ENH: Expansion of Encoders Implementation for Full Flights. #679
Conversation
phmbressan
added
the
Enhancement
Codecov ReportAttention: Patch coverage is
Additional details and impacted files@@ Coverage Diff @@
## develop #679 +/- ##
===========================================
+ Coverage 75.95% 76.46% +0.51%
===========================================
Files 99 95 -4
Lines 11237 11359 +122
===========================================
+ Hits 8535 8686 +151
+ Misses 2702 2673 -29 ☔ View full report in Codecov by Sentry. |
| } | ||
|
|
||
| @classmethod | ||
| def from_dict(cls, func_dict): |
There was a problem hiding this comment.
I`d like to see an example of you loading a class from the dict.
ENH: Simulation Save and Load in JSON Files
|
I have just noticed that I made a mistake when merging CHANGELOG conflicts for this PR. I will fix it before merging. This does not block reviewing.
|
| """Returns the class by importing its signature. | ||
|
|
||
| Parameters | ||
| ---------- | ||
| signature : str | ||
| Signature of the class. | ||
|
|
||
| Returns | ||
| ------- | ||
| type | ||
| Class defined by the signature. | ||
| """ |
There was a problem hiding this comment.
I still cannot know what is a "class signature", is it the signature of the init class?
what do you have in the "signature" dictionary (which bte the docs says it is a str) exactly?
There was a problem hiding this comment.
You are right, I changed what is a signature a few times during development. I will improve the docs.
| @@ -60,3 +60,10 @@ def __str__(self): | |||
| """ | |||
|
|
|||
| return f"Fluid: {self.name}" | |||
|
|
|||
| def to_dict(self, _): | |||
There was a problem hiding this comment.
I usually see this behavior in other packages, what do you think about it? I dont really understand the reasons of that tho.
| def to_dict(self, _): | |
| def to_dict(self, *): |
There was a problem hiding this comment.
Well, actually I think I will simply revert to having the parameter include_outputs in all of them, and disable the warning if the method does not use it. That brings better standardization, since the call to_dict(include_outputs=...) would always work.
The ** would be a better option than I used here, since it allows for the same interface as I mentioned. However, it also allows for other parameter names, and changing it in the future would be a breaking change.
| def to_dict(self, include_outputs=True): | ||
| data = { | ||
| "rocket": self.rocket, | ||
| "env": self.env, | ||
| "rail_length": self.rail_length, | ||
| "inclination": self.inclination, | ||
| "heading": self.heading, | ||
| "initial_solution": self.initial_solution, | ||
| "terminate_on_apogee": self.terminate_on_apogee, | ||
| "max_time": self.max_time, | ||
| "max_time_step": self.max_time_step, | ||
| "min_time_step": self.min_time_step, | ||
| "rtol": self.rtol, | ||
| "atol": self.atol, | ||
| "time_overshoot": self.time_overshoot, | ||
| "name": self.name, | ||
| "equations_of_motion": self.equations_of_motion, | ||
| } | ||
|
|
||
| if include_outputs: | ||
| data.update( | ||
| { | ||
| "time": self.time, | ||
| "out_of_rail_time": self.out_of_rail_time, | ||
| "out_of_rail_velocity": self.out_of_rail_velocity, | ||
| "out_of_rail_state": self.out_of_rail_state, | ||
| "apogee": self.apogee, | ||
| "apogee_time": self.apogee_time, | ||
| "apogee_x": self.apogee_x, | ||
| "apogee_y": self.apogee_y, | ||
| "apogee_state": self.apogee_state, | ||
| "x_impact": self.x_impact, | ||
| "y_impact": self.y_impact, | ||
| "impact_velocity": self.impact_velocity, | ||
| "impact_state": self.impact_state, | ||
| "x": self.x, | ||
| "y": self.y, | ||
| "z": self.z, | ||
| "vx": self.vx, | ||
| "vy": self.vy, | ||
| "vz": self.vz, | ||
| "e0": self.e0, | ||
| "e1": self.e1, | ||
| "e2": self.e2, | ||
| "e3": self.e3, | ||
| "w1": self.w1, | ||
| "w2": self.w2, | ||
