ALS004_Alignment-segment-shape-representation #62

JTurbakiewicz · 2023-04-23T20:39:54Z

No description provided.

Ghesselink

Nice work :)

Should we formulate the rule (and the one of ALB002 I wrote before as well) slightly different to allow for more reuse of the already written code, e.i. 'The value must be X'

ifc-gherkin-rules/features/steps/thens/values.py

Lines 11 to 66 in 47b6a14

    
           @then("The value must {constraint}") 
        
           @then("The values must {constraint}") 
        
           @then('At least "{num:d}" value must {constraint}') 
        
           @then('At least "{num:d}" values must {constraint}') 
        
           def step_impl(context, constraint, num=None): 
        
               errors = [] 
        
               within_model = getattr(context, 'within_model', False) 
        
               #to account for order-dependency of removing characters from constraint 
        
               while constraint.startswith('be ') or constraint.startswith('in '): 
        
                   constraint = constraint[3:] 
        
               if getattr(context, 'applicable', True): 
        
                   stack_tree = list( 
        
                       filter(None, list(map(lambda layer: layer.get('instances'), context._stack)))) 
        
                   instances = [context.instances] if within_model else context.instances 
        
                   if constraint in ('identical', 'unique'): 
        
                       for i, values in enumerate(instances): 
        
                           if not values: 
        
                               continue 
        
                           attribute = getattr(context, 'attribute', None) 
        
                           if constraint == 'identical' and not all([values[0] == i for i in values]): 
        
                               incorrect_values = values  # a more general approach of going through stack frames to return relevant information in error message? 
        
                               incorrect_insts = stack_tree[-1] 
        
                               errors.append(err.IdenticalValuesError(incorrect_insts, incorrect_values, attribute,)) 
        
                           if constraint == 'unique': 
        
                               seen = set() 
        
                               duplicates = [x for x in values if x in seen or seen.add(x)] 
        
                               if not duplicates: 
        
                                   continue 
        
                               inst_tree = [t[i] for t in stack_tree] 
        
                               inst = inst_tree[-1] 
        
                               incorrect_insts = [inst_tree[1][i] for i, x in enumerate(values) if x in duplicates] 
        
                               incorrect_values = duplicates 
        
                               # avoid mentioning ifcopenshell.entity_instance twice in error message 
        
                               report_incorrect_insts = any(misc.map_state(values, lambda v: misc.do_try( 
        
                                   lambda: isinstance(v, ifcopenshell.entity_instance), False))) 
        
                               errors.append(err.DuplicateValueError(inst, incorrect_values, attribute, incorrect_insts, report_incorrect_insts)) 
        
                   if constraint[-5:] == ".csv'": 
        
                       csv_name = constraint.strip("'") 
        
                       for i, values in enumerate(instances): 
        
                           if not values: 
        
                               continue 
        
                           attribute = getattr(context, 'attribute', None) 
        
                           dirname = os.path.dirname(__file__) 
        
                           filename = Path(dirname).parent.parent / "resources" / csv_name 
        
                           valid_values = [row[0] for row in csv.reader(open(filename))] 
        
                           for iv, value in enumerate(values): 
        
                               if not value in valid_values: 
        
                                   errors.append(err.InvalidValueError([t[i] for t in stack_tree][1][iv], attribute, value)) 
        
               misc.handle_errors(context, errors)

An idea would then be (something like) :

Scenario 1 : Agreement on Representation being Axis

Given a file with Schema Identifier "IFC4X3_TC1" or "IFC4X3_ADD1" or "IFC4X3"
And An IfcAlignmentSegment
And Its attribute Representation
And Its attribute Representations
And Its attribute RepresentationIdentifier
Then Its value must be 'Axis'

Scenario 2 : Agreement on RepresentationType being Segment

Given Repeat steps 1,2,3,4,5 #maybe change this to 1-5?
Then The value must be Segment

The functionality of this 'Repeat steps' can be found in the open PR for GEM005

ifc-gherkin-rules/features/steps/giventhens/repeat_return.py

Lines 6 to 17 in 0f2a256

    
           @given("Repeat step {step_count}") 
        
           @given("Repeat steps {step_count}") 
        
           @then("Repeat step {step_count}") 
        
           @then("Repeat steps {step_count}") 
        
           def step_impl(context, step_count): 
        
               step_stack = list( 
        
                   filter(None, list(map(lambda layer: layer.get('step'), context._stack)))) 
        
               step_stack.reverse() 
        
               step_count = list(step_count.split(',')) 
        
               steps = ('\n').join( 
        
                   [(' ').join(['Given', step_stack[int(n)-2].name + ' ']) for n in step_count]) 
        
               context.execute_steps(steps)

Scenario 3 : Agreement on items being IfcCurveSegment

Given repeat steps 1 - 5
Then The value must be 'IfcCurveSegment' #or change to 'items' and add @ to the step impl

This way we get rid of some duplicate code and reuse 'The value must be {constraint}' often. We could even distinquish between 'type' and 'value' in the gherkin implementation by adding 'The {type/value/items} must be {constraint} and perhaps the feature file itself becomes more readable (we could potentially use the scenario description as a description of what is being tested for the platform itself).
But maybe I am making it too complicated, I'm not sure .. What do you think?

features/RI009_Alignment-segment-shape-representation.feature

features/steps/thens/attributes.py

test/files/ri009/pass-ri009-alignment-segment-shape-representation.ifc

evandroAlfieri

Looks good. Just the unfortunate renaming of all RI009.

features/RI009_Alignment-segment-shape-representation.feature

test/files/ri009/README.MD

test/files/ri009/fail-ri009-scenario01-wrong-items-type.ifc

test/files/ri009/fail-ri009-scenario01-wrong-representationidentifier-value.ifc

test/files/ri009/fail-ri009-scenario01-wrong-representationtype-value.ifc

test/files/ri009/pass-ri009-alignment-segment-shape-representation.ifc

JTurbakiewicz · 2023-04-26T21:46:32Z

Nice work :)

Should we formulate the rule (and the one of ALB002 I wrote before as well) slightly different to allow for more reuse of the already written code, e.i. 'The value must be X'

ifc-gherkin-rules/features/steps/thens/values.py

Lines 11 to 66 in 47b6a14

@then("The value must {constraint}")

@then("The values must {constraint}")

@then('At least "{num:d}" value must {constraint}')

@then('At least "{num:d}" values must {constraint}')

def step_impl(context, constraint, num=None):

errors = []

within_model = getattr(context, 'within_model', False)

#to account for order-dependency of removing characters from constraint

while constraint.startswith('be ') or constraint.startswith('in '):

constraint = constraint[3:]

if getattr(context, 'applicable', True):

stack_tree = list(

filter(None, list(map(lambda layer: layer.get('instances'), context._stack))))

instances = [context.instances] if within_model else context.instances

if constraint in ('identical', 'unique'):

for i, values in enumerate(instances):

if not values:

continue

attribute = getattr(context, 'attribute', None)

if constraint == 'identical' and not all([values[0] == i for i in values]):

incorrect_values = values # a more general approach of going through stack frames to return relevant information in error message?

incorrect_insts = stack_tree[-1]

errors.append(err.IdenticalValuesError(incorrect_insts, incorrect_values, attribute,))

if constraint == 'unique':

seen = set()

duplicates = [x for x in values if x in seen or seen.add(x)]

if not duplicates:

continue

inst_tree = [t[i] for t in stack_tree]

inst = inst_tree[-1]

incorrect_insts = [inst_tree[1][i] for i, x in enumerate(values) if x in duplicates]

incorrect_values = duplicates

# avoid mentioning ifcopenshell.entity_instance twice in error message

report_incorrect_insts = any(misc.map_state(values, lambda v: misc.do_try(

lambda: isinstance(v, ifcopenshell.entity_instance), False)))

errors.append(err.DuplicateValueError(inst, incorrect_values, attribute, incorrect_insts, report_incorrect_insts))

if constraint[-5:] == ".csv'":

csv_name = constraint.strip("'")

for i, values in enumerate(instances):

if not values:

continue

attribute = getattr(context, 'attribute', None)

dirname = os.path.dirname(__file__)

filename = Path(dirname).parent.parent / "resources" / csv_name

valid_values = [row[0] for row in csv.reader(open(filename))]

for iv, value in enumerate(values):

if not value in valid_values:

errors.append(err.InvalidValueError([t[i] for t in stack_tree][1][iv], attribute, value))

misc.handle_errors(context, errors)

