def call_as(self, items, meta):
return D.Call(
sp(self.filename, meta),
items[0],
items[1].value,
items[2] if len(items) > 2 else dict(),
)
def document(self, items, meta):
imports = []
structs = {}
tasks = []
workflow = None
for item in items:
if isinstance(item, D.Task):
tasks.append(item)
elif isinstance(item, D.Workflow):
if workflow is not None:
raise Err.MultipleDefinitions(
sp(self.filename, meta),
"Document has multiple workflows")
workflow = item
elif isinstance(item, D.StructTypeDef):
if item.name in structs:
raise Err.MultipleDefinitions(
sp(self.filename, meta),
"multiple structs named " + item.name)
structs[item.name] = item
elif isinstance(item, lark.Tree) and item.data == "version":
pass
elif isinstance(item, D.DocImport):
imports.append(item)
else:
assert False
return D.Document(sp(self.filename, meta), imports, structs, tasks,
workflow)
def task(self, items, meta):
d = {}
for item in items:
if isinstance(item, dict):
for k, v in item.items():
if k in d:
raise Err.MultipleDefinitions(
sp(self.filename, meta), "redundant sections in task"
)
d[k] = v
else:
assert isinstance(item, str)
assert "name" not in d
d["name"] = item.value
return D.Task(
sp(self.filename, meta),
d["name"],
d.get("inputs", []),
d.get("decls", []),
d["command"],
d.get("outputs", []),
d.get("parameter_meta", {}),
d.get("runtime", {}),
d.get("meta", {}),
)
def decl(self, items, meta):
return D.Decl(
sp(self.filename, meta),
items[0],
items[1].value,
(items[2] if len(items) > 2 else None),
)
def parse_wdl(wdl_path: Path,
import_dirs: Optional[Sequence[Path]] = (),
check_quant: bool = False,
**_) -> Document:
return Tree.load(str(wdl_path),
path=[str(path) for path in import_dirs],
check_quant=check_quant)
def parse_document(
txt: str, version: Optional[str] = None, uri: str = "", imported: bool = False
) -> D.Document:
if version is None:
# for now assume the version is 1.0 if the first line is "version <number>"
# otherwise draft-2
version = "draft-2"
for line in txt.split("\n"):
line = line.strip()
if line and line[0] != "#":
if line.startswith("version ") and line[8].isdigit():
version = "1.0"
break
if not txt.strip():
return D.Document(
SourcePosition(filename=uri, line=0, column=0, end_line=0, end_column=0),
[],
[],
None,
imported,
)
try:
return _DocTransformer(uri, imported).transform(parse(txt, "document", version))
except lark.exceptions.UnexpectedCharacters as exn:
raise Err.ParserError(uri if uri != "" else "(in buffer)") from exn
except lark.exceptions.UnexpectedToken as exn:
raise Err.ParserError(uri if uri != "" else "(in buffer)") from exn
def task(self, items, meta):
d = {"noninput_decls": []}
for item in items:
if isinstance(item, dict):
for k, v in item.items():
if k == "noninput_decl":
d["noninput_decls"].append(v)
elif k in d:
raise Err.MultipleDefinitions(
sp(self.filename, meta),
"redundant sections in task")
else:
d[k] = v
else:
assert isinstance(item, str)
assert "name" not in d
d["name"] = item.value
_check_keyword(sp(self.filename, meta), d["name"])
return D.Task(
sp(self.filename, meta),
d["name"],
d.get("inputs", None),
d["noninput_decls"],
d["command"],
d.get("outputs", []),
d.get("parameter_meta", {}),
d.get("runtime", {}),
d.get("meta", {}),
)
def load(uri: str, path: List[str] = [], check_quant: bool = True) -> Document:
"""
Parse a WDL document given filename/URI, recursively descend into imported documents, then typecheck the tasks and workflow.
:param path: local filesystem directories to search for imports, in addition to the current working directory
:param check_quant: set to ``False`` to relax static typechecking of the optional (?) and nonempty (+) type quantifiers. This is discouraged, but may be useful for older WDL workflows which assume less-rigorous static validation of these annotations.
