Closes #305 | Create Multilingual open relation dataloader

[madenindya]

Closes #305

Checkbox

• Confirm that this PR is linked to the dataset issue.
• Create the dataloader script nusantara/nusa_datasets/my_dataset/my_dataset.py (please use only lowercase and underscore for dataset naming).
• Provide values for the _CITATION, _DATASETNAME, _DESCRIPTION, _HOMEPAGE, _LICENSE, _URLs, _SUPPORTED_TASKS, _SOURCE_VERSION, and _NUSANTARA_VERSION variables.
• Implement _info(), _split_generators() and _generate_examples() in dataloader script.
• Make sure that the BUILDER_CONFIGS class attribute is a list with at least one NusantaraConfig for the source schema and one for a nusantara schema.
• Confirm dataloader script works with datasets.load_dataset function.
• Confirm that your dataloader script passes the test suite run with python -m tests.test_nusantara --path=nusantara/nusa_datasets/my_dataset/my_dataset.py.
• If my dataset is local, I have provided an output of the unit-tests in the PR (please copy paste). This is OPTIONAL for public datasets, as we can test these without access to the data files.

[madenindya]

Hi, I found many difficulties with this dataset and would like to have some advice:

1. The data source is from Kaggle, how do I download it here? The original .zip size is huge. In local, I only download Indonesian data and use it for test reference.
2. When creating entities offset, I found it hard to get as the original data didn't show that. When I tried to make it on my own, it was also hard to infer as there are some exact words in a sentence.
3. And because I couldn't get the offset, I don't know whether it's better to have a different EntID for the same word in the same sentence. As it might come from the same/different position word, but I don't have that info.

I mark my line of codes as TODO for the things that need further discussion

Additional Questions:

• Why the entites.text is List & not just a single string? (context: schema's KB)

[muhsatrio]

Hi, I found many difficulties with this dataset and would like to have some advice:

1. The data source is from Kaggle, how do I download it here? The original .zip size is huge. In local, I only download Indonesian data and use it for test reference.
2. When creating entities offset, I found it hard to get as the original data didn't show that. When I tried to make it on my own, it was also hard to infer as there are some exact words in a sentence.
3. And because I couldn't get the offset, I don't know whether it's better to have a different EntID for the same word in the same sentence. As it might come from the same/different position word, but I don't have that info.

I mark my line of codes as TODO for the things that need further discussion

Additional Questions:

• Why the entites.text is List & not just a single string? (context: schema's KB)

Same like bottleneck that I faced.

cc: @SamuelCahyawijaya @holylovenia @bryanwilie

[muhsatrio]

I think you can continue discussing it in slack for the faster response kak @madenindya, thank you!

[SamuelCahyawijaya]

@muhsatrio : Sorry I missed this PR. Has this one been finalized? I can check it right away

[muhsatrio]

@muhsatrio : Sorry I missed this PR. Has this one been finalized? I can check it right away

I think still there is no any changes kak

[SamuelCahyawijaya]

IMO, for this dataset we can just implement the source schema for now, as it will be complicated to extend it to the KB schema. What do you think? @madenindya @muhsatrio