destaq / chess-graph Goto Github PK

When you input a custom way to analyze the data (from a certain position, that is) - the first two positions show up as well.

E.g. if we want to analyze the data from d4 d5, then we will also see d4 d5, but they will be rings that are 100% big.

First of all GREAT WORK ! Amazing tool ;)

I kindly ask you the possibility to print (at least on terminal or on a .txt file) the best (W/B ratio) and worst line for each level.
Something like:

I'm available to help you doing this (just send me an email)

tested on google colab

pip install -U chess-graph 
import chess_graph

result

---------------------------------------------------------------------------
ModuleNotFoundError                       Traceback (most recent call last)
[<ipython-input-4-af3db3b3152d>](https://localhost:8080/#) in <cell line: 1>()
----> 1 import chess_graph

2 frames
[/usr/local/lib/python3.10/dist-packages/chess_graph/new_parser.py](https://localhost:8080/#) in <module>
----> 1 import pgn
      2 
      3 
      4 def parse_games(database, depth, custom_branching, color, name):
      5     database = open(database)

ModuleNotFoundError: No module named 'pgn'

---------------------------------------------------------------------------
NOTE: If your import is failing due to a missing package, you can
manually install dependencies using either !pip or !apt.

To view examples of installing some common dependencies, click the
"Open Examples" button below.
---------------------------------------------------------------------------

The outermost ring is lighter then the rest. Need to make it the same color to prevent confusion and aesthetic interference.

Currently, this tool can analyze games up to any depth. However, the time it takes to do so is exponentially correlated to depth. This is because it is traversing the PGN using brute force.

To put it into perspective, a depth of 5 takes 30 seconds, a depth of 8 about two minutes.

This is low-priority, as it is unlikely people will want to make graphs to such depths, but if there is any way to improve the execution of the code, that would be great.

With the new updates to the code, you can now edit the file and remove any slices that are too small, to prevent the fragmentation look.

This works perfectly well - except for the first level!. On the innermost circle, even if you have a move that is played only once - e.g. a4, with a percentage of 0.1% let's say - it still shows up!

Here is the code for those interested:
`del_list = []
for i in range(len(percentage_holder)):
if percentage_holder[i] <= 5 and parents[i] != '':
del_list.append(i)

for i in range(len(parents)): # loop through all parents
    if i in del_list: # if that parent is in the bin
        death = ids[i] # find that parent's ID so we can locate children
        for x in range(len(parents)): # loop through parents again to find children
            if parents[x] == death: # if child's parent is equal to death
                del_list.append(x) # mark child for death as well


percentage_holder = [percentage_holder[i] for i in range(len(percentage_holder)) if i not in del_list]
ids = [ids[i] for i in range(len(ids)) if i not in del_list]
labels = [labels[i] for i in range(len(labels)) if i not in del_list]
parents = [parents[i] for i in range(len(parents)) if i not in del_list]
values = [values[i] for i in range(len(values)) if i not in del_list]`

Like in the example that you reported from this site: https://blog.ebemunk.com/a-visual-look-at-2-million-chess-games/

you can see that odd/even moves have a darker/lighter background

Is there a command line argument for the code to ignore the clock data in the pgn? From what I understand, PGN's can have clock data. So when I try it with clock data, the graph looks very strange: