Beispiel #1
0
    def __init__(self, toolpath_object: MutableMapping[str, Any],
                 loadingContext: LoadingContext) -> None:
        """Build a Process object from the provided dictionary."""
        super(Process, self).__init__()
        self.metadata = getdefault(loadingContext.metadata,
                                   {})  # type: Dict[str,Any]
        self.provenance_object = None  # type: Optional[ProvenanceProfile]
        self.parent_wf = None  # type: Optional[ProvenanceProfile]
        global SCHEMA_FILE, SCHEMA_DIR, SCHEMA_ANY  # pylint: disable=global-statement
        if SCHEMA_FILE is None or SCHEMA_ANY is None or SCHEMA_DIR is None:
            get_schema("v1.0")
            SCHEMA_ANY = cast(
                Dict[str, Any],
                SCHEMA_CACHE["v1.0"][3].idx["https://w3id.org/cwl/salad#Any"],
            )
            SCHEMA_FILE = cast(
                Dict[str, Any],
                SCHEMA_CACHE["v1.0"][3].idx["https://w3id.org/cwl/cwl#File"],
            )
            SCHEMA_DIR = cast(
                Dict[str, Any],
                SCHEMA_CACHE["v1.0"]
                [3].idx["https://w3id.org/cwl/cwl#Directory"],
            )

        self.names = schema.make_avro_schema(
            [SCHEMA_FILE, SCHEMA_DIR, SCHEMA_ANY], Loader({}))
        self.tool = toolpath_object
        self.requirements = copy.deepcopy(
            getdefault(loadingContext.requirements, []))
        self.requirements.extend(self.tool.get("requirements", []))
        if "id" not in self.tool:
            self.tool["id"] = "_:" + str(uuid.uuid4())
        self.requirements.extend(
            get_overrides(getdefault(loadingContext.overrides_list, []),
                          self.tool["id"]).get("requirements", []))
        self.hints = copy.deepcopy(getdefault(loadingContext.hints, []))
        self.hints.extend(self.tool.get("hints", []))
        # Versions of requirements and hints which aren't mutated.
        self.original_requirements = copy.deepcopy(self.requirements)
        self.original_hints = copy.deepcopy(self.hints)
        self.doc_loader = loadingContext.loader
        self.doc_schema = loadingContext.avsc_names

        self.formatgraph = None  # type: Optional[Graph]
        if self.doc_loader is not None:
            self.formatgraph = self.doc_loader.graph

        checkRequirements(self.tool, supportedProcessRequirements)
        self.validate_hints(
            loadingContext.avsc_names,
            self.tool.get("hints", []),
            strict=getdefault(loadingContext.strict, False),
        )

        self.schemaDefs = {}  # type: Dict[str,Dict[str, Any]]

        sd, _ = self.get_requirement("SchemaDefRequirement")

        if sd is not None:
            sdtypes = avroize_type(sd["types"])
            av = schema.make_valid_avro(sdtypes,
                                        {t["name"]: t
                                         for t in sdtypes}, set())
            for i in av:
                self.schemaDefs[i["name"]] = i  # type: ignore
            schema.make_avsc_object(schema.convert_to_dict(av), self.names)

        # Build record schema from inputs
        self.inputs_record_schema = {
            "name": "input_record_schema",
            "type": "record",
            "fields": [],
        }  # type: Dict[str, Any]
        self.outputs_record_schema = {
            "name": "outputs_record_schema",
            "type": "record",
            "fields": [],
        }  # type: Dict[str, Any]

        for key in ("inputs", "outputs"):
            for i in self.tool[key]:
                c = copy.deepcopy(i)
                c["name"] = shortname(c["id"])
                del c["id"]

                if "type" not in c:
                    raise validate.ValidationException(
                        "Missing 'type' in parameter '{}'".format(c["name"]))

                if "default" in c and "null" not in aslist(c["type"]):
                    nullable = ["null"]
                    nullable.extend(aslist(c["type"]))
                    c["type"] = nullable
                else:
                    c["type"] = c["type"]
                c["type"] = avroize_type(c["type"], c["name"])
                if key == "inputs":
                    self.inputs_record_schema["fields"].append(c)
                elif key == "outputs":
                    self.outputs_record_schema["fields"].append(c)