An idea would then be (something like) :

Scenario 1 : Agreement on Representation being Axis
Given a file with Schema Identifier "IFC4X3_TC1" or "IFC4X3_ADD1" or "IFC4X3"
And An IfcAlignmentSegment
And Its attribute Representation
And Its attribute Representations
And Its attribute RepresentationIdentifier
Then Its value must be 'Axis'
Scenario 2 : Agreement on RepresentationType being Segment
Given Repeat steps 1,2,3,4,5 #maybe change this to 1-5?
Then The value must be Segment
The functionality of this 'Repeat steps' can be found in the open PR for GEM005

ifc-gherkin-rules/features/steps/giventhens/repeat_return.py

Lines 6 to 17 in 0f2a256

@given("Repeat step {step_count}")

@given("Repeat steps {step_count}")

@then("Repeat step {step_count}")

@then("Repeat steps {step_count}")

def step_impl(context, step_count):

step_stack = list(

filter(None, list(map(lambda layer: layer.get('step'), context._stack))))

step_stack.reverse()

step_count = list(step_count.split(','))

steps = ('\n').join(

[(' ').join(['Given', step_stack[int(n)-2].name + ' ']) for n in step_count])

context.execute_steps(steps)

Scenario 3 : Agreement on items being IfcCurveSegment
Given repeat steps 1 - 5
Then The value must be 'IfcCurveSegment' #or change to 'items' and add @ to the step impl
This way we get rid of some duplicate code and reuse 'The value must be {constraint}' often. We could even distinquish between 'type' and 'value' in the gherkin implementation by adding 'The {type/value/items} must be {constraint} and perhaps the feature file itself becomes more readable (we could potentially use the scenario description as a description of what is being tested for the platform itself). But maybe I am making it too complicated, I'm not sure .. What do you think?

Hi!

Thank you for a great comment.

I think to setup multiple scenarios with repeatable code we could also use the Background feature, as experimented here: https://github.com/buildingSMART/ifc-gherkin-rules/pull/37/files#diff-f12e75ce63549024543ab0c71a93b3cc880178205d26599482a386aa6041d205R5

I find it a bit more readable than Given repeat steps 1 - 5, but that's debatable of course.

As for using the already present step: 'The value must be {constraint}' (thens/values.py).
I think this is a good idea, but I would have to add a decorator, to have something like:

@then("The value must be equal to: {constraint}")  # HERE CHANGE
@then("The value must {constraint}")
@then("The values must {constraint}")
@then('At least "{num:d}" value must {constraint}')
@then('At least "{num:d}" values must {constraint}')
def step_impl(context, constraint, num=None):

And I would add the logic checking the value being equal to the provided value to this step, as currently {constraint} checks for 'identical', 'unique' mostly. But I do agree that it would perhaps be cleaner to use this step.

What do you think? @Ghesselink

pjanck · 2023-04-28T10:18:03Z

An idea would then be (something like) :
Scenario 1 : Agreement on Representation being Axis
Given a file with Schema Identifier "IFC4X3_TC1" or "IFC4X3_ADD1" or "IFC4X3"
And An IfcAlignmentSegment
And Its attribute Representation
And Its attribute Representations
And Its attribute RepresentationIdentifier
Then Its value must be 'Axis'
Scenario 2 : Agreement on RepresentationType being Segment
Given Repeat steps 1,2,3,4,5 #maybe change this to 1-5?
Then The value must be Segment

I think to setup multiple scenarios with repeatable code we could also use the Background feature, as experimented here: https://github.com/buildingSMART/ifc-gherkin-rules/pull/37/files#diff-f12e75ce63549024543ab0c71a93b3cc880178205d26599482a386aa6041d205R5

I agree; the way I understand the proposal, I also think that this reinvents an existing Gherkin feature: the Background keyword.

