We Keep Coding

Migrating SimpleITK 1.x to 2.x

I'm working on a project on GitHub that was made with Python 2.7 (https://github.com/AIM-Harvard/DeepCAC)
I've made most relevant changes as to update it to Python 3.7, but I'm fixed on an error message regarding simpleITK.
Error message:
TypeError: Execute() takes 2 positional arguments but 10 were given
It stems from this code:
res_filter = sitk.ResampleImageFilter()
----> img_sitk = res_filter.Execute(img_sitk, curated_size, sitk.Transform(), method, img_sitk.GetOrigin(), curated_spacing, img_sitk.GetDirection(), 0, img_sitk.GetPixelIDValue())
According to the simpleITK document on switching from 1.x to 2.x (version that is available with python 3.7) it should be done like this (https://simpleitk.readthedocs.io/en/master/migrationGuide2.0.html#filter-s-execute-method) but I can't quite grasp it.
Can someone help out?
Thanks

It looks like you're trying to use the procedural version, Resample, not the class version, ResampleImageFilter.
You can see the documentation for the Resample function here:
https://simpleitk.org/doxygen/latest/html/namespaceitk_1_1simple.html#aadbb243c10d1aedea8e955e8beda4df0
You want to the second version of Resample. So your code would look like this:
img_sitk = sitk.Resample(img_sitk, curated_size, sitk.Transform(), method, img_sitk.GetOrigin(), curated_spacing, img_sitk.GetDirection(), 0, img_sitk.GetPixelIDValue())
If you want to use the ResampleImageFilter, you would set all the parameters using the filter's various Set methods, and then just call the Execute method with the input image as the only parameter.

kubeflow AttributeError: 'ComponentStore' object has no attribute 'uri_search_template'

I am trying to load some prebuit gcp kubeflow components using kfp.components.ComponentStore. However I am getting this error:
line 180, in _load_component_spec_in_component_ref
if self.uri_search_template:
AttributeError: 'ComponentStore' object has no attribute 'uri_search_template'
when at this line of code:
mlengine_train_op = component_store.load_component('ml_engine/train')
KFP version:
kfp 1.8.10
kfp-pipeline-spec 0.1.13
kfp-server-api 1.7.1
Steps to reproduce
import kfp
from kfp.components import func_to_container_op
COMPONENT_URL_SEARCH_PREFIX = "https://raw.githubusercontent.com/kubeflow/pipelines/1.7.1/components/gcp/"
component_store = kfp.components.ComponentStore(
local_search_paths=None, url_search_prefixes=[COMPONENT_URL_SEARCH_PREFIX])
mlengine_train_op = component_store.load_component('ml_engine/train')
mlengine_deploy_op = component_store.load_component('ml_engine/deploy')
I got this code from an example that was running on kfp 0.2.5. This is probably some version problem but does anyone knows how to update this code to the newer version?

I solved the issue. It is something that #Kabilan Mohanraj also mentioned is his comment. It looks like there is a bug if you don't set up uri_search_template when creating the ComponentStore.
I solved it by providing it as parameter, it doesn't really matter what you pass to the constructor really, but it is required because in the code there is the following
if self.url_search_prefixes:
but if you don't pass something that attribute doesn't get created in the constructor.
I fixed it like this:
component_store = kfp.components.ComponentStore(
local_search_paths=["local_search_path"], url_search_prefixes=[COMPONENT_URL_SEARCH_PREFIX]), uri_search_template="{name}/"
I hope it will help someone in the future.
I suggested in the github issue to address this by either make the error more clear or to have better check on uri_search_template attribute.

No module named 'sklearn.utils.linear_assignment_'

I am trying to run a project from github , every object counter applications using sort algorithm. I can't run any of them because of a specific error, attaching errors screenshot. Can anyone help me about fixing this issue?

