Use tesserac (apt install tesseract-ocr) for OCR.

dataLoad.py

@@ -4,6 +4,7 @@ from PIL import Image
 from pprint import pprint
 
 import sys, time, os, requests, random, json, argparse
+import subprocess
 
 parser = argparse.ArgumentParser(description="S0urce.io utility program.")
 parser.add_argument("--download", help="Download Images", action="store_true")
@@ -14,10 +15,6 @@ parser.add_argument(
 args = parser.parse_args()
 # pprint(args)
 
-# Should we spend the time to download image, and process it? (True = Yes, False = No)
-# DOWNLOAD = False
-DOWNLOAD = args.download
-
 # Should we add the JSON in a file? (True is filename, False = do not do)
 # JSONME = 'test.js'
 JSONME = args.JSON
@@ -257,8 +254,40 @@ def run(difficult):
         # print(f"File '{fname}' automatically removed")
 
 
+def autotrain(difficult):
+    """
+    run, represents a single execution of components to the image, (Actuall we do it 1 category at a time instead of just 1 single execution )
+    those components do the following...                           (Each category has around 70 items so we standardize on 70, but           )
+                                                                   (not all of the categories have 70 and thus we print a File does not exist)
+    We open and load the image, and get it's size,
+    then we scan_img for dark and light pixels, <-- This narrows the image down to just the majority of dark pixels
+    then from that we output the image line by line onto the screen after it has been output_image d into list form,
+    Where we ask the user what the word is, and after that we save all that to a file in the data directory.
+    """
+
+    for x in range(0, 70):
+        fname = f"in/{difficult}_{x}.png"
+
+        if not os.path.exists(fname):
+            print("Could not find '{0}'".format(fname))
+            continue
+
+        print(f"Loading: {fname}")
+        fileout = "data/{0}_{1}".format(difficult, x)
+        output = subprocess.run(
+            ["tesseract", fname, fileout],
+            stderr=subprocess.DEVNULL,
+            # capture_output=False,
+            shell=False,
+        )
+
+        with open(fileout + ".txt", "r") as fp:
+            word = fp.read().strip()
+            print(word)
+
+
 # Now to call all the previous functions
-if DOWNLOAD == True:
+if args.download:
     print("Downloading s0urce.io Words")
     print("EASY")
     # time.sleep(5)
@@ -280,9 +309,8 @@ if DOWNLOAD == True:
 
 if args.train:
     # Img Processing
-    run("e")  # Answer the questions
-    run("m")
-    run("h")
+    for level in ["e", "m", "h"]:
+        autotrain(level)
 
 # ----------------------------------------------------------------------------------------
 # All below was in a seperate dataJS.py file... but now I have fixed it so it's 1 script!