I would have to add a decorator

This is a code-architecture decision. I don't have any strong opinions on this.

Ghesselink · 2023-04-30T11:20:51Z

I also think that this reinvents an existing Gherkin feature

think to setup multiple scenarios with repeatable code we could also use the Background feature

Agree with both of you @pjanck @JTurbakiewicz . In this case it seems clearer.

So then it would be something like :

Background : AlignmentSegment Representation

Given a file with Schema Identifier "IFC4X3_TC1" or "IFC4X3_ADD1" or "IFC4X3"
And An IfcAlignmentSegment
And Its attribute Representation
And Its attribute Representations
And Its attribute RepresentationIdentifier

Scenario : Agreemen on Representation being Axis
Then The value must be 'Axis'

Scenario: Agreement on RepresentationType being Segment 
Then: The value must be 'Segment' 

Scenario: Agreement on IfcCurveSegment
Then: The value must be 'IfcCurveSegment'

This would look very clear indeed, nice :-)

I think this is a good idea, but I would have to add a decorator, to have something like

That's not required I think as it's already covered by the @The value must {constraint} . In the example of 'The value must be 'Axis', the constraint is 'be Axis' and 'be' part will be stripped from the constraint, leaving only 'Axis' in this case. https://github.com/JTurbakiewicz/ifc-gherkin-rules/blob/bb4a7f3c2b3db3a65e20f5b93e5a3e3fa39cd165/features/steps/thens/values.py#L21-L22 .
Then there seems to be two options of implementation:

If there is nothing 'extra' stated in the constraint (such as identical/unique), as is the case above, implement it as 'equal to'. That also leaves the option open to implement the specific 'equal to' as actually equality between to values if this use case comes up. Then we'll need to check for the combination of a literal 'be' followed by a value.
Require for the gherkin statement to specifically state 'value must be equal to Axis' and implement it in a similar way as identical and unique. https://github.com/JTurbakiewicz/ifc-gherkin-rules/blob/bb4a7f3c2b3db3a65e20f5b93e5a3e3fa39cd165/features/steps/thens/values.py#L29 .

While I am writing this down, I think I'm leaning towards the first option. It seems to me that it will more understandable for someone else reading or making own gherkin statements (including future us).
The implementation in python would then be to change https://github.com/JTurbakiewicz/ifc-gherkin-rules/blob/bb4a7f3c2b3db3a65e20f5b93e5a3e3fa39cd165/features/steps/thens/values.py#L21-L22 With something like :

while constraint.startswith('be'):
   constraint = constraint[3:]
   if re.search(r'\s', constraint): #check if str containts one value, i.e. not containing whitespaces
      if not constraint in context.instances: # untested yet, perhaps needs to be modified to know which instances is causing the error). 
          errors.append(some_error_category) #tbd 
 while constraint.startswith('in'):
     constraint = constraint[3:]

I'll test and work this out further, but would first like to know if you agree/I make sense :) @JTurbakiewicz

JTurbakiewicz · 2023-04-30T20:28:41Z

I also think that this reinvents an existing Gherkin feature

think to setup multiple scenarios with repeatable code we could also use the Background feature

Agree with both of you @pjanck @JTurbakiewicz . In this case it seems clearer.

