_here}")
+
             for i in range(12):
                 self.getScreen()
                 chapters = ocr_utils.find_chapter_headers(self.screenRGB)
                 self.logger.debug(f"Iteration {i}: OCR chapters {chapters}")
                 if chapter_num in chapters:
                     pos = np.array(chapters[chapter_num])
                     self.logger.info(f"Found chapter {chapter_num} at {pos}")
-                    if not expanded:
-                        expanded = 1
-                        self.click_button(pos + [500, 90])
+                    # Expand chapter only if a collapse icon is detected
+                    icons = self.get_current_icons(available=True)
+                    chapter_icon = f"chapter_{chapter_num}.png"
+                    if (icons["icon"] == chapter_icon).any():
+                        self.logger.debug(f"Chapter {chapter_num} collapsed, expanding")
+                        self.click_button(pos)
                         self.getScreen()
                     if pos[1] < 550 and floor % 3 != 0:
                         break
@@ -760,7 +783,10 @@ def play_dungeon(self, floor=5):
                         chosen_offset = slot_offsets[3]
 
                 self.click_button(pos + chosen_offset)
-                self.click_button((500, 600))
+                # Play selected floor then choose random partner
+                self.click_button((500, 600))  # Play
+                time.sleep(0.5)
+                self.click_button((500, 800))  # Random
                 for i in range(10):
                     time.sleep(2)
                     avail_buttons = self.get_current_icons(available=True)

Src/ocr_utils.py

@@ -7,9 +7,12 @@
 
 from typing import Dict, Optional, Tuple
 
+import logging
 import cv2
 import numpy as np
 
+logger = logging.getLogger(__name__)
+
 try:
     import os
     import shutil
@@ -21,6 +24,7 @@
     _tess_path = os.getenv("TESSERACT_PATH")
     if _tess_path and os.path.exists(_tess_path):
         pytesseract.pytesseract.tesseract_cmd = _tess_path
+        logger.debug("Using tesseract executable at %s", _tess_path)
     else:
         # Common Windows locations
         candidates = [
@@ -31,35 +35,47 @@
         for c in candidates:
             if os.path.exists(c):
                 pytesseract.pytesseract.tesseract_cmd = c
+                logger.debug("Found tesseract executable at %s", c)
                 found = True
                 break
         if not found:
             which = shutil.which("tesseract.exe") or shutil.which("tesseract")
             if which:
                 pytesseract.pytesseract.tesseract_cmd = which
+                logger.debug("Using tesseract executable from PATH: %s", which)
 except Exception:
     _TESS = False
+    logger.warning("Tesseract not available, OCR functions will be disabled")
 
 
 def _prep_digits(img_bgr: np.ndarray) -> np.ndarray:
     """Preprocess ROI to improve OCR on digits: grayscale, CLAHE, blur, threshold, morph."""
+    logger.debug("Preprocessing digits ROI with shape %s", img_bgr.shape)
     gray = cv2.cvtColor(img_bgr, cv2.COLOR_BGR2GRAY)
+    logger.debug("Converted ROI to grayscale")
     # Local contrast boost
     clahe = cv2.createCLAHE(clipLimit=2.0, tileGridSize=(8, 8))
     gray = clahe.apply(gray)
+    logger.debug("Applied CLAHE for local contrast enhancement")
     # De-noise but keep edges
     gray = cv2.bilateralFilter(gray, d=5, sigmaColor=40, sigmaSpace=40)
+    logger.debug("Applied bilateral filter to reduce noise")
     # Adaptive threshold
     th = cv2.adaptiveThreshold(gray, 255, cv2.ADAPTIVE_THRESH_GAUSSIAN_C, cv2.THRESH_BINARY, 31, 5)
+    logger.debug("Applied adaptive thresholding")
     # Morph to unify characters
     kernel = np.ones((2, 2), np.uint8)
     th = cv2.morphologyEx(th, cv2.MORPH_OPEN, kernel, iterations=1)
     th = cv2.dilate(th, kernel, iterations=1)
+    logger.debug("Performed morphological operations")
     # Scale up for better OCR
     h, w = th.shape
     scale = 2 if max(h, w) < 80 else 1
     if scale != 1:
         th = cv2.resize(th, (w * scale, h * scale), interpolation=cv2.INTER_CUBIC)
+        logger.debug("Resized ROI by factor %s", scale)
+    else:
+        logger.debug("No resizing applied (scale=1)")
     return th
 
