app/services/metadata_service.py

from pathlib import Path

from git import Repo

from app.models.metadata import (
    RepositoryMetadata,
    RepositoryInfo,
    StructureInfo,
    ConfigurationInfo,
)


class MetadataService:

    def extract(self, repo_path: Path):

        repo = Repo(repo_path)

        branch = repo.active_branch.name

        latest_commit = repo.head.commit.hexsha

        total_files = 0
        python_files = 0
        directories = 0

        size = 0

        has_readme = False
        has_license = False
        has_dockerfile = False
        has_docker_compose = False

        requirements = []

        for path in repo_path.rglob("*"):

            if path.is_dir():
                directories += 1
                continue

            total_files += 1

            size += path.stat().st_size

            if path.suffix == ".py":
                python_files += 1

            filename = path.name.lower()

            if filename.startswith("readme"):
                has_readme = True

            elif filename.startswith("license"):
                has_license = True

            elif filename == "dockerfile":
                has_dockerfile = True

            elif filename == "docker-compose.yml":
                has_docker_compose = True

            elif filename == "requirements.txt":

                with open(path) as f:

                    for line in f:

                        line = line.strip()

                        if line and not line.startswith("#"):

                            requirements.append(line)

        return RepositoryMetadata(

            repository=RepositoryInfo(

                name=repo_path.name,

                path=str(repo_path),

                branch=branch,

                latest_commit=latest_commit,

                size_mb=round(size / 1024 / 1024, 2),

            ),

            structure=StructureInfo(

                total_files=total_files,

                python_files=python_files,

                directories=directories,

            ),

            configuration=ConfigurationInfo(

                has_readme=has_readme,

                has_license=has_license,

                has_dockerfile=has_dockerfile,

                has_docker_compose=has_docker_compose,

                requirements=requirements,

            ),
        )