fix: improve word generation cyclic file reading (#54)

This pull request includes several improvements to the word_generator function in the bing_rewards/__init__.py file. The changes enhance error handling, ensure uniform random distribution, and improve code readability.

Enhancements to word_generator function:
- Type hints: specified the return type as Generator[str, None, None].
- Added yields and potential exceptions in the docstring.
- Implemented error handling for FileNotFoundError and IOError, with appropriate error messages and re-raising of exceptions.
- Ensured the function handles empty keyword files by raising a ValueError if the file size is zero. (Useful when implementing a feature to add your own .txt files for searches)
- Improved the random starting position logic (to capture the whole txt file) and ensured clean line boundaries by discarding partial lines and skipping empty lines.

bing_rewards/__init__.py

@@ -44,25 +44,58 @@
 import bing_rewards.options as app_options
 
 
-def word_generator() -> Generator[str]:
+def word_generator() -> Generator[str, None, None]:
     """Infinitely generate terms from the word file.
 
     Starts reading from a random position in the file.
     If end of file is reached, close and restart.
+    Handles file operations safely and ensures uniform random distribution.
+
+    Yields:
+        str: A random keyword from the file, stripped of whitespace.
+
+    Raises:
+        OSError: If there are issues accessing or reading the file.
     """
     word_data = resources.files('bing_rewards').joinpath('data', 'keywords.txt')
-    while True:
-        with (
-            resources.as_file(word_data) as p,
-            p.open(mode='r', encoding='utf-8') as fh,
-        ):
-            fh.seek(0, SEEK_END)
-            size = fh.tell()  # Get the filesize of the Keywords file
-            # Start at a random position in the stream
-            fh.seek(random.randint(0, (size * 3 // 4)), SEEK_SET)
-            for line in fh:
-                # Use the built in file handler generator
-                yield line.strip()
+
+    try:
+        while True:
+            with (
+                resources.as_file(word_data) as p,
+                p.open(mode='r', encoding='utf-8') as fh,
+            ):
+                # Get the file size of the Keywords file
+                fh.seek(0, SEEK_END)
+                size = fh.tell()
+
+                if size == 0:
+                    raise ValueError('Keywords file is empty')
+
+                # Start at a random position in the stream
+                fh.seek(random.randint(0, size - 1), SEEK_SET)
+
+                # Read and discard partial line to ensure we start at a clean line boundary
+                fh.readline()
+
+                # Read lines until EOF
+                for raw_line in fh:
+                    stripped_line = raw_line.strip()
+                    if stripped_line:  # Skip empty lines
+                        yield stripped_line
+
+                # If we hit EOF, seek back to start and continue until we've yielded enough words
+                fh.seek(0)
+                for raw_line in fh:
+                    stripped_line = raw_line.strip()
+                    if stripped_line:
+                        yield stripped_line
+    except OSError as e:
+        print(f'Error accessing keywords file: {e}')
+        raise
+    except Exception as e:
+        print(f'Unexpected error in word generation: {e}')
+        raise
 
 
 def browser_cmd(exe: Path, agent: str, profile: str = '') -> list[str]: