Warning

This document is for an old release of Galaxy. You can alternatively view this page in the latest release if it exists or view the top of the latest release's documentation.

Source code for galaxy.tool_util.linters.inputs

"""This module contains a linting functions for tool inputs."""
import re

from galaxy.util import string_as_bool
from ._util import (
    is_datasource,
    is_valid_cheetah_placeholder
)
from ..parser.util import _parse_name

FILTER_TYPES = [
    'data_meta',
    'param_value',
    'static_value',
    'regexp',
    'unique_value',
    'multiple_splitter',
    'attribute_value_splitter',
    'add_value',
    'remove_value',
    'sort_by',
]

ATTRIB_VALIDATOR_COMPATIBILITY = {
    "check": ["metadata"],
    "expression": ["substitute_value_in_message"],
    "table_name": ["dataset_metadata_in_data_table", "dataset_metadata_not_in_data_table", "value_in_data_table", "value_not_in_data_table"],
    "filename": ["dataset_metadata_in_file"],
    "metadata_name": ["dataset_metadata_in_data_table", "dataset_metadata_not_in_data_table", "dataset_metadata_in_file"],
    "metadata_column": ["dataset_metadata_in_data_table", "dataset_metadata_not_in_data_table", "value_in_data_table", "value_not_in_data_table", "dataset_metadata_in_file"],
    "line_startswith": ["dataset_metadata_in_file"],
    "min": ["in_range", "length", "dataset_metadata_in_range"],
    "max": ["in_range", "length", "dataset_metadata_in_range"],
    "exclude_min": ["in_range", "dataset_metadata_in_range"],
    "exclude_max": ["in_range", "dataset_metadata_in_range"],
    "split": ["dataset_metadata_in_file"],
    "skip": ["metadata"]
}

PARAMETER_VALIDATOR_TYPE_COMPATIBILITY = {
    "integer": ["in_range", "expression"],
    "float": ["in_range", "expression"],
    "data": ["metadata", "unspecified_build", "dataset_ok_validator", "dataset_metadata_in_range", "dataset_metadata_in_file", "dataset_metadata_in_data_table", "dataset_metadata_not_in_data_table", "expression"],
    "data_collection": ["metadata", "unspecified_build", "dataset_ok_validator", "dataset_metadata_in_range", "dataset_metadata_in_file", "dataset_metadata_in_data_table", "dataset_metadata_not_in_data_table", "expression"],
    "text": ["regex", "length", "empty_field", "value_in_data_table", "value_not_in_data_table", "expression"],
    "select": ["no_options", "regex", "length", "empty_field", "value_in_data_table", "value_not_in_data_table", "expression"],
    "drill_down": ["no_options", "regex", "length", "empty_field", "value_in_data_table", "value_not_in_data_table", "expression"],
    "data_column": ["no_options", "regex", "length", "empty_field", "value_in_data_table", "value_not_in_data_table", "expression"]
}


