Ejemplo n.º 1
0
def install_tools() -> NoReturn:
    """
    Install the build tools. Doesn't work inside a Docker container.
    """
    info('"course install-tools" is deprecated, because it is not possible '
         'to update a Docker container from within itself. To update the '
         'build tools, run the following command:\n\n'
         'curl -L https://git.io/fhaLg | bash')
Ejemplo n.º 2
0
    def import_dbc(dbc: str, build: bdc.BuildData) -> NoReturn:
        '''
        Import a single DBC.

        Assumes (a) the working directory is the build directory, and
        (b) that the remote target path has already been created.
        '''
        w = databricks.Workspace(profile=db_profile)
        if build.has_profiles:
            parent_subpath = os.path.dirname(dbc)
            dir_to_make = f'{remote_target}/{os.path.dirname(parent_subpath)}'
            w.mkdirs(dir_to_make)
            remote_path = f'{remote_target}/{parent_subpath}'
        else:
            remote_path = remote_target

        info(f'Importing "{dbc}" to "{remote_path}"...')
        w.import_dbc(dbc, remote_path)
Ejemplo n.º 3
0
def import_dbcs(cfg: Dict[str, str], build_dir: str,
                build_file: str) -> NoReturn:
    """
    Find all DBC files under the build output directory for the current course,
    and upload them (import them) into the Databricks instance.

    :param cfg:       The config. COURSE_NAME, COURSE_REMOTE_TARGET, and
                      DB_PROFILE are assumed to be set.
    :param build_dir: The path to the build directory.

    :return: NOthing
    """
    check_config(cfg)
    remote_target = cfg['COURSE_REMOTE_TARGET']
    db_profile = cfg['DB_PROFILE']

    def import_dbc(dbc: str, build: bdc.BuildData) -> NoReturn:
        '''
        Import a single DBC.

        Assumes (a) the working directory is the build directory, and
        (b) that the remote target path has already been created.
        '''
        w = databricks.Workspace(profile=db_profile)
        if build.has_profiles:
            parent_subpath = os.path.dirname(dbc)
            dir_to_make = f'{remote_target}/{os.path.dirname(parent_subpath)}'
            w.mkdirs(dir_to_make)
            remote_path = f'{remote_target}/{parent_subpath}'
        else:
            remote_path = remote_target

        info(f'Importing "{dbc}" to "{remote_path}"...')
        w.import_dbc(dbc, remote_path)

    # Get the build information. We'll need it later.
    build = bdc.bdc_load_build(build_file)

    print(
        f'Importing all DBCs under "{build_dir}" to remote "{remote_target}"')
    dbcs = []
    with working_directory(build_dir) as pwd:
        for dirpath, _, filenames in os.walk('.'):
            for filename in filenames:
                _, ext = os.path.splitext(filename)
                if ext != '.dbc':
                    continue
                dbcs.append(os.path.normpath(os.path.join(dirpath, filename)))

        if not dbcs:
            warn('No DBCs found.')
        else:
            clean(cfg)
            w = databricks.Workspace(profile=db_profile)
            # If we're doing a profile-based build, create the remote target.
            # The import operations will implicitly create the remote
            # subfolders. However, if we're not doing profile-based builds,
            # then creating the remote target ahead of time will cause the
            # import to fail, so don't do that.
            if build.has_profiles:
                w.mkdirs(remote_target)

            for dbc in dbcs:
                info(f'\nIn "{pwd}":')
                import_dbc(dbc, build)