Feature iteration
Our geometric interrogation performs extensive feature and manufacturability feedback detection on CAD files that are uploaded to the system. You can even create custom interrogations to specify what features and feedback look like for a part based on process and material information. Previously, the only place to access the information associated with these features and manufacturability feedback was in our part viewer. Now, with feature iteration, we give you all of this information from within P3L so you can use it to influence pricing logic.
If you want to implement custom feature-based runtime/pricing algorithms, this is where to start.
In this article, we will discuss high level principles and how to access feature and manufacturability feedback information. In related articles, we will break down the specifics for each type of manufacturing process, as well as provide more specific examples. Combine feature iteration with custom interrogations for an even more streamlined workflow and customized experience.
Overview
When you request interrogation information within P3L using an analyze_*() command (analyze_mill3(), analyze_sheet_metal(), analyze_lathe(), analyze_wire_edm(), analyze_additive(), or analyze_casting()), it returns a blob of information about the part in context with that manufacturing process type. Lets take the output of an analyze_sheet_metal() call for example. In this blob of information exists important pricing information like total_cut_length, total_hole_setups, thickness, etc that you can use for straightforward pricing logic. However, when you want to break pricing down into a feature by feature exercise, these basic numbers fall short of providing the information you need. Now with feature iteration, we provide helper functions for you to extract all that in depth information about each feature and manufacturability feedback you can see in our part viewer.
How to access information
While the way to access this information is slightly different for each process type, you use four basic iterator functions:
get_setups(analyze_output) -> P3LList
Given the output of an analyze_*() call, this function returns a P3LList of all setups for the part. This method is useful for analyze_mill3() and analyze_wire_edm() outputs, as with these processes, we store features and feedback within the actual setups they pertain to, as well as storing all features and feedback with the top level analyze_output including those that do not pertain to any individual setup.
get_features(analyze_output_or_setup, name=None) -> P3LList
Given the output of an analyze_*() or a setup object from the get_setups iterator, this function returns a P3LList of all features within the object. Setups accessed using a CNC-style process with the INDEX value also can be passed as an input to this function. You can optionally specify a name that will return a list of only the features with a name that matches the specified value. Just calling get_features(analyze_output_or_setup) will return an array of all features.
get_feedback(analyze_output_or_setup, name=None) -> P3LList
Given the output of an analyze_*() or a setup object from the get_setups iterator, this function returns a P3LList of all manufacturability feedback found within the object. Setups accessed using a CNC-style process with the INDEX value also can be passed as an input to this function. You can optionally specify a name that will return a list of only the feedback with a name that matches the specified value. Just calling get_feedback(analyze_output_or_setup) will return an array of all feedback.
How to use information
Every feature instance that comes out of the get_features() function has a name and properties attribute accessed by dot notation:
sheet_metal = analyze_sheet_metal() for feature in get_features(sheet_metal): feature.name feature.properties
Every feedback instance that comes out of the get_feedback() function has a name, level, and properties attribute accessed by dot notation:
sheet_metal = analyze_sheet_metal() for feedback in get_feedback(sheet_metal): feedback.name feedback.level feedback.properties
The level of a feedback object is a string that can hold one of three values: cost_driver, manufacturing_issue, or complete_exclusion. These are assigned internally by Paperless Parts that dictates the general severity of manufacturability feedback instance, where things like deep holes are cost drivers, tight corners are manufacturing issues, and exceeding size limits are complete exclusions. You can use these levels as a general indicator of how problematic a part is for a certain process. A basic usage for the level is seen below using sheet metal:
sheet_metal = analyze_sheet_metal()
complexity_metric = 0
for feedback in get_feedback(sheet_metal):
if feedback.level == 'cost_driver':
complexity_metric += 1
elif feedback.level == 'manufacturing_issue':
complexity_metric += 2
elif feedback.level == 'complete_exclusion':
complexity_metric += 3
Accessing this information from get_setups() works in a very similar manner.
mill = analyze_mill3()
for setup in get_setups(mill):
for feature in get_features(setup):
feature.name
feature.properties
for feedback in get_feedback(setup):
feedback.name
feedback.level
feedback.properties
The properties from a feature or feedback object has a range of attributes that depend on the type of feature or feedback. All of these properties are accessed via dot notation. For example, features with the name bend from a sheet metal output have area, length, radius, angle, and k_factor attributes are accessed and used like the following snippet:
sheet_metal = analyze_sheet_metal() for bend in get_features(sheet_metal, name='bend'): bend.properties.area bend.properties.length bend.properties.radius bend.properties.angle bend.properties.k_factor
You additionally can use all four iterator functions to quickly check the count of setups, features, feedback, or bends:
mill = analyze_mill3() number_of_setups = len(get_setups(mill)) sheet_metal = analyze_sheet_metal() number_of_bends = len(get_bends(sheet_metal)) number_of_feedback = len(get_feedback(sheet_metal)) number_of_features = len(get_features(sheet_metal))
You also can call the features and feedback arrays directly. These will return a P3LList copy of the feature and feedback objects which you can then manipulate with list methods. Check out our article on P3L Lists for more information on how to use these.
sheet_metal = analyze_mill3()
for f in iterate(sheet_metal.features):
do_something = 1
for f in iterate(sheet_metal.feedback):
do_something = 1
bends = sheet_metal.features.filter(lambda x: x.name in create_list('bend', 'curl', 'open_hem', 'tear_drop', 'offset')
for b in iterate(bends):
do_something = 1
Checkout our articles on each process to get lists of feature and feedback names as well as their associated properties and levels. These articles also have interesting examples to work from.