[docs]def lint_inputs(tool_xml, lint_ctx): """Lint parameters in a tool's inputs block.""" datasource = is_datasource(tool_xml) inputs = tool_xml.findall("./inputs//param") # determine node to report for general problems with outputs tool_node = tool_xml.find("./inputs") if tool_node is None: tool_node = tool_xml.getroot() num_inputs = 0 for param in inputs: num_inputs += 1 param_attrib = param.attrib if "name" not in param_attrib and "argument" not in param_attrib: lint_ctx.error("Found param input with no name specified.", node=param) continue param_name = _parse_name(param_attrib.get("name"), param_attrib.get("argument")) if "name" in param_attrib and "argument" in param_attrib: if param_attrib.get("name") == _parse_name(None, param_attrib.get("argument")): lint_ctx.warn(f"Param input [{param_name}] 'name' attribute is redundant if argument implies the same name.", node=param) if param_name.strip() == "": lint_ctx.error("Param input with empty name.", node=param) elif not is_valid_cheetah_placeholder(param_name): lint_ctx.warn(f"Param input [{param_name}] is not a valid Cheetah placeholder.", node=param) # TODO lint for params with duplicated name (in inputs & outputs) if "type" not in param_attrib: lint_ctx.error(f"Param input [{param_name}] input with no type specified.", node=param) continue elif param_attrib["type"].strip() == "": lint_ctx.error(f"Param input [{param_name}] with empty type specified.", node=param) continue param_type = param_attrib["type"] # TODO lint for valid param type - attribute combinations if param_type == "data": if "format" not in param_attrib: lint_ctx.warn(f"Param input [{param_name}] with no format specified - 'data' format will be assumed.", node=param) elif param_type == "select": # get dynamic/statically defined options dynamic_options = param.get("dynamic_options", None) options = param.findall("./options") filters = param.findall("./options/filter") select_options = param.findall('./option') if dynamic_options is not None: lint_ctx.warn(f"Select parameter [{param_name}] uses deprecated 'dynamic_options' attribute.", node=param) # check if options are defined by exactly one possibility if param.getparent().tag != "conditional": if (dynamic_options is not None) + (len(options) > 0) + (len(select_options) > 0) != 1: lint_ctx.error(f"Select parameter [{param_name}] options have to be defined by either 'option' children elements, a 'options' element or the 'dynamic_options' attribute.", node=param) else: if len(select_options) == 0: lint_ctx.error(f"Select parameter of a conditional [{param_name}] options have to be defined by 'option' children elements.", node=param) # lint dynamic options if len(options) == 1: filters = options[0].findall("./filter") # lint filters # TODO check if dataset is available for filters referring other datasets filter_adds_options = False for f in filters: ftype = f.get("type", None) if ftype is None: lint_ctx.error(f"Select parameter [{param_name}] contains filter without type.", node=f) continue if ftype not in FILTER_TYPES: lint_ctx.error(f"Select parameter [{param_name}] contains filter with unknown type '{ftype}'.", node=f) continue if ftype in ['add_value', 'data_meta']: filter_adds_options = True # TODO more linting of filters from_file = options[0].get("from_file", None) from_parameter = options[0].get("from_parameter", None) from_dataset = options[0].get("from_dataset", None) from_data_table = options[0].get("from_data_table", None) # TODO check if input param is present for from_dataset if (from_file is None and from_parameter is None and from_dataset is None and from_data_table is None and not filter_adds_options): lint_ctx.error(f"Select parameter [{param_name}] options tag defines no options. Use 'from_dataset', 'from_data_table', or a filter that adds values.", node=options[0]) for deprecated_attr in ["from_file", "from_parameter", "options_filter_attribute", "transform_lines"]: if options[0].get(deprecated_attr) is not None: lint_ctx.warn(f"Select parameter [{param_name}] options uses deprecated '{deprecated_attr}' attribute.", node=options[0]) if from_dataset is not None and from_data_table is not None: lint_ctx.error(f"Select parameter [{param_name}] options uses 'from_dataset' and 'from_data_table' attribute.", node=options[0]) if options[0].get("meta_file_key", None) is not None and from_dataset is None: lint_ctx.error(f"Select parameter [{param_name}] 'meta_file_key' is only compatible with 'from_dataset'.", node=options[0]) elif len(options) > 1: lint_ctx.error(f"Select parameter [{param_name}] contains multiple options elements.", node=options[1]) # lint statically defined options if any('value' not in option.attrib for option in select_options): lint_ctx.error(f"Select parameter [{param_name}] has option without value", node=param) if any(option.text is None for option in select_options): lint_ctx.warn(f"Select parameter [{param_name}] has option without text", node=param) select_options_texts = list() select_options_values = list() for option in select_options: value = option.attrib.get("value", "") if option.text is None: text = value.capitalize() else: text = option.text select_options_texts.append((text, option.attrib.get("selected", "false"))) select_options_values.append((value, option.attrib.get("selected", "false"))) if len(set(select_options_texts)) != len(select_options_texts): lint_ctx.error(f"Select parameter [{param_name}] has multiple options with the same text content", node=param) if len(set(select_options_values)) != len(select_options_values): lint_ctx.error(f"Select parameter [{param_name}] has multiple options with the same value", node=param) multiple = string_as_bool(param_attrib.get("multiple", "false")) optional = string_as_bool(param_attrib.get("optional", multiple)) if param_attrib.get("display") == "checkboxes": if not multiple: lint_ctx.error(f'Select [{param_name}] `display="checkboxes"` is incompatible with `multiple="false"`, remove the `display` attribute', node=param) if not optional: lint_ctx.error(f'Select [{param_name}] `display="checkboxes"` is incompatible with `optional="false"`, remove the `display` attribute', node=param) if param_attrib.get("display") == "radio": if multiple: lint_ctx.error(f'Select [{param_name}] display="radio" is incompatible with multiple="true"', node=param) if optional: lint_ctx.error(f'Select [{param_name}] display="radio" is incompatible with optional="true"', node=param) # TODO: Validate type, much more... # lint validators # TODO check if dataset is available for validators referring other datasets validators = param.findall("./validator") for validator in validators: vtype = validator.attrib['type'] if param_type in PARAMETER_VALIDATOR_TYPE_COMPATIBILITY: if vtype not in PARAMETER_VALIDATOR_TYPE_COMPATIBILITY[param_type]: lint_ctx.error(f"Parameter [{param_name}]: validator with an incompatible type '{vtype}'", node=validator) for attrib in ATTRIB_VALIDATOR_COMPATIBILITY: if attrib in validator.attrib and vtype not in ATTRIB_VALIDATOR_COMPATIBILITY[attrib]: lint_ctx.error(f"Parameter [{param_name}]: attribute '{attrib}' is incompatible with validator of type '{vtype}'", node=validator) if vtype == "expression": if validator.text is None: lint_ctx.error(f"Parameter [{param_name}]: expression validators are expected to contain text", node=validator) else: try: re.compile(validator.text) except Exception as e: lint_ctx.error(f"Parameter [{param_name}]: '{validator.text}' is no valid regular expression: {str(e)}", node=validator) if vtype not in ["expression", "regex"] and validator.text is not None: lint_ctx.warn(f"Parameter [{param_name}]: '{vtype}' validators are not expected to contain text (found '{validator.text}')", node=validator) if vtype in ["in_range", "length", "dataset_metadata_in_range"] and ("min" not in validator.attrib and "max" not in validator.attrib): lint_ctx.error(f"Parameter [{param_name}]: '{vtype}' validators need to define the 'min' or 'max' attribute(s)", node=validator) if vtype in ["metadata"] and ("check" not in validator.attrib and "skip" not in validator.attrib): lint_ctx.error(f"Parameter [{param_name}]: '{vtype}' validators need to define the 'check' or 'skip' attribute(s)", node=validator) if vtype in ["value_in_data_table", "value_not_in_data_table", "dataset_metadata_in_data_table", "dataset_metadata_not_in_data_table"] and "table_name" not in validator.attrib: lint_ctx.error(f"Parameter [{param_name}]: '{vtype}' validators need to define the 'table_name' attribute", node=validator) conditional_selects = tool_xml.findall("./inputs//conditional") for conditional in conditional_selects: conditional_name = conditional.get('name') if not conditional_name: lint_ctx.error("Conditional without a name", node=conditional) if conditional.get("value_from"): # Probably only the upload tool use this, no children elements continue first_param = conditional.findall("param") if len(first_param) != 1: lint_ctx.error(f"Conditional [{conditional_name}] needs exactly one child <param> found {len(first_param)}", node=conditional) continue first_param = first_param[0] first_param_type = first_param.get('type') if first_param_type == 'boolean': lint_ctx.warn(f'Conditional [{conditional_name}] first param of type="boolean" is discouraged, use a select', node=first_param) elif first_param_type != 'select': lint_ctx.error(f'Conditional [{conditional_name}] first param should have type="select"', node=first_param) continue if first_param_type == 'select': select_options = _find_with_attribute(first_param, 'option', 'value') option_ids = [option.get('value') for option in select_options] else: # boolean option_ids = [ first_param.get('truevalue', 'true'), first_param.get('falsevalue', 'false') ] for incomp in ["optional", "multiple"]: if string_as_bool(first_param.get(incomp, False)): lint_ctx.warn(f'Conditional [{conditional_name}] test parameter cannot be {incomp}="true"', node=first_param) whens = conditional.findall('./when') if any('value' not in when.attrib for when in whens): lint_ctx.error(f"Conditional [{conditional_name}] when without value", node=conditional) when_ids = [w.get('value') for w in whens if w.get('value') is not None] for option_id in option_ids: if option_id not in when_ids: lint_ctx.warn(f"Conditional [{conditional_name}] no <when /> block found for {first_param_type} option '{option_id}'", node=conditional) for when_id in when_ids: if when_id not in option_ids: if first_param_type == 'select': lint_ctx.warn(f"Conditional [{conditional_name}] no <option /> found for when block '{when_id}'", node=conditional) else: lint_ctx.warn(f"Conditional [{conditional_name}] no truevalue/falsevalue found for when block '{when_id}'", node=conditional) if datasource: # TODO only display is subtag of inputs, uihints is a separate top level tag (supporting only attrib minwidth) for datasource_tag in ('display', 'uihints'): if not any(param.tag == datasource_tag for param in inputs): lint_ctx.info(f"{datasource_tag} tag usually present in data sources", node=tool_node) if num_inputs: lint_ctx.info(f"Found {num_inputs} input parameters.", node=tool_node) else: if datasource: lint_ctx.info("No input parameters, OK for data sources", node=tool_node) else: lint_ctx.warn("Found no input parameters.", node=tool_node)
[docs]def lint_repeats(tool_xml, lint_ctx): """Lint repeat blocks in tool inputs.""" repeats = tool_xml.findall("./inputs//repeat") for repeat in repeats: if "name" not in repeat.attrib: lint_ctx.error("Repeat does not specify name attribute.", node=repeat) if "title" not in repeat.attrib: lint_ctx.error("Repeat does not specify title attribute.", node=repeat)
def _find_with_attribute(element, tag, attribute, test_value=None): rval = [] for el in (element.findall(f'./{tag}') or []): if attribute not in el.attrib: continue value = el.attrib[attribute] if test_value is not None: if value == test_value: rval.append(el) else: rval.append(el) return rval