"""
return Tree.load(uri, path, check_quant)
def _get_workflow_name(self, wdl_path: Path, kwargs: dict):
if "workflow_name" in kwargs:
return kwargs["workflow_name"]
elif Tree:
if "check_quant" not in kwargs:
kwargs["check_quant"] = False
doc = Tree.load(str(wdl_path),
path=[str(path) for path in self.import_dirs],
**kwargs)
return doc.workflow.name
else: # TODO: test this
return safe_string(wdl_path.stem)
def struct(self, items, meta):
assert len(items) >= 1
name = items[0]
_check_keyword(sp(self.filename, meta), name)
members = {}
for d in items[1:]:
assert not d.expr
if d.name in members:
raise Err.MultipleDefinitions(sp(self.filename, meta),
"duplicate members in struct")
members[d.name] = d.type
return D.StructTypeDef(sp(self.filename, meta), name, members)
def workflow(self, items, meta):
elements = []
inputs = None
outputs = None
output_idents = None
output_idents_pos = None
parameter_meta = None
meta_section = None
for item in items[1:]:
if isinstance(item, dict):
if "inputs" in item:
assert inputs is None
inputs = item["inputs"]
elif "outputs" in item:
if outputs is not None:
raise Err.MultipleDefinitions(
sp(self.filename, meta),
"redundant sections in workflow")
outputs = item["outputs"]
if "output_idents" in item:
assert output_idents is None
output_idents = item["output_idents"]
output_idents_pos = item["pos"]
elif "meta" in item:
if meta_section is not None:
raise Err.MultipleDefinitions(
sp(self.filename, meta),
"redundant sections in workflow")
meta_section = item["meta"]
elif "parameter_meta" in item:
if parameter_meta is not None:
raise Err.MultipleDefinitions(
sp(self.filename, meta),
"redundant sections in workflow")
parameter_meta = item["parameter_meta"]
else:
assert False
elif isinstance(item, (D.Call, D.Conditional, D.Decl, D.Scatter)):
elements.append(item)
else:
assert False
_check_keyword(sp(self.filename, meta), items[0].value)
return D.Workflow(
sp(self.filename, meta),
items[0].value,
inputs,
elements,
outputs,
parameter_meta or dict(),
meta_section or dict(),
output_idents,
output_idents_pos,
)
def import_doc(self, items, meta):
uri = items[0]
if len(items) > 1 and isinstance(items[1], str):
namespace = items[1].value
else:
namespace = uri
try:
namespace = namespace[namespace.rindex("/") + 1:]
except ValueError:
pass
if namespace.endswith(".wdl"):
namespace = namespace[:-4]
_check_keyword(sp(self.filename, meta), namespace)
aliases = [p for p in items[1:] if isinstance(p, tuple)]
return D.DocImport(pos=sp(self.filename, meta),
uri=uri,
namespace=namespace,
aliases=aliases,
doc=None)
def document(self, items, meta):
imports = []
tasks = []
workflow = None
for item in items:
if isinstance(item, D.Task):
tasks.append(item)
elif isinstance(item, D.Workflow):
if workflow is not None:
raise Err.MultipleDefinitions(
sp(self.filename, meta), "Document has multiple workflows"
)
workflow = item
elif isinstance(item, lark.Tree) and item.data == "version":
pass
elif isinstance(item, dict) and "import" in item:
imports.append(item["import"])
else:
assert False
return D.Document(sp(self.filename, meta), imports, tasks, workflow, self.imported)
def parse_document(txt: str,
version: Optional[str] = None,
uri: str = "") -> D.Document:
if version is None:
# for now assume the version is 1.0 if the first line is "version <number>"
# otherwise draft-2
version = "draft-2"
for line in txt.split("\n"):
line = line.strip()
if line and line[0] != "#":
if line.startswith("version ") and line[8].isdigit():
version = "1.0"
break
if not txt.strip():
return D.Document(
SourcePosition(filename=uri,
line=0,
column=0,
end_line=0,
end_column=0),
[],
{},
[],
None,
)
try:
return _DocTransformer(uri).transform(parse(txt, "document", version))
except lark.exceptions.UnexpectedInput as exn:
pos = SourcePosition(
filename=(uri if uri != "" else "(buffer)"),
line=getattr(exn, "line", "?"),
column=getattr(exn, "column", "?"),
end_line=getattr(exn, "line", "?"),
end_column=getattr(exn, "column", "?"),
)
raise Err.SyntaxError(pos, str(exn)) from None
except lark.exceptions.VisitError as exn:
raise exn.__context__
def scatter(self, items, meta):
return D.Scatter(sp(self.filename, meta), items[0].value, items[1], items[2:])
def conditional(self, items, meta):
return D.Conditional(sp(self.filename, meta), items[0], items[1:])
def scatter(self, items, meta):
_check_keyword(sp(self.filename, meta), items[0].value)
return D.Scatter(sp(self.filename, meta), items[0].value, items[1],
items[2:])
def call(self, items, meta):
return D.Call(sp(self.filename, meta), items[0], None,
items[1] if len(items) > 1 else dict())