        with SourceLine(toolpath_object, "inputs",
                        validate.ValidationException):
            self.inputs_record_schema = cast(
                Dict[str, Any],
                schema.make_valid_avro(self.inputs_record_schema, {}, set()),
            )
            schema.make_avsc_object(
                schema.convert_to_dict(self.inputs_record_schema), self.names)
        with SourceLine(toolpath_object, "outputs",
                        validate.ValidationException):
            self.outputs_record_schema = cast(
                Dict[str, Any],
                schema.make_valid_avro(self.outputs_record_schema, {}, set()),
            )
            schema.make_avsc_object(
                schema.convert_to_dict(self.outputs_record_schema), self.names)

        if toolpath_object.get("class") is not None and not getdefault(
                loadingContext.disable_js_validation, False):
            if loadingContext.js_hint_options_file is not None:
                try:
                    with open(loadingContext.js_hint_options_file
                              ) as options_file:
                        validate_js_options = json.load(options_file)
                except (OSError, ValueError) as err:
                    _logger.error(
                        "Failed to read options file %s",
                        loadingContext.js_hint_options_file,
                    )
                    raise
            else:
                validate_js_options = None
            if self.doc_schema is not None:
                validate_js_expressions(
                    cast(CommentedMap, toolpath_object),
                    self.doc_schema.names[toolpath_object["class"]],
                    validate_js_options,
                )

        dockerReq, is_req = self.get_requirement("DockerRequirement")

        if (dockerReq is not None and "dockerOutputDirectory" in dockerReq
                and is_req is not None and not is_req):
            _logger.warning(
                SourceLine(item=dockerReq, raise_type=str).makeError(
                    "When 'dockerOutputDirectory' is declared, DockerRequirement "
                    "should go in the 'requirements' section, not 'hints'."
                    ""))

        if (dockerReq is not None and is_req is not None and
                dockerReq.get("dockerOutputDirectory") == "/var/spool/cwl"):
            if is_req:
                # In this specific case, it is legal to have /var/spool/cwl, so skip the check.
                pass
            else:
                # Must be a requirement
                var_spool_cwl_detector(self.tool)
        else:
            var_spool_cwl_detector(self.tool)
Beispiel #2
0
    def __init__(self, toolpath_object, **kwargs):
        # type: (Dict[Text, Any], **Any) -> None
        """
        kwargs:

        metadata: tool document metadata
        requirements: inherited requirements
        hints: inherited hints
        loader: schema_salad.ref_resolver.Loader used to load tool document
        avsc_names: CWL Avro schema object used to validate document
        strict: flag to determine strict validation (fail on unrecognized fields)
        """

        self.metadata = kwargs.get("metadata", {})  # type: Dict[Text,Any]
        self.names = None  # type: schema.Names

        global SCHEMA_FILE, SCHEMA_DIR, SCHEMA_ANY  # pylint: disable=global-statement
        if SCHEMA_FILE is None:
            get_schema("v1.0")
            SCHEMA_ANY = cast(
                Dict[Text, Any],
                SCHEMA_CACHE["v1.0"][3].idx["https://w3id.org/cwl/salad#Any"])
            SCHEMA_FILE = cast(
                Dict[Text, Any],
                SCHEMA_CACHE["v1.0"][3].idx["https://w3id.org/cwl/cwl#File"])
            SCHEMA_DIR = cast(
                Dict[Text, Any], SCHEMA_CACHE["v1.0"]
                [3].idx["https://w3id.org/cwl/cwl#Directory"])