So then it would be something like :
Background : AlignmentSegment Representation

Given a file with Schema Identifier "IFC4X3_TC1" or "IFC4X3_ADD1" or "IFC4X3"
And An IfcAlignmentSegment
And Its attribute Representation
And Its attribute Representations
And Its attribute RepresentationIdentifier

Scenario : Agreemen on Representation being Axis
Then The value must be 'Axis'

Scenario: Agreement on RepresentationType being Segment 
Then: The value must be 'Segment' 

Scenario: Agreement on IfcCurveSegment
Then: The value must be 'IfcCurveSegment'
This would look very clear indeed, nice :-)

I think this is a good idea, but I would have to add a decorator, to have something like

That's not required I think as it's already covered by the @The value must {constraint} . In the example of 'The value must be 'Axis', the constraint is 'be Axis' and 'be' part will be stripped from the constraint, leaving only 'Axis' in this case. https://github.com/JTurbakiewicz/ifc-gherkin-rules/blob/bb4a7f3c2b3db3a65e20f5b93e5a3e3fa39cd165/features/steps/thens/values.py#L21-L22 . Then there seems to be two options of implementation:

If there is nothing 'extra' stated in the constraint (such as identical/unique), as is the case above, implement it as 'equal to'. That also leaves the option open to implement the specific 'equal to' as actually equality between to values if this use case comes up. Then we'll need to check for the combination of a literal 'be' followed by a value.

Require for the gherkin statement to specifically state 'value must be equal to Axis' and implement it in a similar way as identical and unique. https://github.com/JTurbakiewicz/ifc-gherkin-rules/blob/bb4a7f3c2b3db3a65e20f5b93e5a3e3fa39cd165/features/steps/thens/values.py#L29 .

While I am writing this down, I think I'm leaning towards the first option. It seems to me that it will more understandable for someone else reading or making own gherkin statements (including future us). The implementation in python would then be to change https://github.com/JTurbakiewicz/ifc-gherkin-rules/blob/bb4a7f3c2b3db3a65e20f5b93e5a3e3fa39cd165/features/steps/thens/values.py#L21-L22 With something like :
while constraint.startswith('be'):
   constraint = constraint[3:]
   if re.search(r'\s', constraint): #check if str containts one value, i.e. not containing whitespaces
      if not constraint in context.instances: # untested yet, perhaps needs to be modified to know which instances is causing the error). 
          errors.append(some_error_category) #tbd 
 while constraint.startswith('in'):
     constraint = constraint[3:]
I'll test and work this out further, but would first like to know if you agree/I make sense :) @JTurbakiewicz

Makes sense. I'll play with it a little bit as well.

The additional constraint of ours is to have all the scenarios, so:


    Then The value of attribute RepresentationIdentifier must be Axis
    And The value of attribute RepresentationType must be Segment
    And The type of attribute Items must be IfcCurveSegment

Be present in the same time. I'll check if background scenario lets us do that.

JTurbakiewicz added 4 commits April 20, 2023 01:14

Initial commit

5918c6f

Initial full commit

a05217b

Should to must

06d7f47

Fix

e89dba3

JTurbakiewicz requested review from pjanck, aothms, Ghesselink and evandroAlfieri April 23, 2023 23:13

Ghesselink reviewed Apr 24, 2023

View reviewed changes

pjanck suggested changes Apr 24, 2023

View reviewed changes

features/RI009_Alignment-segment-shape-representation.feature Outdated Show resolved Hide resolved

features/steps/thens/attributes.py Outdated Show resolved Hide resolved

test/files/ri009/pass-ri009-alignment-segment-shape-representation.ifc Outdated Show resolved Hide resolved

evandroAlfieri reviewed Apr 25, 2023

View reviewed changes

MR comments

bb4a7f3

JTurbakiewicz changed the title ~~RI009_Alignment-segment-shape-representation~~ ALS004_Alignment-segment-shape-representation Apr 26, 2023

JTurbakiewicz requested a review from evandroAlfieri April 28, 2023 09:57

aothms added 2 commits September 20, 2023 09:19

Merge branch 'main' into ri09_alignment-segment-shape-representation

be0f59e

Resolve duplicated impl after merge

169e50b

aothms merged commit fcc1070 into buildingSMART:main Sep 20, 2023
1 check passed

Ghesselink mentioned this pull request Sep 24, 2023

Code structure and parsing of 'entity to value' and 'entity -> relationship -> entity' #90