 
@@ -68,21 +84,26 @@ def ocr_digits(img_bgr: np.ndarray, psm: int = 7) -> Tuple[Optional[int], float]
     psm 7=single line, 6=block; tries strict whitelist for 0-9.
     """
     if not _TESS:
+        logger.debug("Tesseract not available, skipping OCR")
         return None, 0.0
+    logger.debug("Running digit OCR with psm=%s", psm)
     proc = _prep_digits(img_bgr)
+    logger.debug("Prepared ROI for OCR with shape %s", proc.shape)
     cfg = f"--psm {psm} -c tessedit_char_whitelist=0123456789"
     try:
         txt = pytesseract.image_to_string(proc, config=cfg).strip()
+        logger.debug("Tesseract raw output: '%s'", txt)
         # Optional: confidences via image_to_data
         data = pytesseract.image_to_data(proc, config=cfg, output_type=pytesseract.Output.DICT)
         confs = [float(c) for c in data.get("conf", []) if c not in ("-1", None)]
         conf = (sum(confs) / len(confs)) / 100.0 if confs else 0.0
-        # Keep only digits
         digits = "".join(ch for ch in txt if ch.isdigit())
+        logger.debug("Filtered digits='%s', confidence=%.2f", digits, conf)
         if digits == "":
             return None, conf
         return int(digits), conf
-    except Exception:
+    except Exception as e:
+        logger.debug("pytesseract.image_to_string failed: %s", e)
         return None, 0.0
 
 
@@ -107,17 +128,23 @@ def read_floor_from_chapter(
     for slot, (rx, ry, rw, rh) in rois.items():
         rx0, ry0 = max(0, rx), max(0, ry)
         rx1, ry1 = min(w, rx + rw), min(h, ry + rh)
+        logger.debug("Slot %s ROI: x=%s y=%s w=%s h=%s", slot, rx, ry, rw, rh)
         if rx1 <= rx0 or ry1 <= ry0:
+            logger.debug("Slot %s ROI out of bounds", slot)
             results[slot] = (None, 0.0)
             continue
         roi = screen_bgr[ry0:ry1, rx0:rx1]
         # Try psm 7 first, fallback psm 6
         val, conf = ocr_digits(roi, psm=7)
+        logger.debug("Slot %s psm7 result: %s (conf=%.2f)", slot, val, conf)
         if val is None or conf < 0.55:
+            logger.debug("Slot %s falling back to psm6", slot)
             val2, conf2 = ocr_digits(roi, psm=6)
+            logger.debug("Slot %s psm6 result: %s (conf=%.2f)", slot, val2, conf2)
             if (val2 is not None and conf2 >= conf) or val is None:
                 val, conf = val2, conf2
         results[slot] = (val, conf)
+        logger.debug("Slot %s final result: %s (conf=%.2f)", slot, val, conf)
     return results
 
 
@@ -126,22 +153,28 @@ def find_chapter_headers(screen_bgr: np.ndarray) -> Dict[int, Tuple[int, int]]:
     Returns mapping {chapter_number: (x, y)} for the top-left of the word 'Chapter'.
     """
     if not _TESS:
+        logger.debug("Tesseract not available, cannot find chapter headers")
         return {}
+    logger.debug("Running OCR to find chapter headers")
     try:
         data = pytesseract.image_to_data(screen_bgr, output_type=pytesseract.Output.DICT)
     except Exception as e:
         if isinstance(e, (KeyboardInterrupt, SystemExit)):
             raise
+        logger.debug("image_to_data failed: %s", e)
         return {}
     results: Dict[int, Tuple[int, int]] = {}
     words = [w.strip().lower() for w in data.get("text", [])]
+    logger.debug("OCR detected %d words", len(words))
     for i, w in enumerate(words):
         if w == "chapter" and i + 1 < len(words):
             try:
                 num = int(words[i + 1])
                 x = int(data["left"][i])
                 y = int(data["top"][i])
                 results[num] = (x, y)
+                logger.debug("Found chapter %s at (%s, %s)", num, x, y)
             except (ValueError, TypeError):
+                logger.debug("Skipping invalid chapter entry at index %s", i)
                 continue
     return results