Source code for src.utils

import logging
from pathlib import Path
from typing import Callable, Iterable, List, Optional

from git.repo import Repo

logger: logging.Logger = logging.getLogger(__name__)


# shamelessly stolen from
# https://stackoverflow.com/questions/1175208/elegant-python-function-to-convert-camelcase-to-snake-case/44969381#44969381


def camel_to_snake(s: str) -> str:
    """
    Converts a string from CamelCase to snake_case. Does not handle many of the
    special cases.
    """
    return "".join(["_" + c.lower() if c.isupper() else c for c in s]).lstrip(
        "_"
    )





def dir_list(
    repo: Repo, dir_predicate: Callable[[str], bool]
) -> Iterable[Path]:
    """
    Searches the project for directories that match the `dir_predicate`. If a
    directory matches its subtree is not explored for more matching directories.

    :param repo: The local checkout of the project.
    :param dir_predicate: Function that receives the directory name and returns
      True iff the spanned subtree should be yielded.
    """
    try:
        logger.debug(f"Assuming fs repo base at {repo.working_dir}")
        trees: Optional[List[Path]] = [Path(repo.working_dir)]
    except ValueError as E:
        logger.error(f"Unable to obtain tree for {repo}: {E}")
        trees: Optional[List[Path]] = None
    while trees:
        tree: Path = trees.pop()
        for f in tree.iterdir():
            if not f.is_dir():
                continue
            if dir_predicate(f.name):
                yield f
            trees.append(f)





def file_list(
    repo: Repo,
    file_name_filter: Callable[[str], bool] = lambda _: False,
    path_component_filter: Callable[[str], bool] = lambda _: False,
    root: Optional[Path] = None,
    recursive: bool = True,
) -> Iterable[Path]:
    """
    Generates a list of all files in a project. Optionally, the yielded
    files can be filtered based on path components or file name. The function
    can also be restricted to a subtree.

    :param repo: The local checkout of the project.
    :param file_name_filter: Function that receives a file name and returns
      True iff the file should be skipped.
    :param path_component_filter: Function that receives the name of a directory
      and returns True iff the whole directory should be skipped.
    :param root: Optionally, search only the subtree rooted at root
    :param recursive: Iterate through folders recursively
    :return: Iterator over all files in the project
    """
    try:
        logger.debug(f"Assuming fs repo base at {repo.working_dir}")
        trees: Optional[List[Path]] = [
            root if root else Path(str(repo.working_dir))
        ]
    except ValueError as E:
        logger.error(f"Unable to obtain tree for {repo}: {E}")
        trees: Optional[List[Path]] = None
    while trees:
        tree: Path = trees.pop()
        for f in tree.iterdir():
            if f.is_symlink():
                continue
            if f.is_file() and file_name_filter(f.name):
                logger.debug(f"Skipping filtered file {f.as_posix()}")
                continue
            if f.is_dir() and recursive:
                if path_component_filter(f.name):
                    logger.debug(f"Skipping filtered subtree {f.as_posix()}")
                else:
                    trees.append(f)
                continue
            if not f.is_file():
                continue
            logger.debug(f"Yielding {f.as_posix()}")
            yield f





def get_publiccode_cfg(repo: Repo) -> Path | None:
    """
    Tries to find the OpenCoDE configuration file in the repository.

    :param repo: The local checkout of the project.
    :return: path of the file or None if it does not exist.
    """
    files_in_repo_root = {f.name: f for f in file_list(repo, recursive=False)}
    for path in ("publiccode.yml", "publiccode.yaml"):
        if path in files_in_repo_root:
            return files_in_repo_root[path]
    return None