Open

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

ALS004_Alignment-segment-shape-representation #62

ALS004_Alignment-segment-shape-representation #62

JTurbakiewicz commented Apr 23, 2023

Ghesselink left a comment •

edited

Loading

evandroAlfieri left a comment

JTurbakiewicz commented Apr 26, 2023

pjanck commented Apr 28, 2023

Ghesselink commented Apr 30, 2023 •

edited

Loading

JTurbakiewicz commented Apr 30, 2023

	@then("The value must {constraint}")
	@then("The values must {constraint}")
	@then('At least "{num:d}" value must {constraint}')
	@then('At least "{num:d}" values must {constraint}')
	def step_impl(context, constraint, num=None):
	errors = []

	within_model = getattr(context, 'within_model', False)

	#to account for order-dependency of removing characters from constraint
	while constraint.startswith('be ') or constraint.startswith('in '):
	constraint = constraint[3:]

	if getattr(context, 'applicable', True):
	stack_tree = list(
	filter(None, list(map(lambda layer: layer.get('instances'), context._stack))))
	instances = [context.instances] if within_model else context.instances

	if constraint in ('identical', 'unique'):
	for i, values in enumerate(instances):
	if not values:
	continue
	attribute = getattr(context, 'attribute', None)
	if constraint == 'identical' and not all([values[0] == i for i in values]):
	incorrect_values = values # a more general approach of going through stack frames to return relevant information in error message?
	incorrect_insts = stack_tree[-1]
	errors.append(err.IdenticalValuesError(incorrect_insts, incorrect_values, attribute,))
	if constraint == 'unique':
	seen = set()
	duplicates = [x for x in values if x in seen or seen.add(x)]
	if not duplicates:
	continue
	inst_tree = [t[i] for t in stack_tree]
	inst = inst_tree[-1]
	incorrect_insts = [inst_tree[1][i] for i, x in enumerate(values) if x in duplicates]
	incorrect_values = duplicates
	# avoid mentioning ifcopenshell.entity_instance twice in error message
	report_incorrect_insts = any(misc.map_state(values, lambda v: misc.do_try(
	lambda: isinstance(v, ifcopenshell.entity_instance), False)))
	errors.append(err.DuplicateValueError(inst, incorrect_values, attribute, incorrect_insts, report_incorrect_insts))
	if constraint[-5:] == ".csv'":
	csv_name = constraint.strip("'")
	for i, values in enumerate(instances):
	if not values:
	continue
	attribute = getattr(context, 'attribute', None)

	dirname = os.path.dirname(__file__)
	filename = Path(dirname).parent.parent / "resources" / csv_name
	valid_values = [row[0] for row in csv.reader(open(filename))]

	for iv, value in enumerate(values):
	if not value in valid_values:
	errors.append(err.InvalidValueError([t[i] for t in stack_tree][1][iv], attribute, value))

	misc.handle_errors(context, errors)

	@given("Repeat step {step_count}")
	@given("Repeat steps {step_count}")
	@then("Repeat step {step_count}")
	@then("Repeat steps {step_count}")
	def step_impl(context, step_count):
	step_stack = list(
	filter(None, list(map(lambda layer: layer.get('step'), context._stack))))
	step_stack.reverse()
	step_count = list(step_count.split(','))
	steps = ('\n').join(
	[(' ').join(['Given', step_stack[int(n)-2].name + ' ']) for n in step_count])
	context.execute_steps(steps)

ALS004_Alignment-segment-shape-representation #62

ALS004_Alignment-segment-shape-representation #62

Conversation

JTurbakiewicz commented Apr 23, 2023

Ghesselink left a comment • edited Loading

Choose a reason for hiding this comment

evandroAlfieri left a comment

Choose a reason for hiding this comment

JTurbakiewicz commented Apr 26, 2023

pjanck commented Apr 28, 2023

Ghesselink commented Apr 30, 2023 • edited Loading

JTurbakiewicz commented Apr 30, 2023

Ghesselink left a comment •

edited

Loading

Ghesselink commented Apr 30, 2023 •

edited

Loading