harsh19/chesscommentarygeneration

Harsh Jhamtani*, Varun Gangal*, Eduard Hovy, Graham Neubig, Taylor Berg-Kirkpatrick. Learning to Generate Move-by-Move Commentary for Chess Games from Large-Scale Social Forum Data. ACL 2018

ChessCommentaryGeneration

Harsh Jhamtani*, Varun Gangal*, Eduard Hovy, Graham Neubig, Taylor Berg-Kirkpatrick. Learning to Generate Move-by-Move Commentary for Chess Games from Large-Scale Social Forum Data. ACL 2018

Dataset

Code for crawling and preprocessing data: Data/crawler/

Code

Code/

Requirements

Python 2.7
PyTorch 0.3 (along with torchvision)
PyQt4

Reference

If you use the data or code, please consider citing

@inproceedings{jhamtani2018acl,
    title = {Learning to Generate Move-by-Move Commentary for Chess Games from Large-Scale Social Forum Data},
    author = {Harsh Jhamtani and Varun Gangal and Eduard Hovy and Graham Neubig and Taylor Berg-Kirkpatrick},
    booktitle = {The 56th Annual Meeting of the Association for Computational Linguistics (ACL)},
    address = {Melbourne, Australia},
    month = {July},
    year = {2018}
}

chesscommentarygeneration's Issues

Can you provide processed dataset to me?

I follow the /Data/crawler/README to run the code, but cannot run the code in /Code/chess_features/.
Can you provide processed dataset to me? My e-mail is: [email protected]
Thank you very much!

Code Help

Hi,

Thank you for this amazing project.

I have currently extracted all my data and split into the .en files etc. for validation test and training. I now want to split it into the different categories but I'm unsure of how to work your code for the SVM. Will you be updating this?

I am using the commentary as part of a dataset for a sentiment analysis research project, meaning we only want to focus on quality moves. Thus, being able to retrieve the quality moves would be great.

Thanks,

`ImportError: No module named PyQt4.QtGui`

It seems that this package is using an outdated package (I don't believe pyqt is maintained anymore)

Would it be possible to get the data from you @harsh19 ? I'd like to experiment with it!

ValueError: too many values to unpack for "preprocess.py"

When I run python preprocess.py train, it says:

Traceback (most recent call last):
  File "preprocess.py", line 103, in <module>
    for i,link in all_links:
ValueError: too many values to unpack

Do you mean: for i,link in enumerate(all_links): actually? Or if I got some mistakes when I downloaded the dataset?

Thank you very much.

annotation2.tsv, rules.txt

Dear authors,

I am a student in my last year of my Bachelors Degree in Computer Science, and I would like to attempt building an ai with a similar purpose to that described in this paper. While looking through papers, I have also noticed https://github.com/zhyack/SCC, which seems to make use of this dataset, and there the README.md instructs us to request access to "Ask the permission from Jhamtani et.al for the distribution of annotation2.tsv, rules.txt, and their processed dataset with [train/test/valid]_[0/1/2].en for data pre-processing.".

Would it be possible for me to obtain these files?

Best wishes,
George

(cc: @harsh19)

HF dataset

Dear authors, have you considered publishing the dataset as a HF dataset for all to use and research?

Trouble running the code with python 3.5

Hi Harsh, first of all, amazing work there, I am interested in running this for a side project of my own, and am having trouble converting the code for python 3.5 compatible as PyQt4 is not available for python 3 and PyQt5 has done a lot of changes with the classes and method, I would be grateful if you can mail me the processed datasets, and would be happy to contribute to making it 3.5 compatible later on.
email : [email protected]