def pack(document_loader, # type: Loader
processobj, # type: Union[Dict[Text, Any], List[Dict[Text, Any]]]
uri, # type: Text
metadata, # type: Dict[Text, Text]
rewrite_out=None # type: Dict[Text, Text]
): # type: (...) -> Dict[Text, Any]
document_loader = SubLoader(document_loader)
document_loader.idx = {}
if isinstance(processobj, dict):
document_loader.idx[processobj["id"]] = CommentedMap(iteritems(processobj))
elif isinstance(processobj, list):
_, frag = urllib.parse.urldefrag(uri)
for po in processobj:
if not frag:
if po["id"].endswith("#main"):
uri = po["id"]
document_loader.idx[po["id"]] = CommentedMap(iteritems(po))
def loadref(base, uri):
# type: (Optional[Text], Text) -> Union[Dict, List, Text, None]
return document_loader.resolve_ref(uri, base_url=base)[0]
ids = set() # type: Set[Text]
find_ids(processobj, ids)
runs = {uri}
find_run(processobj, loadref, runs)
for f in runs:
find_ids(document_loader.resolve_ref(f)[0], ids)
names = set() # type: Set[Text]
if rewrite_out is None:
rewrite = {} # type: Dict[Text, Text]
else:
rewrite = rewrite_out
mainpath, _ = urllib.parse.urldefrag(uri)
def rewrite_id(r, mainuri):
# type: (Text, Text) -> None
if r == mainuri:
rewrite[r] = "#main"
elif r.startswith(mainuri) and r[len(mainuri)] in ("#", "/"):
if r[len(mainuri):].startswith("#main/"):
rewrite[r] = "#" + uniquename(r[len(mainuri)+1:], names)
else:
rewrite[r] = "#" + uniquename("main/"+r[len(mainuri)+1:], names)
else:
path, frag = urllib.parse.urldefrag(r)
if path == mainpath:
rewrite[r] = "#" + uniquename(frag, names)
else:
if path not in rewrite:
rewrite[path] = "#" + uniquename(shortname(path), names)
sortedids = sorted(ids)
for r in sortedids:
rewrite_id(r, uri)
packed = {"$graph": [], "cwlVersion": metadata["cwlVersion"]
} # type: Dict[Text, Any]
namespaces = metadata.get('$namespaces', None)
schemas = set() # type: Set[Text]
for r in sorted(runs):
dcr, metadata = document_loader.resolve_ref(r)
if isinstance(dcr, CommentedSeq):
dcr = dcr[0]
dcr = cast(CommentedMap, dcr)
if not isinstance(dcr, dict):
continue
for doc in (dcr, metadata):
if "$schemas" in doc:
for s in doc["$schemas"]:
schemas.add(s)
if dcr.get("class") not in ("Workflow", "CommandLineTool", "ExpressionTool"):
continue
dc = cast(Dict[Text, Any], copy.deepcopy(dcr))
v = rewrite[r]
dc["id"] = v
for n in ("name", "cwlVersion", "$namespaces", "$schemas"):
if n in dc:
del dc[n]
packed["$graph"].append(dc)
if schemas:
packed["$schemas"] = list(schemas)
for r in list(rewrite.keys()):
v = rewrite[r]
replace_refs(packed, rewrite, r + "/" if "#" in r else r + "#", v + "/")
import_embed(packed, set())
if len(packed["$graph"]) == 1:
# duplicate 'cwlVersion' inside $graph when there is a single item
# because we're printing contents inside '$graph' rather than whole dict
packed["$graph"][0]["cwlVersion"] = packed["cwlVersion"]
if namespaces:
packed["$graph"][0]["$namespaces"] = dict(cast(Dict, namespaces))
return packed
def pack(
loadingContext: LoadingContext,
uri: str,
rewrite_out: Optional[Dict[str, str]] = None,
loader: Optional[Loader] = None,
) -> CWLObjectType:
# The workflow document we have in memory right now may have been
# updated to the internal CWL version. We need to reload the
# document to go back to its original version.
#
# What's going on here is that the updater replaces the
# documents/fragments in the index with updated ones, the
# index is also used as a cache, so we need to go through the
# loading process with an empty index and updating turned off
# so we have the original un-updated documents.
#
loadingContext = loadingContext.copy()
document_loader = SubLoader(loader or loadingContext.loader or Loader({}))
loadingContext.do_update = False
loadingContext.loader = document_loader
loadingContext.loader.idx = {}
loadingContext.metadata = {}
loadingContext, docobj, uri = fetch_document(uri, loadingContext)
loadingContext, fileuri = resolve_and_validate_document(
loadingContext, docobj, uri, preprocess_only=True
)
if loadingContext.loader is None:
raise Exception("loadingContext.loader cannot be none")
processobj, metadata = loadingContext.loader.resolve_ref(uri)
document_loader = loadingContext.loader
if isinstance(processobj, MutableMapping):
document_loader.idx[processobj["id"]] = CommentedMap(processobj.items())
elif isinstance(processobj, MutableSequence):
_, frag = urllib.parse.urldefrag(uri)
for po in processobj:
if not frag:
if po["id"].endswith("#main"):
uri = po["id"]
document_loader.idx[po["id"]] = CommentedMap(po.items())
document_loader.idx[metadata["id"]] = CommentedMap(metadata.items())
found_versions = {
cast(str, loadingContext.metadata["cwlVersion"])
} # type: Set[str]
def loadref(base: Optional[str], lr_uri: str) -> ResolveType:
lr_loadingContext = loadingContext.copy()
lr_loadingContext.metadata = {}
lr_loadingContext, lr_workflowobj, lr_uri = fetch_document(
lr_uri, lr_loadingContext
)
lr_loadingContext, lr_uri = resolve_and_validate_document(
lr_loadingContext, lr_workflowobj, lr_uri
)
found_versions.add(cast(str, lr_loadingContext.metadata["cwlVersion"]))
if lr_loadingContext.loader is None:
raise Exception("loader should not be None")
return lr_loadingContext.loader.resolve_ref(lr_uri, base_url=base)[0]
input_ids: Set[str] = set()
output_ids: Set[str] = set()
if isinstance(processobj, MutableSequence):
mainobj = processobj[0]
else:
mainobj = processobj
find_ids(cast(Dict[str, Any], mainobj)["inputs"], input_ids)
find_ids(cast(Dict[str, Any], mainobj)["outputs"], output_ids)
runs = {uri}
find_run(processobj, loadref, runs)
# Figure out the highest version, everything needs to be updated
# to it.
m = 0
for fv in found_versions:
m = max(m, ORDERED_VERSIONS.index(fv))
update_to_version = ORDERED_VERSIONS[m]
for f in runs:
find_ids(document_loader.resolve_ref(f)[0], input_ids)
input_names: Set[str] = set()
output_names: Set[str] = set()
rewrite_inputs: Dict[str, str] = {}
rewrite_outputs: Dict[str, str] = {}
mainpath, _ = urllib.parse.urldefrag(uri)
def rewrite_id(
r: str, mainuri: str, rewrite: Dict[str, str], names: Set[str]
) -> None:
if r == mainuri:
rewrite[r] = "#main"
elif r.startswith(mainuri) and r[len(mainuri)] in ("#", "/"):
if r[len(mainuri) :].startswith("#main/"):
rewrite[r] = "#" + uniquename(r[len(mainuri) + 1 :], names)
else:
rewrite[r] = "#" + uniquename("main/" + r[len(mainuri) + 1 :], names)
else:
path, frag = urllib.parse.urldefrag(r)
if path == mainpath:
rewrite[r] = "#" + uniquename(frag, names)
else:
if path not in rewrite:
rewrite[path] = "#" + uniquename(shortname(path), names)
sorted_input_ids = sorted(input_ids)
sorted_output_ids = sorted(output_ids)
for r in sorted_input_ids:
rewrite_id(r, uri, rewrite_inputs, input_names)
for r in sorted_output_ids:
rewrite_id(r, uri, rewrite_outputs, output_names)
packed = CommentedMap(
(("$graph", CommentedSeq()), ("cwlVersion", update_to_version))
)
namespaces = metadata.get("$namespaces", None)
schemas: Set[str] = set()
if "$schemas" in metadata:
for each_schema in metadata["$schemas"]:
schemas.add(each_schema)
for r in sorted(runs):
dcr, metadata = document_loader.resolve_ref(r)
if isinstance(dcr, CommentedSeq):
dcr = dcr[0]
dcr = cast(CommentedMap, dcr)
if not isinstance(dcr, MutableMapping):
continue
dcr = update(
dcr,
document_loader,
r,
loadingContext.enable_dev,
metadata,
update_to_version,
)
if ORIGINAL_CWLVERSION in metadata:
del metadata[ORIGINAL_CWLVERSION]
if ORIGINAL_CWLVERSION in dcr:
del dcr[ORIGINAL_CWLVERSION]
if "$schemas" in metadata:
for s in metadata["$schemas"]:
schemas.add(s)
if dcr.get("class") not in (
"Workflow",
"CommandLineTool",
"ExpressionTool",
"Operation",
):
continue
dc = cast(Dict[str, Any], copy.deepcopy(dcr))
v = rewrite_inputs[r]
dc["id"] = v
for n in ("name", "cwlVersion", "$namespaces", "$schemas"):
if n in dc:
del dc[n]
packed["$graph"].append(dc)
if schemas:
packed["$schemas"] = list(schemas)
if namespaces:
packed["$namespaces"] = namespaces
save_outputs = packed["$graph"][0].pop("outputs")
for r in list(rewrite_inputs.keys()):
v = rewrite_inputs[r]
replace_refs(packed, rewrite_inputs, r + "/" if "#" in r else r + "#", v + "/")
import_embed(packed, set())
packed["$graph"][0]["outputs"] = save_outputs
for r in list(rewrite_outputs.keys()):
v = rewrite_outputs[r]
replace_refs(
packed["$graph"][0]["outputs"],
rewrite_outputs,
r + "/" if "#" in r else r + "#",
v + "/",
)
for r in list(
rewrite_inputs.keys()
): # again, to process the outputSource references
v = rewrite_inputs[r]
replace_refs(packed, rewrite_inputs, r + "/" if "#" in r else r + "#", v + "/")
if len(packed["$graph"]) == 1:
# duplicate 'cwlVersion', '$schemas', and '$namespaces' inside '$graph'
# when there is only a single item because main.print_pack() will print
# the contents inside '$graph' rather than whole dict in this case
packed["$graph"][0]["cwlVersion"] = packed["cwlVersion"]
if schemas:
packed["$graph"][0]["$schemas"] = list(schemas)
if namespaces:
packed["$graph"][0]["$namespaces"] = namespaces
if rewrite_out is not None:
rewrite_out.update(rewrite_inputs)
rewrite_out.update(rewrite_outputs)
return packed
def pack(
loadingContext: LoadingContext,
uri, # type: str
rewrite_out=None, # type: Optional[Dict[str, str]]
loader=None, # type: Optional[Loader]
): # type: (...) -> Dict[str, Any]
# The workflow document we have in memory right now may have been
# updated to the internal CWL version. We need to reload the
# document to go back to its original version.
#
# What's going on here is that the updater replaces the
# documents/fragments in the index with updated ones, the
# index is also used as a cache, so we need to go through the
# loading process with an empty index and updating turned off
# so we have the original un-updated documents.
#
loadingContext = loadingContext.copy()
document_loader = SubLoader(loader or loadingContext.loader or Loader({}))
loadingContext.do_update = False
loadingContext.loader = document_loader
loadingContext.loader.idx = {}
loadingContext.metadata = {}
loadingContext, docobj, uri = fetch_document(uri, loadingContext)
loadingContext, fileuri = resolve_and_validate_document(
loadingContext, docobj, uri, preprocess_only=True
)
if loadingContext.loader is None:
raise Exception("loadingContext.loader cannot be none")
processobj, metadata = loadingContext.loader.resolve_ref(uri)
document_loader = loadingContext.loader
if isinstance(processobj, MutableMapping):
document_loader.idx[processobj["id"]] = CommentedMap(processobj.items())
elif isinstance(processobj, MutableSequence):
_, frag = urllib.parse.urldefrag(uri)
for po in processobj:
if not frag:
if po["id"].endswith("#main"):
uri = po["id"]
document_loader.idx[po["id"]] = CommentedMap(po.items())
document_loader.idx[metadata["id"]] = CommentedMap(metadata.items())
def loadref(base, uri):
# type: (Optional[str], str) -> ResolveType
return document_loader.resolve_ref(uri, base_url=base)[0]
ids = set() # type: Set[str]
find_ids(processobj, ids)
runs = {uri}
find_run(processobj, loadref, runs)
for f in runs:
find_ids(document_loader.resolve_ref(f)[0], ids)
names = set() # type: Set[str]
if rewrite_out is None:
rewrite = {} # type: Dict[str, str]
else:
rewrite = rewrite_out
mainpath, _ = urllib.parse.urldefrag(uri)
def rewrite_id(r, mainuri):
# type: (str, str) -> None
if r == mainuri:
rewrite[r] = "#main"
elif r.startswith(mainuri) and r[len(mainuri)] in ("#", "/"):
if r[len(mainuri) :].startswith("#main/"):
rewrite[r] = "#" + uniquename(r[len(mainuri) + 1 :], names)
else:
rewrite[r] = "#" + uniquename("main/" + r[len(mainuri) + 1 :], names)
else:
path, frag = urllib.parse.urldefrag(r)
if path == mainpath:
rewrite[r] = "#" + uniquename(frag, names)
else:
if path not in rewrite:
rewrite[path] = "#" + uniquename(shortname(path), names)
sortedids = sorted(ids)
for r in sortedids:
rewrite_id(r, uri)
packed = CommentedMap(
(("$graph", CommentedSeq()), ("cwlVersion", metadata["cwlVersion"]))
)
namespaces = metadata.get("$namespaces", None)
schemas = set() # type: Set[str]
if "$schemas" in metadata:
for each_schema in metadata["$schemas"]:
schemas.add(each_schema)
for r in sorted(runs):
dcr, metadata = document_loader.resolve_ref(r)
if isinstance(dcr, CommentedSeq):
dcr = dcr[0]
dcr = cast(CommentedMap, dcr)
if not isinstance(dcr, MutableMapping):
continue
if "$schemas" in metadata:
for s in metadata["$schemas"]:
schemas.add(s)
if dcr.get("class") not in ("Workflow", "CommandLineTool", "ExpressionTool"):
continue
dc = cast(Dict[str, Any], copy.deepcopy(dcr))
v = rewrite[r]
dc["id"] = v
for n in ("name", "cwlVersion", "$namespaces", "$schemas"):
if n in dc:
del dc[n]
packed["$graph"].append(dc)
if schemas:
packed["$schemas"] = list(schemas)
for r in list(rewrite.keys()):
v = rewrite[r]
replace_refs(packed, rewrite, r + "/" if "#" in r else r + "#", v + "/")
import_embed(packed, set())
if len(packed["$graph"]) == 1:
