Bump signpath/github-action-submit-signing-request from 2.3 to 3.0

Bumps [signpath/github-action-submit-signing-request](https://github.com/signpath/github-action-submit-signing-request) from 2.3 to 3.0.
- [Release notes](https://github.com/signpath/github-action-submit-signing-request/releases)
- [Commits](https://github.com/signpath/github-action-submit-signing-request/compare/v2.3...v3.0)

---
updated-dependencies:
- dependency-name: signpath/github-action-submit-signing-request
  dependency-version: '3.0'
  dependency-type: direct:production
  update-type: version-update:semver-major
...

Signed-off-by: dependabot[bot] <support@github.com>

.github/workflows/package-windows.yml

@@ -58,7 +58,7 @@ jobs:
           name: windows-build-${{ matrix.entry }}
           path: dist/*.exe
       - id: optional_step_id
-        uses: signpath/github-action-submit-signing-request@v2.3
+        uses: signpath/github-action-submit-signing-request@v3.0
         if: ${{ github.repository == 'ciromattia/kcc' }}
         with:
          api-token: '${{ secrets.SIGNPATH_API_TOKEN }}'

.github/workflows/package-windows7.yml

@@ -51,7 +51,7 @@ jobs:
           name: windows7-build
           path: dist/*.exe
       - id: optional_step_id
-        uses: signpath/github-action-submit-signing-request@v2.3
+        uses: signpath/github-action-submit-signing-request@v3.0
         if: ${{ github.repository == 'ciromattia/kcc' }}
         with:
          api-token: '${{ secrets.SIGNPATH_API_TOKEN }}'

README.md

@@ -55,19 +55,23 @@ You can view the split halves either before or after combined spreads:
 
 Just drop your input files into the KCC window, hit convert, and USB drop the output files onto your device's `documents` folder!
 
-https://github.com/user-attachments/assets/da73d625-e082-482d-91a4-ae4765e96fd7
+![settings](images/easy.png)
+
+There are a lot of options, but most people only care about the ones checked above.
+
+Hover over each option to see details in a tooltip.
+
+NOTE: some checkboxes can have a half-check marks.
 
 Then all your usb loaded comic files will show up next to all your normal books in your ereader library!
 
+Here's a quick video demo of how the process works (using expert mode):
+
+https://github.com/user-attachments/assets/da73d625-e082-482d-91a4-ae4765e96fd7
+
 The final result looks incredible on my 10" Kindle Scribe: https://www.youtube.com/watch?v=2CIHW2N9Enc
 
-**WARNING**: Kindle Scribe 2025 MOBI support may have blank pages. Use PDF instead.
-
-The GUI looks like this, built in Qt6. 
-There are a lot of options, but most people only care about the ones checked below.
-Hover over each option to see details in a tooltip. 
-
-![settings](images/settings.png)
+**WARNING**: Kindle Scribe 2025 MOBI support is limited, converted books may have blank pages. Use PDF instead.
 
 If you are on macOS and use a 2022+ Kindle, you may need to use Amazon USB File Manager for Mac or OpenMTP.
 
@@ -77,7 +81,7 @@ Installation tutorial: https://www.youtube.com/watch?v=IR2Fhcm9658
 
 ## Kindle Jailbreak for KOreader
 
-Jailbreaking to use KOreader is optional. KCC works just fine with the native mobi manga viewer.
+Jailbreaking to use KOreader is optional. MOBI manga converted by KCC can be opened in Kindle’s native reader just fine
 
 But if you jailbreak your Kindle, you'll unlock additional features like:
 
@@ -86,6 +90,10 @@ But if you jailbreak your Kindle, you'll unlock additional features like:
 2) no dependency on kindlegen, which can mean **twice as fast conversions** since you can convert to CBZ directly.
 3) You can dither by partially checking the JPG/PNG box and can crop in KCC instead of KOreader. This may increase page turn speed and battery.
 
+Common KOreader manga mistakes: https://youtu.be/gaakIh9p8tw?si=X7OgPJlAvL9XGbLR
+
+Set per folder setting defaults: https://youtu.be/O9p6HVWIQ6s?si=Jtx607erCBA_auWM
+
 ### A word of warning
 **KCC** _is not_ [Amazon's Kindle Comic Creator](http://www.amazon.com/gp/feature.html?ie=UTF8&docId=1001103761) nor is in any way endorsed by Amazon.
 Amazon's tool is for comic publishers and involves a lot of manual effort, while **KCC** is for comic/manga readers.

kindlecomicconverter/page_number_crop_alg.py

@@ -1,4 +1,5 @@
 from PIL import ImageOps, ImageFilter, ImageFile
+from PIL.Image import Image
 import numpy as np
 from .common_crop import threshold_from_power, group_close_values
 
@@ -149,22 +150,39 @@ def get_bbox_crop_margin(img, power=1, background_color='white'):
 
     return bw_img.getbbox()
 
-def ignore_pixels_near_edge(bw_img):
+def ignore_pixels_near_edge(bw_img: Image):
     w, h = bw_img.size
+    # going 0.02-0.03 was too much
+    if int(0.02 * h) == int(0.025 * h):
+        return
+    if int(0.02 * w) == int(0.025 * w):
+        return
     edge_bbox = [
         (0, 0, w, int(0.02 * h)),
         (0, int(0.98 * h), w, h),
         (0, 0, int(0.02 * w), h),
         (int(0.98 * w), 0, w, h)
     ]
-    for box in edge_bbox:
-        edge = bw_img.crop(box)
-        h = edge.histogram()
-        if not edge.height or not edge.width:
-            continue
-        imperfections = h[255] / (edge.height * edge.width)
-        if imperfections > 0 and imperfections < .02:
-            bw_img.paste(im=0, box=box)
+
+    inner_bbox = [
+        (int(0.02 * w), int(0.02 * h), int(0.98 * w), int(0.025 * h)), # top
+        (int(0.02 * w), int(0.975 * h), int(0.98 * w), int(0.98 * h)), # lower
+        (int(0.02 * w), int(0.02 * h), int(0.025 * w), int(0.98 * h)), # left
+        (int(0.975 * w), int(0.02 * h), int(0.98 * w), int(0.98 * h)), # right
+    ]
+
+    for edge_box, inner_box in zip(edge_bbox, inner_bbox):
+        inner_edge = bw_img.crop(inner_box)
+        h = inner_edge.histogram()
+        imperfections = h[255] / (inner_edge.height * inner_edge.width)
+
+        if imperfections > 0 and imperfections < .001:
+            bw_img.paste(im=0, box=inner_box)
+        if imperfections < .001:
+            edge = bw_img.crop(edge_box)
+            edge_h = edge.histogram()
+            if edge_h[-1] != 0:
+                bw_img.paste(im=0, box=edge_box)
 
 
 def box_intersect(box1, box2, max_dist):