        names = schema.make_avro_schema([SCHEMA_FILE, SCHEMA_DIR, SCHEMA_ANY],
                                        Loader({}))[0]
        if isinstance(names, schema.SchemaParseException):
            raise names
        else:
            self.names = names
        self.tool = toolpath_object
        self.requirements = (kwargs.get("requirements", []) + self.tool.get(
            "requirements", []) + get_overrides(kwargs.get(
                "overrides", []), self.tool["id"]).get("requirements", []))
        self.hints = kwargs.get("hints", []) + self.tool.get("hints", [])
        self.formatgraph = None  # type: Graph
        if "loader" in kwargs:
            self.formatgraph = kwargs["loader"].graph

        self.doc_loader = kwargs["loader"]
        self.doc_schema = kwargs["avsc_names"]

        checkRequirements(self.tool, supportedProcessRequirements)
        self.validate_hints(kwargs["avsc_names"],
                            self.tool.get("hints", []),
                            strict=kwargs.get("strict"))

        self.schemaDefs = {}  # type: Dict[Text,Dict[Text, Any]]

        sd, _ = self.get_requirement("SchemaDefRequirement")

        if sd:
            sdtypes = sd["types"]
            av = schema.make_valid_avro(
                sdtypes, {t["name"]: t
                          for t in avroize_type(sdtypes)}, set())
            for i in av:
                self.schemaDefs[i["name"]] = i  # type: ignore
            schema.AvroSchemaFromJSONData(av, self.names)  # type: ignore

        # Build record schema from inputs
        self.inputs_record_schema = {
            "name": "input_record_schema",
            "type": "record",
            "fields": []
        }  # type: Dict[Text, Any]
        self.outputs_record_schema = {
            "name": "outputs_record_schema",
            "type": "record",
            "fields": []
        }  # type: Dict[Text, Any]

        for key in ("inputs", "outputs"):
            for i in self.tool[key]:
                c = copy.copy(i)
                c["name"] = shortname(c["id"])
                del c["id"]

                if "type" not in c:
                    raise validate.ValidationException(u"Missing 'type' in "
                                                       "parameter '%s'" %
                                                       c["name"])

                if "default" in c and "null" not in aslist(c["type"]):
                    c["type"] = ["null"] + aslist(c["type"])
                else:
                    c["type"] = c["type"]
                c["type"] = avroize_type(c["type"], c["name"])
                if key == "inputs":
                    self.inputs_record_schema["fields"].append(c)
                elif key == "outputs":
                    self.outputs_record_schema["fields"].append(c)

        with SourceLine(toolpath_object, "inputs",
                        validate.ValidationException):
            self.inputs_record_schema = cast(
                Dict[six.text_type, Any],
                schema.make_valid_avro(self.inputs_record_schema, {}, set()))
            schema.AvroSchemaFromJSONData(self.inputs_record_schema,
                                          self.names)
        with SourceLine(toolpath_object, "outputs",
                        validate.ValidationException):
            self.outputs_record_schema = cast(
                Dict[six.text_type, Any],
                schema.make_valid_avro(self.outputs_record_schema, {}, set()))
            schema.AvroSchemaFromJSONData(self.outputs_record_schema,
                                          self.names)

        if toolpath_object.get("class") is not None and not kwargs.get(
                "disable_js_validation", False):
            if kwargs.get("js_hint_options_file") is not None:
                try:
                    with open(kwargs["js_hint_options_file"]) as options_file:
                        validate_js_options = json.load(options_file)
                except (OSError, ValueError) as e:
                    _logger.error("Failed to read options file %s" %
                                  kwargs["js_hint_options_file"])
                    raise e
            else:
                validate_js_options = None

            validate_js_expressions(
                cast(CommentedMap, toolpath_object),
                self.doc_schema.names[toolpath_object["class"]],
                validate_js_options)

        dockerReq, is_req = self.get_requirement("DockerRequirement")

        if dockerReq and dockerReq.get("dockerOutputDirectory") and not is_req:
            _logger.warn(
                SourceLine(item=dockerReq, raise_type=Text).makeError(
                    """When 'dockerOutputDirectory' is declared, DockerRequirement
  should go in the 'requirements' section, not 'hints'."""))

