fix: explain xpdf install failure

Log list of GPUs for tensorflow.

install_xpdf.py

@@ -117,7 +117,7 @@ def run(self) -> bool:
             raise ValueError(msg)
 
         logger.info("Finished.")
-        return False
+        return True
 
     def _get_url_file_name(self, url: str) -> str:
         parsed_url = urlparse(url)
@@ -163,6 +163,7 @@ def _gpg_verify(
             result_import = subprocess.run(cmd_import, capture_output=True)
             expected_import_return_code = 2
             if result_import.returncode != expected_import_return_code:
+                logger.warning("Unexpected return code '%s'.", result_import)
                 # expecting return code 2 because there is no ultimately trusted key
                 return False
             contains_email = [
@@ -171,6 +172,7 @@ def _gpg_verify(
                 if expected_email in line.decode()
             ]
             if len(contains_email) != 1:
+                logger.warning("Unexpected stderr '%s'.", result_import)
                 # gpg key must contain the expected email address
                 return False
 
@@ -185,6 +187,7 @@ def _gpg_verify(
             ]
             result_verify = subprocess.run(cmd_verify, capture_output=True)
             if result_verify.returncode != 0:
+                logger.warning("Unexpected return code '%s'.", result_verify)
                 # expecting status code 0
                 return False
             contains_good_email = [
@@ -193,6 +196,7 @@ def _gpg_verify(
                 if f"Good signature from {expected_email}" in line.decode()
             ]
             if len(contains_good_email) != 1:
+                logger.warning("Unexpected stderr code '%s'.", result_verify)
                 # must be a 'good' signature with the expected key
                 return False
 

src/leaf_focus/ocr/keras_ocr.py

@@ -48,6 +48,10 @@ def engine_create(self) -> None:
 
         tf.get_logger().setLevel(log_level)
 
+        # check the CPU / GPU in use
+        gpus = tf.config.list_physical_devices("GPU")
+        logger.info("GPUs in use: '%s'.", gpus)
+
         try:
             import keras_ocr  # pylint: disable=import-outside-toplevel
         except ModuleNotFoundError as error: