Book: Add a sphinx preprocessor

This is to avoid having to use build_book.py script.
1. Preprocessor is faster and doesn't use disk, as all communication is
   trough stdio.
2. It doesn't need to be executed explicitly. You can just use
   `mdbook build docs` and it will create correct documentation.
3. New script is shorter and simpler than old.

Script is based on example from https://rust-lang.github.io/mdBook/for_developers/preprocessors.html
and uses a function from build_book.py

In book.toml the command used is a simple bash script instead of just
`python ./sphinx_preprocessor.py` because the command is run in the
directory the user executed it in, not the directory that book.toml
is in (which is imho a poor design in this case), so in order to be able
to run the command both from main dir and from docs dir, such a script
locating the preprocessor is necessary.

Author: Lorak-mmk

docs/book.toml

@@ -11,4 +11,13 @@ edition = "2021"
 
 [output.html]
 default-theme = "ayu"
-git-repository-url = "https://github.com/scylladb/scylla-rust-driver"
+git-repository-url = "https://github.com/scylladb/scylla-rust-driver"
+
+[preprocessor.sphinx]
+command = '''bash -c '
+if test -f "./docs/sphinx_preprocessor.py"; then
+    python ./docs/sphinx_preprocessor.py "$@" 
+else 
+    python ./sphinx_preprocessor.py "$@"
+fi' run_preprocessor
+'''

docs/sphinx_preprocessor.py

@@ -0,0 +1,34 @@
+import json
+import sys
+
+def remove_sphinx_markdown(md_file_text):
+    text_split = md_file_text.split("```eval_rst")
+    
+    result = text_split[0]
+
+    for i in range(1, len(text_split)):
+        cur_chunk = text_split[i]
+        result += cur_chunk[cur_chunk.find("```") + 3:]
+
+    return result
+
+def process_section(section):
+    if 'Chapter' in section:
+        section['Chapter']['content'] = remove_sphinx_markdown(section['Chapter']['content'])
+        for s in section['Chapter']['sub_items']:
+            process_section(s)
+
+if __name__ == '__main__':
+    if len(sys.argv) > 1: # we check if we received any argument
+        if sys.argv[1] == "supports": 
+            # then we are good to return an exit status code of 0, since the other argument will just be the renderer's name
+            sys.exit(0)
+
+    # load both the context and the book representations from stdin
+    context, book = json.load(sys.stdin)
+
+    for section in book['sections']:
+        process_section(section)
+
+    # we are done with the book's modification, we can just print it to stdout
+    print(json.dumps(book))