py : type-check all Python scripts with Pyright (#8341)

* py : type-check all Python scripts with Pyright

* server-tests : use trailing slash in openai base_url

* server-tests : add more type annotations

* server-tests : strip "chat" from base_url in oai_chat_completions

* server-tests : model metadata is a dict

* ci : disable pip cache in type-check workflow

The cache is not shared between branches, and it's 250MB in size,
so it would become quite a big part of the 10GB cache limit of the repo.

* py : fix new type errors from master branch

* tests : fix test-tokenizer-random.py

Apparently, gcc applies optimisations even when pre-processing,
which confuses pycparser.

* ci : only show warnings and errors in python type-check

The "information" level otherwise has entries
from 'examples/pydantic_models_to_grammar.py',
which could be confusing for someone trying to figure out what failed,
considering that these messages can safely be ignored
even though they look like errors.

scripts/compare-llama-bench.py

@@ -123,13 +123,13 @@ builds = cursor.execute("SELECT DISTINCT build_commit FROM test;").fetchall()
 
 try:
     repo = git.Repo(".", search_parent_directories=True)
-except git.exc.InvalidGitRepositoryError:
+except git.InvalidGitRepositoryError:
     repo = None
 
 
-def find_parent_in_data(commit):
+def find_parent_in_data(commit: git.Commit):
     """Helper function to find the most recent parent measured in number of commits for which there is data."""
-    heap = [(0, commit)]
+    heap: list[tuple[int, git.Commit]] = [(0, commit)]
     seen_hexsha8 = set()
     while heap:
         depth, current_commit = heapq.heappop(heap)
@@ -144,7 +144,7 @@ def find_parent_in_data(commit):
     return None
 
 
-def get_all_parent_hexsha8s(commit):
+def get_all_parent_hexsha8s(commit: git.Commit):
     """Helper function to recursively get hexsha8 values for all parents of a commit."""
     unvisited = [commit]
     visited   = []