        if dockerReq and dockerReq.get(
                "dockerOutputDirectory") == "/var/spool/cwl":
            if is_req:
                # In this specific case, it is legal to have /var/spool/cwl, so skip the check.
                pass
            else:
                # Must be a requirement
                var_spool_cwl_detector(self.tool)
        else:
            var_spool_cwl_detector(self.tool)
Beispiel #3
0
    def __init__(self,
                 toolpath_object,      # type: Dict[Text, Any]
                 loadingContext        # type: LoadingContext
                ):  # type: (...) -> None
        self.metadata = getdefault(loadingContext.metadata, {})  # type: Dict[Text,Any]
        self.provenance_object = None  # type: Optional[CreateProvProfile]
        self.parent_wf = None          # type: Optional[CreateProvProfile]
        global SCHEMA_FILE, SCHEMA_DIR, SCHEMA_ANY  # pylint: disable=global-statement
        if SCHEMA_FILE is None or SCHEMA_ANY is None or SCHEMA_DIR is None:
            get_schema("v1.0")
            SCHEMA_ANY = cast(Dict[Text, Any],
                              SCHEMA_CACHE["v1.0"][3].idx["https://w3id.org/cwl/salad#Any"])
            SCHEMA_FILE = cast(Dict[Text, Any],
                               SCHEMA_CACHE["v1.0"][3].idx["https://w3id.org/cwl/cwl#File"])
            SCHEMA_DIR = cast(Dict[Text, Any],
                              SCHEMA_CACHE["v1.0"][3].idx["https://w3id.org/cwl/cwl#Directory"])

        names = schema.make_avro_schema([SCHEMA_FILE, SCHEMA_DIR, SCHEMA_ANY],
                                        Loader({}))[0]
        if isinstance(names, schema.SchemaParseException):
            raise names
        else:
            self.names = names
        self.tool = toolpath_object
        self.requirements = (getdefault(loadingContext.requirements, []) +
                             self.tool.get("requirements", []) +
                             get_overrides(getdefault(loadingContext.overrides_list, []),
                                           self.tool["id"]).get("requirements", []))
        self.hints = getdefault(loadingContext.hints, []) + self.tool.get("hints", [])
        # Versions of requirements and hints which aren't mutated.
        self.original_requirements = copy.deepcopy(self.requirements)
        self.original_hints = copy.deepcopy(self.hints)
        self.doc_loader = loadingContext.loader
        self.doc_schema = loadingContext.avsc_names

        self.formatgraph = None  # type: Optional[Graph]
        if self.doc_loader:
            self.formatgraph = self.doc_loader.graph

        checkRequirements(self.tool, supportedProcessRequirements)
        self.validate_hints(loadingContext.avsc_names, self.tool.get("hints", []),
                            strict=getdefault(loadingContext.strict, False))

        self.schemaDefs = {}  # type: Dict[Text,Dict[Text, Any]]

        sd, _ = self.get_requirement("SchemaDefRequirement")

        if sd:
            sdtypes = sd["types"]
            av = schema.make_valid_avro(sdtypes, {t["name"]: t for t in avroize_type(sdtypes)}, set())
            for i in av:
                self.schemaDefs[i["name"]] = i  # type: ignore
            schema.AvroSchemaFromJSONData(av, self.names)  # type: ignore

        # Build record schema from inputs
        self.inputs_record_schema = {
            "name": "input_record_schema", "type": "record",
            "fields": []}  # type: Dict[Text, Any]
        self.outputs_record_schema = {
            "name": "outputs_record_schema", "type": "record",
            "fields": []}  # type: Dict[Text, Any]

        for key in ("inputs", "outputs"):
            for i in self.tool[key]:
                c = copy.copy(i)
                c["name"] = shortname(c["id"])
                del c["id"]

                if "type" not in c:
                    raise validate.ValidationException(
                        u"Missing 'type' in parameter '{}'".format(c["name"]))

                if "default" in c and "null" not in aslist(c["type"]):
                    c["type"] = ["null"] + aslist(c["type"])
                else:
                    c["type"] = c["type"]
                c["type"] = avroize_type(c["type"], c["name"])
                if key == "inputs":
                    self.inputs_record_schema["fields"].append(c)
                elif key == "outputs":
                    self.outputs_record_schema["fields"].append(c)

        with SourceLine(toolpath_object, "inputs", validate.ValidationException):
            self.inputs_record_schema = cast(
                Dict[six.text_type, Any], schema.make_valid_avro(
                    self.inputs_record_schema, {}, set()))
            schema.AvroSchemaFromJSONData(self.inputs_record_schema, self.names)
        with SourceLine(toolpath_object, "outputs", validate.ValidationException):
            self.outputs_record_schema = cast(
                Dict[six.text_type, Any],
                schema.make_valid_avro(self.outputs_record_schema, {}, set()))
            schema.AvroSchemaFromJSONData(self.outputs_record_schema, self.names)

        if toolpath_object.get("class") is not None \
                and not getdefault(loadingContext.disable_js_validation, False):
            if loadingContext.js_hint_options_file is not None:
                try:
                    with open(loadingContext.js_hint_options_file) as options_file:
                        validate_js_options = json.load(options_file)
                except (OSError, ValueError) as err:
                    _logger.error(
                        "Failed to read options file %s",
                        loadingContext.js_hint_options_file)
                    raise err
            else:
                validate_js_options = None
            if self.doc_schema is not None:
                validate_js_expressions(cast(CommentedMap, toolpath_object), self.doc_schema.names[toolpath_object["class"]], validate_js_options)

        dockerReq, is_req = self.get_requirement("DockerRequirement")

        if dockerReq and dockerReq.get("dockerOutputDirectory") and not is_req:
            _logger.warning(SourceLine(
                item=dockerReq, raise_type=Text).makeError(
                "When 'dockerOutputDirectory' is declared, DockerRequirement "
                "should go in the 'requirements' section, not 'hints'."""))

        if dockerReq and dockerReq.get("dockerOutputDirectory") == "/var/spool/cwl":
            if is_req:
                # In this specific case, it is legal to have /var/spool/cwl, so skip the check.
                pass
            else:
                # Must be a requirement
                var_spool_cwl_detector(self.tool)
        else:
            var_spool_cwl_detector(self.tool)
Beispiel #4
0
    def __init__(self,
                 toolpath_object,      # type: Dict[Text, Any]
                 loadingContext        # type: LoadingContext
                ):  # type: (...) -> None
        self.metadata = getdefault(loadingContext.metadata, {})  # type: Dict[Text,Any]
        self.provenance_object = None  # type: Optional[CreateProvProfile]
        self.parent_wf = None          # type: Optional[CreateProvProfile]
        global SCHEMA_FILE, SCHEMA_DIR, SCHEMA_ANY  # pylint: disable=global-statement
        if SCHEMA_FILE is None or SCHEMA_ANY is None or SCHEMA_DIR is None:
            get_schema("v1.0")
            SCHEMA_ANY = cast(Dict[Text, Any],
                              SCHEMA_CACHE["v1.0"][3].idx["https://w3id.org/cwl/salad#Any"])
            SCHEMA_FILE = cast(Dict[Text, Any],
                               SCHEMA_CACHE["v1.0"][3].idx["https://w3id.org/cwl/cwl#File"])
            SCHEMA_DIR = cast(Dict[Text, Any],
                              SCHEMA_CACHE["v1.0"][3].idx["https://w3id.org/cwl/cwl#Directory"])

        names = schema.make_avro_schema([SCHEMA_FILE, SCHEMA_DIR, SCHEMA_ANY],
                                        Loader({}))[0]
        if isinstance(names, schema.SchemaParseException):
            raise names
        else:
            self.names = names
        self.tool = toolpath_object
        self.requirements = (getdefault(loadingContext.requirements, []) +
                             self.tool.get("requirements", []) +
                             get_overrides(getdefault(loadingContext.overrides_list, []),
                                           self.tool["id"]).get("requirements", []))
        self.hints = getdefault(loadingContext.hints, []) + self.tool.get("hints", [])
        # Versions of requirements and hints which aren't mutated.
        self.original_requirements = copy.deepcopy(self.requirements)
        self.original_hints = copy.deepcopy(self.hints)
        self.doc_loader = loadingContext.loader
        self.doc_schema = loadingContext.avsc_names