The linear_assignment function is deprecated in 0.21 and will be removed from 0.23, but sklearn.utils.linear_assignment_ can be replaced by scipy.optimize.linear_sum_assignment.
You can use:
from scipy.optimize import linear_sum_assignment as linear_assignment
then you can run the file and don't need to change the code.

pip install scikit-learn==0.22.2

As yiakwy points out in a github comment the scipy.optimize.linear_sum_assignment is not the perfect replacement:
I am concerned that linear_sum_assignment is not equivalent to linear_assignment which later implements "maximum values" matching strategy not "complete matching" strategy, i.e. in tracking problem maybe an old landmark lost and a new detection coming in. We don't have to make a complete assignment, just match as more as possible.
I have found this out while trying to use it inside SORT-based yolo tracking code which that replacement broke (I was lucky that it did otherwise, I would get wrong results from the experiments without realising it...)
Instead, I suggest copying the module itself to the last version of sklearn and include as module in your code.
https://github.com/scikit-learn/scikit-learn/blob/0.22.X/sklearn/utils/linear_assignment_.py
For instance if you copy this file into an utils directory import with from utils.linear_assignment_ import linear_assignment

Solution
Use pip to install lap and optionally scipy
Uncomment the import and use the following function
def linear_assignment(cost_matrix):
try:
import lap
_, x, y = lap.lapjv(cost_matrix, extend_cost=True)
return np.array([[y[i], i] for i in x if i >= 0])
except ImportError:
from scipy.optimize import linear_sum_assignment
x, y = linear_sum_assignment(cost_matrix)
return np.array(list(zip(x, y)))

You are getting this error because you haven't install scikit module yet.
Install scikit-learn module from https://pypi.org/project/scikit-learn/

How do I tell which module a function comes from?

I'm trying to read the help on what various things do as I'm reading through code. I'm getting a bit lost in how to determine which module a function comes from. Here is my current example:
import quandl
import numpy as np
import matplotlib.pyplot as plt
amzn = quandl.get("WIKI/AMZN", start_date="2018-01-01", end_date="2019-01-01")
amzn_daily_close = amzn[['Adj. Close']]
amzn_daily_log_returns = np.log(amzn_daily_close.pct_change()+1)
monthly = amzn.resample('BM').apply(lambda x: x[-1])
So given this block of code, I can do help (quandl.get) to see information about that and help (np.log) to see what that does. But when I get to amzn.resample, where is that resample coming from? What should I be entering to see some help information on the resample stuff?

Look at the docstring of quandl.get method to get the help message about the return object. This will contain a statement as returns x-object. Googling about x-object will give you more info on this.
Alternatively, you can do this. To identify what is the object you can do the below.
amzn_type = type(amzn)
This gives the monthly object type. Googling for this type value will give you more insights about that object.Example -
a = 10
print(type(a))
The above code returns <class 'int'> output. Googling about int class in python3 will be helpful.

Inspection
You can 'inspect' the method to find the implementation:
import inspect
print(inspect.getfile(amzn.resample))
# /opt/miniconda/envs/stackoverflow/lib/python3.6/site-packages/pandas/core/generic.py
IDE
Or you can use a good IDE (e.g. PyCharm or IntelliJ) which supports you with some neat features:

Generally, these modules should be documented somewhere. They are usually "packaged" and made available on Python Package Index (pypi). You can search there for your package name and find the quandl page. That may have a link to the projects home page with more documentation.

Weird error in sklearn.cluster.KMeans

I want to use mr. yasaichi's implementation of x-means written in Python for my master's thesis (yasaichi's x-means: https://gist.github.com/yasaichi/254a060eff56a3b3b858) . For the last few weeks there have been no problem and I have been running the algorithm several times on various data sets. Today, however, a weird error popped up:
AttributeError: 'KMeans' object has no attribute 'get_params'.
The error comes from line 75 in the yasaichi's implementation:
labels = range(0, k_means.get_params()["n_clusters"])
Originally I thought it was me who had done some weird changes to the code, but when I re-downloaded the original again it came up with the same error.
Any ideas?

It sounds like the KMeans object you are trying to use doesn't have the method get_params.
I just tested the code at https://gist.github.com/yasaichi/254a060eff56a3b3b858 and it worked for me. So, my best guess is that you are somehow overwriting the KMeans object or that your code is using a cached version of the code that defines the KMeans object.
To verify this, try adding print dir(k_means) before line 75 of yasaichi's implementation. You should also see that print k_means.__module__ should show sklearn.cluster.k_means_. If this is the case, the final thing I would recommend would be deleting the compiled Python file implementing the k_means_ module. This can be found by running the following:
import sklearn.cluster.k_means_
print sklearn.cluster.__file__