rodrigodlpontes / visualization-tool Goto Github PK

As seen here, the Rabin-Karp algorithm erroneously skips over matches when the base constant is set to a number other than one. See the image for an example (a match occurs in the last sequence but is skipped due to differing obtained hashes):

As seen in the side view, the last hash calculation appears to be incorrect. Assuming a = 0, b = 1, c = 2 when base = 17, one can easily verify that cba has a hash of 2 * 289 + 1 * 17 + 0 * 1 = 595, which is dramatically different from either 27763 or 30083 as seen in the picture.

Interestingly, the algorithm works properly when a base constant of 1 is set.

It would be great if we could pick our own pivot for each iteration of QuickSort and QuickSelect, so we can visualize it better.

Doesn't resize after reaching maximum capacity - sure implementation is similar to the Queue Array one

When inserting values in sequence into the 2-4 tree: 52, 14, 34, 76, 3. The 3 ends up in the wrong place. It should be before the 14.

To reproduce:

1. Insert an element that hashes to index 0.
2. Delete this Element.

It will not be found and thus not be deleted.

Fix: In ClosedHash.js, in deleteElement(), replace if (index > 0) by if (index >= 0).

To reproduce (there are probably simpler ways):

1. set CLOSED_HASH_TABLE_SIZE = 7
2. activate Hash Strings
3. insert A
4. insert H (same hash code as A)
5. insert O (same hash code as A)
6. delete A
7. delete H
8. insert H

At this point, getEmptyIndex() enters into an endless loop.

Moveover, I observed confusions between empty and deleted cells in the explanations given by ExplainLabel, also in this method.

Using "Skip Back" button in Array-backed queues messes stuff up. Reproducible by enqueueing [0,1,2], skip back * 2, then enqueueing 1:

When BuildHeap Max-Heap is selected for Heaps/PQs, only some swaps are made correctly. Seems that sometimes it is working as a max heap and other times as a min heap. Has happened with more than one input list.

What license do the source code here and the resulting visualizations fall under? Would it be compatible with CC BY-SA 3.0? I'm wondering if it would be okay to reuse the visualizations on Wikipedia. Apparently the David Galles visualization tool code is licensed under FreeBSD.

Please delete auto-generated files from git and .gitignore them; they unnecessarily give rise to merge conflicts.

Setup: Add 1, add 30, delete 1

Reproduce: Add 30

Expect: Adding 30 should check the delete flag, store it, then continue to the next index and see the duplicate 30

Actual: 30 gets added as soon as it sees the delete flag

After enqueuing elements in the QueueLL algo, selecting clear crosses out the head as null but not the tail.

see title

For floyd warshall, letting negative edge weights be present in the randomization with an option for negative cycles would be nice (and smaller graphs cause v^3).

Traversals should stay on screen after the traversal is performed

This breaks on all sorting algorithms, not just quicksort

An key feature of the KMP algorithm is the consultation of f[k] on encountering a mismatch following one or more matches, where f is the failure table and k is the last-matched index into the pattern. This is not currently visualized, and the role of f[k] in advancing the pattern is not obvious.

I implemented such a visualization in #110.

First of all, thanks for a great tool for visualizing data structures. It is very helpful to solidify the concepts.

When creating a Binary Search Tree with data 15,2,26,1,9,19,30,7,12,17,20,22, performing a Levelorder Traversal correctly visits the nodes. However, when the entire traversal is completed, the sequence of visited nodes displayed in the bottom left corner of the canvas is not retained.

The other traversal types do retain the returned sequence correctly so it can be viewed after completion.

I implemented Lomuto's partition scheme for my own needs, hard-coding the pivot to the last element, because it is the simplest to write down, explain, and analyze. Currently it only lives on my local system in a separate helperLomuto() function.

Let me know if you're interested; if not, just close this feature request :-).

Hello, I'm having issues skipping back using the LSD radix sort. I tried to add the data: 100, 10, 1. I paused the animation, but I couldn't step back to the beginning. The linked lists did not appear when stepping back. To reproduce this issue, you may add whatever data you please. Once you allow the first iteration of LSD to complete and the data from the buckets are placed back into the array, you are unable to put the data back into the buckets when stepping back. I thank all the TAs who worked on the visualization tool. It has been beneficial to my peers and me.

To reproduce:

1. Insert two elements that hash to the same entry.
2. Delete the first.
3. Find the second.

It is not found.

Fix: In ClosedHash.js, in getElemIndex(index, elem), remove || this.deleted[candidateIndex] from the else if condition.