| "w3": self.w3, | ||
| "ax": self.ax, | ||
| "ay": self.ay, | ||
| "az": self.az, | ||
| "alpha1": self.alpha1, | ||
| "alpha2": self.alpha2, | ||
| "alpha3": self.alpha3, | ||
| "altitude": self.altitude, | ||
| "mach_number": self.mach_number, | ||
| "stream_velocity_x": self.stream_velocity_x, | ||
| "stream_velocity_y": self.stream_velocity_y, | ||
| "stream_velocity_z": self.stream_velocity_z, | ||
| "free_stream_speed": self.free_stream_speed, | ||
| "angle_of_attack": self.angle_of_attack, | ||
| "static_margin": self.static_margin, | ||
| "stability_margin": self.stability_margin, | ||
| "latitude": self.latitude, | ||
| "longitude": self.longitude, | ||
| } | ||
| ) | ||
|
|
||
| return data |
There was a problem hiding this comment.
If these are just the __init__ arguments why not just use `inspect.signature(Flight.init).parameters to construct the data dict?
Also, I am pretty sure that all attributes/cached_properties can be accessed using dir() or something similar...
I think some of the to_dict and from_dict can be generalized and might not need to be defined in every class. This reduces maintenance
| @@ -373,6 +400,7 @@ def __init__(self, radius, height, spherical_caps=False, geometry_dict=None): | |||
| """ | |||
| geometry_dict = geometry_dict or {} | |||
| super().__init__(geometry_dict) | |||
| self.__input_radius = radius | |||
There was a problem hiding this comment.
We don't have support for tanks in Monte Carlo yet, but in Stochastic classes I believe that we require the input arguments of the __init__ method to be saved in the class as an attribute with the same name, e.g.:
def __init__(self, radius):
self.radius=radiusis there a way to avoid creating self.__input_radius?
| def __init__(self, *args, **kwargs): | ||
| self.include_outputs = kwargs.pop("include_outputs", True) | ||
| super().__init__(*args, **kwargs) |
There was a problem hiding this comment.
I feel like making the default value of include_outputs as False will be much more useful. There will be more use cases for when it is false
I strongly advise to not include dosctrings yet. Also, they exists mainly for private reasons: we need them in order have the JSONEncoder working. Btw the methods are so simple that they don't even need docstrings. But that's a discussion for future sessions... |
Pull request type
Checklist
black rocketpy/ tests/) has passed locallypytest tests -m slow --runslow) have passed locallyCHANGELOG.mdhas been updated (if relevant)Current behavior
Currently the
_encodingmodule is just a stub and cannot still encode full simulations into a JSON like format.New behavior
This PR expand this module bringing the following:
to_dict;Functionencoding.This PR also introduces decoding functionalities. The way this works is:
to_dictorfrom_dictmethod, they are used to perform the (de)encoding;Breaking change
Additional information
There is much to discuss here on the implementation and maintainability side of things. Main discussion points:
to_dictand (in the future) afrom_dictmethod to some (or all) rocketpy classes?_encodersprivate.simplejson: brings some more types and speed into encoding, but does not help much with custom types;jsonpickle: really great for the typing handler (already writes by default each object type) and is able to serialize. The problem is that is does not handlefunctions/lambdasand I did not find a good way to implement custom handlers for our use case;pickle / dill: the most plug and play solution, however the output is not human readable and not generally compatible between different rocketpy versions.Of course, this is what I understood upon researching and testing, feel free to make any comments or suggest other modules.
Remarks and Future Considerations
Here are some of the steps that could be improved:
AirBrakes+Controllersand new classes.Some important remarks: