Merge pull request #83 from dynobo/dev

Add changelog; Update readme

CHANGELOG.md

@@ -1,5 +1,16 @@
 # Changelog
 
+## v0.1.5 (2020-10-31)
+
+**Changes:**
+
+- Fixes Issue #81: Unhandled exception when user selected no region
+- Related to the one above: Skip OCR if selected area is tiny (below 25 square pixels)
+
+**Development related:**
+- Updates dependencies which also now get pinned
+- Move away from pipenv, using plain setuptools and requirements.txt's instead.
+
 ## v0.1.4 (2020-04-25)
 
 **Changes:**

README.md

@@ -34,9 +34,7 @@
 2. Select a region on the screen
 3. Retrieve recognized text in clipboard
 
-![](https://raw.githubusercontent.com/dynobo/normcap/master/ressource/normcap.gif)
-
-([Screencast on GitHub](https://raw.githubusercontent.com/dynobo/normcap/master/ressource/normcap.gif))
+[![Screencast](https://raw.githubusercontent.com/dynobo/normcap/master/ressource/normcap)](https://raw.githubusercontent.com/dynobo/normcap/master/ressource/normcap)
 
 ## Installation
 
@@ -67,7 +65,7 @@ pip install normcap
 
 (**_OR_** download and extract binary package from the [latest release](https://github.com/dynobo/normcap/releases))
 
-3\. Execute `./normcap`
+3\. Execute `normcap`
 
 ### On Windows _(recommended method)_
 

normcap/handlers/ocr_handler.py

@@ -40,8 +40,10 @@ def handle(self, request: NormcapData) -> NormcapData:
         return request
 
     def img_to_dict(self, img, lang) -> list:
-        if not img.size or img.size[0] * img.size[1] <= 10:
-            self._logger.warning("Image too small for OCR. Dimensions: %s", img.size)
+        if not img.size or img.size[0] * img.size[1] <= 25:
+            self._logger.info(
+                "Selected region of %s too small for OCR. Skipping.", img.size
+            )
             return []
 
         img.format = "PNG"  # WORKAROUND for a pyinstaller bug on Win