        self.formatgraph = None  # type: Optional[Graph]
        if self.doc_loader:
            self.formatgraph = self.doc_loader.graph

        checkRequirements(self.tool, supportedProcessRequirements)
        self.validate_hints(loadingContext.avsc_names, self.tool.get("hints", []),
                            strict=getdefault(loadingContext.strict, False))

        self.schemaDefs = {}  # type: Dict[Text,Dict[Text, Any]]

        sd, _ = self.get_requirement("SchemaDefRequirement")

        if sd:
            sdtypes = sd["types"]
            av = schema.make_valid_avro(sdtypes, {t["name"]: t for t in avroize_type(sdtypes)}, set())
            for i in av:
                self.schemaDefs[i["name"]] = i  # type: ignore
            schema.AvroSchemaFromJSONData(av, self.names)  # type: ignore

        # Build record schema from inputs
        self.inputs_record_schema = {
            "name": "input_record_schema", "type": "record",
            "fields": []}  # type: Dict[Text, Any]
        self.outputs_record_schema = {
            "name": "outputs_record_schema", "type": "record",
            "fields": []}  # type: Dict[Text, Any]

        for key in ("inputs", "outputs"):
            for i in self.tool[key]:
                c = copy.copy(i)
                c["name"] = shortname(c["id"])
                del c["id"]

                if "type" not in c:
                    raise validate.ValidationException(u"Missing 'type' in "
                            "parameter '%s'" % c["name"])

                if "default" in c and "null" not in aslist(c["type"]):
                    c["type"] = ["null"] + aslist(c["type"])
                else:
                    c["type"] = c["type"]
                c["type"] = avroize_type(c["type"], c["name"])
                if key == "inputs":
                    self.inputs_record_schema["fields"].append(c)
                elif key == "outputs":
                    self.outputs_record_schema["fields"].append(c)

        with SourceLine(toolpath_object, "inputs", validate.ValidationException):
            self.inputs_record_schema = cast(
                Dict[six.text_type, Any], schema.make_valid_avro(
                    self.inputs_record_schema, {}, set()))
            schema.AvroSchemaFromJSONData(self.inputs_record_schema, self.names)
        with SourceLine(toolpath_object, "outputs", validate.ValidationException):
            self.outputs_record_schema = cast(Dict[six.text_type, Any],
                    schema.make_valid_avro(self.outputs_record_schema, {}, set()))
            schema.AvroSchemaFromJSONData(self.outputs_record_schema, self.names)

        if toolpath_object.get("class") is not None and not getdefault(loadingContext.disable_js_validation, False):
            if loadingContext.js_hint_options_file is not None:
                try:
                    with open(loadingContext.js_hint_options_file) as options_file:
                        validate_js_options = json.load(options_file)
                except (OSError, ValueError) as e:
                    _logger.error("Failed to read options file %s" % loadingContext.js_hint_options_file)
                    raise e
            else:
                validate_js_options = None
            if self.doc_schema is not None:
                validate_js_expressions(cast(CommentedMap, toolpath_object), self.doc_schema.names[toolpath_object["class"]], validate_js_options)

        dockerReq, is_req = self.get_requirement("DockerRequirement")

        if dockerReq and dockerReq.get("dockerOutputDirectory") and not is_req:
            _logger.warn(SourceLine(
                item=dockerReq, raise_type=Text).makeError(
                "When 'dockerOutputDirectory' is declared, DockerRequirement "
                "should go in the 'requirements' section, not 'hints'."""))

        if dockerReq and dockerReq.get("dockerOutputDirectory") == "/var/spool/cwl":
            if is_req:
                # In this specific case, it is legal to have /var/spool/cwl, so skip the check.
                pass
            else:
                # Must be a requirement
                var_spool_cwl_detector(self.tool)
        else:
            var_spool_cwl_detector(self.tool)