import json
import os
from typing import List

import datasets
from datasets import Features, Value, Image, ClassLabel


_CITATION = """\
@dataset{african_medical_multimodal_fracture_2024,
  title={African Medical Multimodal Bone Fracture Dataset},
  author={Electric Sheep Africa},
  year={2024},
  publisher={Hugging Face},
  url={https://huggingface.co/datasets/electricsheepafrica/african-medical-multimodal-fracture}
}
"""

_DESCRIPTION = """\
A comprehensive, multimodal bone break classification dataset specifically designed for African healthcare contexts. 
This dataset addresses critical gaps in medical AI for resource-constrained environments while ensuring cultural 
sensitivity and local relevance.

The dataset includes 1,129 multimodal medical cases with X-ray images, clinical measurements, patient demographics, 
and contextual information across 18 African countries and 11 African languages.
"""

_HOMEPAGE = "https://huggingface.co/datasets/electricsheepafrica/african-medical-multimodal-fracture"

_LICENSE = "CC BY 4.0"

_URLS = {
    "train": "data/train.jsonl",
    "validation": "data/validation.jsonl", 
    "test": "data/test.jsonl",
}


class AfricanMedicalMultimodalFractureDataset(datasets.GeneratorBasedBuilder):
    """African Medical Multimodal Bone Fracture Dataset"""

    VERSION = datasets.Version("1.0.0")

    def _info(self):
        features = Features({
            "record_id": Value("string"),
            "patient_id": Value("string"),
            "image": Image(),
            "fracture_type": ClassLabel(names=[
                "Avulsion fracture", "Comminuted fracture", "Fracture Dislocation",
                "Greenstick fracture", "Hairline Fracture", "Impacted fracture",
                "Longitudinal fracture", "Oblique fracture", "Pathological fracture",
                "Spiral fracture"
            ]),
            "age": Value("int32"),
            "gender": Value("string"),
            "country": Value("string"),
            "region_type": Value("string"),
            "primary_language": Value("string"),
            "chief_complaint": Value("string"),
            "complaint_language": Value("string"),
            "heart_rate": Value("int32"),
            "blood_pressure_systolic": Value("int32"),
            "blood_pressure_diastolic": Value("int32"),
            "temperature": Value("float32"),
            "height": Value("float32"),
            "weight": Value("float32"),
            "pain_scale": Value("int32"),
            "season": Value("string"),
            "phone_type": Value("string"),
            "digital_literacy": Value("string"),
        })

        return datasets.DatasetInfo(
            description=_DESCRIPTION,
            features=features,
            homepage=_HOMEPAGE,
            license=_LICENSE,
            citation=_CITATION,
        )

    def _split_generators(self, dl_manager):
        urls = _URLS
        data_files = dl_manager.download_and_extract(urls)
        
        return [
            datasets.SplitGenerator(
                name=datasets.Split.TRAIN,
                gen_kwargs={
                    "filepath": data_files["train"],
                    "split": "train",
                    "dl_manager": dl_manager,
                },
            ),
            datasets.SplitGenerator(
                name=datasets.Split.VALIDATION,
                gen_kwargs={
                    "filepath": data_files["validation"],
                    "split": "validation",
                    "dl_manager": dl_manager,
                },
            ),
            datasets.SplitGenerator(
                name=datasets.Split.TEST,
                gen_kwargs={
                    "filepath": data_files["test"],
                    "split": "test",
                    "dl_manager": dl_manager,
                },
            ),
        ]

    def _generate_examples(self, filepath, split, dl_manager):
        """Generate examples from the JSONL file."""
        
        with open(filepath, "r", encoding="utf-8") as f:
            for key, line in enumerate(f):
                data = json.loads(line)
                
                # Build a fully resolvable URL for the image and download/stream it via dl_manager
                image_rel_path = data["image"]

                # Determine repository base path (strip the last two segments e.g., "data/train.jsonl")
                base_path_parts = filepath.split("/")[:-2]
                base_path = "/".join(base_path_parts)
                image_url = f"{base_path}/{image_rel_path}"

                # dl_manager will return a local cached path or a streaming URL depending on context
                downloaded_image_path = dl_manager.download(image_url)

                data["image"] = downloaded_image_path

                yield key, data