# duplicate 'cwlVersion' and $schemas inside $graph when there is only
# a single item because we will print the contents inside '$graph'
# rather than whole dict
packed["$graph"][0]["cwlVersion"] = packed["cwlVersion"]
if schemas:
packed["$graph"][0]["$schemas"] = list(schemas)
# always include $namespaces in the #main
if namespaces:
packed["$graph"][0]["$namespaces"] = namespaces
return packed
def pack(
document_loader: Loader,
processobj, # type: Union[Dict[str, Any], List[Dict[str, Any]]]
uri, # type: str
metadata, # type: Dict[str, str]
rewrite_out=None, # type: Optional[Dict[str, str]]
): # type: (...) -> Dict[str, Any]
document_loader = SubLoader(document_loader)
document_loader.idx = {}
if isinstance(processobj, MutableMapping):
document_loader.idx[processobj["id"]] = CommentedMap(
processobj.items())
elif isinstance(processobj, MutableSequence):
_, frag = urllib.parse.urldefrag(uri)
for po in processobj:
if not frag:
if po["id"].endswith("#main"):
uri = po["id"]
document_loader.idx[po["id"]] = CommentedMap(po.items())
document_loader.idx[metadata["id"]] = CommentedMap(metadata.items())
def loadref(base, uri):
# type: (Optional[str], str) -> Union[Dict[str, Any], List[Dict[str, Any]], str, None]
return document_loader.resolve_ref(uri, base_url=base)[0]
ids = set() # type: Set[str]
find_ids(processobj, ids)
runs = {uri}
find_run(processobj, loadref, runs)
for f in runs:
find_ids(document_loader.resolve_ref(f)[0], ids)
names = set() # type: Set[str]
if rewrite_out is None:
rewrite = {} # type: Dict[str, str]
else:
rewrite = rewrite_out
mainpath, _ = urllib.parse.urldefrag(uri)
def rewrite_id(r, mainuri):
# type: (str, str) -> None
if r == mainuri:
rewrite[r] = "#main"
elif r.startswith(mainuri) and r[len(mainuri)] in ("#", "/"):
if r[len(mainuri):].startswith("#main/"):
rewrite[r] = "#" + uniquename(r[len(mainuri) + 1:], names)
else:
rewrite[r] = "#" + uniquename("main/" + r[len(mainuri) + 1:],
names)
else:
path, frag = urllib.parse.urldefrag(r)
if path == mainpath:
rewrite[r] = "#" + uniquename(frag, names)
else:
if path not in rewrite:
rewrite[path] = "#" + uniquename(shortname(path), names)
sortedids = sorted(ids)
for r in sortedids:
rewrite_id(r, uri)
packed = CommentedMap(
(("$graph", CommentedSeq()), ("cwlVersion", metadata["cwlVersion"])))
namespaces = metadata.get("$namespaces", None)
schemas = set() # type: Set[str]
if "$schemas" in metadata:
for each_schema in metadata["$schemas"]:
schemas.add(each_schema)
for r in sorted(runs):
dcr, metadata = document_loader.resolve_ref(r)
if isinstance(dcr, CommentedSeq):
dcr = dcr[0]
dcr = cast(CommentedMap, dcr)
if not isinstance(dcr, MutableMapping):
continue
metadata = cast(Dict[str, Any], metadata)
if "$schemas" in metadata:
for s in metadata["$schemas"]:
schemas.add(s)
if dcr.get("class") not in ("Workflow", "CommandLineTool",
"ExpressionTool"):
continue
dc = cast(Dict[str, Any], copy.deepcopy(dcr))
v = rewrite[r]
dc["id"] = v
for n in ("name", "cwlVersion", "$namespaces", "$schemas"):
if n in dc:
del dc[n]
packed["$graph"].append(dc)
if schemas:
packed["$schemas"] = list(schemas)
for r in list(rewrite.keys()):
v = rewrite[r]
replace_refs(packed, rewrite, r + "/" if "#" in r else r + "#",
v + "/")
import_embed(packed, set())
if len(packed["$graph"]) == 1:
# duplicate 'cwlVersion' and $schemas inside $graph when there is only
# a single item because we will print the contents inside '$graph'
# rather than whole dict
packed["$graph"][0]["cwlVersion"] = packed["cwlVersion"]
if schemas:
packed["$graph"][0]["$schemas"] = list(schemas)
# always include $namespaces in the #main
if namespaces:
packed["$graph"][0]["$namespaces"] = namespaces
return packed