def load(self, context, cache):
if cache is not None and path.isdir(cache.get("directory", "")):
# Check if the requested version changes.
url_template = context.expand_variables(cache.get("url_template", ""))
if url_template == cache.get("url"):
self.directory = cache["directory"]
logger.info("Reusing cached directory: {}".format(path.rel(self.directory, nopar=True)))
return cache
else:
logger.info("Cached URL is outdated:", cache.get("url"))
directory = None
archive = None
delete_after_extract = True
for url_template in self.urls:
url = context.expand_variables(url_template)
if not url:
continue
if url.startswith("file://"):
name = url[7:]
if path.isdir(name):
logger.info("Using directory", url)
directory = name
break
elif path.isfile(name):
logger.info("Using archive", url)
archive = name
delete_after_extract = False
break
error = None
else:
error = None
try:
progress = lambda d: self._download_progress(url, context, d)
archive, reused = httputils.download_file(
url, directory=context.get_temporary_directory(), on_exists="skip", progress=progress
)
except (httputils.URLError, httputils.HTTPError) as exc:
error = exc
except self.DownloadAlreadyExists as exc:
directory = exc.directory
logger.info("Reusing existing directory", directory)
else:
if reused:
logger.info("Reusing cached download", path.basename(archive))
break
if error:
logger.info("Error reading", url, ":", error)
if directory or archive:
logger.debug("URL applies: {}".format(url))
if not directory and archive:
suffix, directory = self._get_archive_unpack_info(context, archive)
logger.info(
'Unpacking "{}" to "{}" ...'.format(path.rel(archive, nopar=True), path.rel(directory, nopar=True))
)
nr.misc.archive.extract(
archive,
directory,
suffix=suffix,
unpack_single_dir=True,
check_extract_file=self._check_extract_file,
progress_callback=self._extract_progress,
)
elif not directory:
raise LoaderError(self, "no URL matched")
self.directory = directory
with open(path.join(self.directory, ".craftr_downloadurl"), "w") as fp:
fp.write(url)
return {"directory": directory, "url_template": url_template, "url": url}
def external_file(*urls, filename = None, directory = None,
copy_file_url = False, name = None):
"""
Downloads a file from the first valid URL and saves it into *directory*
under the specified *filename*.
:param urls: One or more URLs. Supports ``http://``, ``https://``,
``ftp://`` and ``file://`. Note that if a ``file://`` URL is
specified, the file is not copied to the output filename unless
*copy_file_url* is True.
:param filename: The output filename of the downloaded file. Defaults
to the filename of the downloaded file.
:param directory: The directory to save the file to. If *filename*
is a relative path, it will be joined with this directory. Defaults
to a path in the build directory.
:param copy_file_url: If True, ``file://`` URLs will be copied instead
of used as-is.
:param name: The name of the loader action. This name is used to store
information in the :attr:`Session.cache` so we can re-use existing
downloaded data. :func:`~craftr.defaults.gtn` will be used to
retrieve the default value for this parameter.
:return: The path to the downloaded file.
"""
name = gtn(name)
if not directory and not filename:
directory = buildlocal('data')
cache = get_loader_cache(name)
# TODO: expand variables of the current module.
target_filename = None
exceptions = []
for url in urls:
if url == cache.get('download_url'):
existing_file = cache.get('download_file')
if existing_file and path.isfile(existing_file):
return existing_file
progress_info = 'Downloading {} ...'.format(url)
if url.startswith('file://'):
source_file = url[7:]
if path.isfile(source_file):
if not copy_file_url:
return source_file
if not filename:
filename = path.basename(source_file)
# TODO: Use httputils.download_file() for this as well?
logger.progress_begin(progress_info)
path.makedirs(directory)
target_filename = path.join(directory, filename)
with open(source_file, 'rb') as sfp:
with open(target_filename, 'wb') as dfp:
for bytes_copied, size in pyutils.copyfileobj(sfp, dfp):
logger.progress_update(float(bytes_copied) / size)
logger.progress_end()
# TODO: Copy file permissions
break
else:
exceptions.append(FileNotFoundError(url))
else:
progress = lambda data: _external_file_download_callback(
progress_info, directory, filename, cache, data)
try:
target_filename, reused = httputils.download_file(
url, filename = filename, directory = directory, on_exists = 'skip',
progress = progress)
except (httputils.URLError, httputils.HTTPError) as exc:
exceptions.append(exc)
else:
break
finally:
logger.progress_end()
if target_filename:
cache['download_url'] = url
cache['download_file'] = target_filename
return target_filename
raise NoExternalFileMatch(name, urls, exceptions)
