Update test_tesseract.py

Funtionality cannot be tested in GitHub actions because no Tesseract is installed.

Correctly locate tessdata from installation folder

If tesseract-ocr is not version 4.00 in a Unix-like platform, we incorrectly used pathlib.Path for locating tessdata.
This fix correctly determines the tessdata folder name for any version of tesseract-ocr.
Note however, that we cannot guarantee that MuPDF's OCR code (which is on version 4.00) can cope with tessdata content contained in a different version than 4.00.

src/__init__.py

@@ -18041,7 +18041,7 @@ def make_utf16be(s):
     return "(" + r + ")"
 
 
-def get_tessdata() -> str:
+def get_tessdata():
     """Detect Tesseract-OCR and return its language support folder.
 
     This function can be used to enable OCR via Tesseract even if the
@@ -18053,17 +18053,9 @@ def get_tessdata() -> str:
         Folder name of tessdata if Tesseract-OCR is available, otherwise False.
     """
     TESSDATA_PREFIX = os.getenv("TESSDATA_PREFIX")
-    if TESSDATA_PREFIX is not None:
+    if TESSDATA_PREFIX:  # use environment variable if set
         return TESSDATA_PREFIX
 
-    if sys.platform == "win32":
-        tessdata = "C:\\Program Files\\Tesseract-OCR\\tessdata"
-    else:
-        tessdata = "/usr/share/tesseract-ocr/4.00/tessdata"
-
-    if os.path.exists(tessdata):
-        return tessdata
-
     """
     Try to locate the tesseract-ocr installation.
     """
@@ -18090,18 +18082,17 @@ def get_tessdata() -> str:
         message("Tesseract-OCR is not installed")
         return False
 
-    # determine tessdata via iteration over subfolders
-    tessdata = None
-    for sub_response in response.iterdir():
-        for sub_sub in sub_response.iterdir():
-            if str(sub_sub).endswith("tessdata"):
-                tessdata = sub_sub
-                break
-    if tessdata is not None:
+    # search tessdata in folder structure
+    dirname = response[1]  # contains tesseract-ocr installation folder
+    if isinstance(dirname, bytes):  # has probably 'bytes' format
+        dirname = dirname.decode()
+    dirname = pathlib.Path(dirname)  # make a Path from it
+    sub = list(dirname.iterdir())[0]  # includes the version sub-folder
+    tessdata = str(list(sub.iterdir())[0])  # this should be the tessdata
+    if tessdata.endswith("tessdata"):  # expected name suffix
         return tessdata
     else:
         message("unexpected: tesseract-ocr has no 'tessdata' folder")
-        return False
     return False
 
 

tests/test_tesseract.py

@@ -65,3 +65,4 @@ def test_tesseract():
                         'dropping unclosed output'
                         )
 
+