Improve the python implementation of some search algorithms

[codefeeder]

Issue Type

Code improvement

Changes

Returning the complete path from the root node to the goal node from existing functions for search algorithms in search.py

Descriptions

The current recursive and iterative algorithms for breadth-first search, depth-first search, etc. only returns the node or None when goal node is found or not found respectively.
I propose to improve the algorithm to return the complete path from the start node to the goal node in case a path is found otherwise return None.

I also propose to modify driver functions accordingly.

@MrDupin may I work on this?

[antmarakis]

This sounds good. What do you mean by 'driver functions' though?

[codefeeder]

By 'driver functions', I am referring to the functions calling the search functions as their code would also need to be modified to accommodate the changes in the return values of the search functions.

[antmarakis]

Hmmm, if the driver function changes are only in the way they get return values, then it's fine. I don't want to modify them that much, since they must be as close to the pseudocode as possible.

[codefeeder]

@MrDupin I have made the changes and as suggested by you lso, I made minimal changes in the test_search.py file in PR #1070
My changes pass all the test cases provided in test_search.py but during final testing the output is expected in the old format and therefore, some test cases failed.
Should I format the output according to the old requirements?

[antmarakis]

What is the old format and what is the new format of the output?

[codefeeder]

The old format("<Node ()>") is created by using the single node returned by the search functions.
But, the newly modified search program is directly returning the list of nodes in the path without any further computation.
But due to the difference in the formats, the above test case failed even though the values were same.

[antmarakis]

Hmm, you can change the format, sure. What I'm not sure about now is the implementation itself. It seems that you are using quite a lot more lines to find the full path. Can we not use Node's parent attribute to backtrack through the path?

[codefeeder]

Yes, sir, we can but that would increase the computation by again backtracking from the current node to the root which according to me can be done at the same time when we are finding a path.

[antmarakis]

The issue is that the code now gets more complicated, which is not desirable. Remember that this project is educational in nature and efficiency comes second. Operations should be done in the simplest and most intuitive manner possible.

[codefeeder]

Sure sir, I got your point. Sir, any suggestions on as to how I should proceed in order to add the functionality without adding any further complexity would be appreciated.

[antmarakis]

I am not that familiar with the search module, but can't we backtrack from node.parent to node.parent?

[dipzzzz]

@codefeeder were you able to improve your algorithm? Is this issue closed?

[dipzzzz]

As per me, create custom test cases or use regex to remove <Node ()> and then compare
Or simplest way is to add in your output