def collect_granular(
self,
target: str = None,
with_deps: bool = False,
recursive: bool = False,
graph: "DiGraph" = None,
accept_group: bool = False,
) -> List[StageInfo]:
"""Collects a list of (stage, filter_info) from the given target.
Priority is in the order of following in case of ambiguity:
- .dvc file or .yaml file
- dir if recursive and directory exists
- stage_name
- output file
Args:
target: if not provided, all of the stages without any filters are
returned.
If `target` is a path to a dvc-tracked output,
a (stage, output_path_info) is returned.
Otherwise, the details above for `target` in `collect()`
applies.
(see `collect()` for other arguments)
"""
if not target:
return [StageInfo(stage) for stage in self.repo.stages]
stages, file, _ = _collect_specific_target(self, target, with_deps,
recursive, accept_group)
if not stages:
if not (recursive and self.tree.isdir(target)):
try:
(out, ) = self.repo.find_outs_by_path(target, strict=False)
filter_info = PathInfo(os.path.abspath(target))
return [StageInfo(out.stage, filter_info)]
except OutputNotFoundError:
pass
try:
stages = self.collect(
target,
with_deps,
recursive,
graph,
accept_group=accept_group,
)
except StageFileDoesNotExistError as exc:
# collect() might try to use `target` as a stage name
# and throw error that dvc.yaml does not exist, whereas it
# should say that both stage name and file does not exist.
if file and is_valid_filename(file):
raise
raise NoOutputOrStageError(target, exc.file) from exc
except StageNotFound as exc:
raise NoOutputOrStageError(target, exc.file) from exc
return [StageInfo(stage) for stage in stages]
def collect_granular(self,
target=None,
with_deps=False,
recursive=False,
graph=None):
"""
Priority is in the order of following in case of ambiguity:
- .dvc file or .yaml file
- dir if recursive and directory exists
- stage_name
- output file
"""
if not target:
return [(stage, None) for stage in self.stages]
file, name = parse_target(target)
stages = []
# Optimization: do not collect the graph for a specific target
if not file:
# parsing is ambiguous when it does not have a colon
# or if it's not a dvcfile, as it can be a stage name
# in `dvc.yaml` or, an output in a stage.
logger.debug("Checking if stage '%s' is in '%s'", target,
PIPELINE_FILE)
if not (recursive and os.path.isdir(target)):
stage = self._collect_from_default_dvcfile(target)
if stage:
stages = (self._collect_pipeline(stage)
if with_deps else [stage])
elif not with_deps and is_valid_filename(file):
stages = self.get_stages(file, name)
if not stages:
if not (recursive and os.path.isdir(target)):
try:
(out, ) = self.find_outs_by_path(target, strict=False)
filter_info = PathInfo(os.path.abspath(target))
return [(out.stage, filter_info)]
except OutputNotFoundError:
pass
try:
stages = self.collect(target, with_deps, recursive, graph)
except StageFileDoesNotExistError as exc:
# collect() might try to use `target` as a stage name
# and throw error that dvc.yaml does not exist, whereas it
# should say that both stage name and file does not exist.
if file and is_valid_filename(file):
raise
raise NoOutputOrStageError(target, exc.file) from exc
except StageNotFound as exc:
raise NoOutputOrStageError(target, exc.file) from exc
return [(stage, None) for stage in stages]