sync ms

.gitignore

@@ -0,0 +1,2 @@
+test.*
+*__pycache__*

README.md

@@ -1,3 +1,30 @@
 ---
 license: cc-by-nc-nd-4.0
+viewer: false
 ---
+
+# Intro
+The SF Soft Soundfont Dataset is a comprehensive sample library designed for intelligent music synthesis and digital audio workstation (DAW) development, featuring a diverse collection of high-quality instrument sounds spanning multiple categories including piano, strings, woodwinds, brass, percussion, electronic synthesizers, and ethnic instruments, while supporting both mainstream soundfont formats: SF2, which utilizes uncompressed PCM storage to ensure maximum fidelity and tonal authenticity, and SF3, which employs Ogg Vorbis compression to achieve reduced file sizes and faster loading times; each soundfont is accompanied by structured metadata encompassing instrument classification, pitch range, sample rate, polyphony count, velocity layering, and loop markers, enabling advanced applications such as automatic arrangement, real-time performance, timbre transfer, and cross-platform sequencer development, making it well-suited for mobile music applications, game audio engines, and AI-driven composition systems to facilitate high-fidelity, low-latency intelligent music generation experiences.
+
+## Usage
+```python
+from datasets import load_dataset
+
+ds = load_dataset(
+    "Genius-Society/soundfonts",
+    split="train",
+    cache_dir="./__pycache__",
+    trust_remote_code=True,
+)
+for item in ds:
+    print(item)
+```
+
+## Maintenance
+```bash
+git clone git@hf.co:datasets/Genius-Society/soundfonts
+cd soundfonts
+```
+
+## Mirror
+<https://www.modelscope.cn/datasets/Genius-Society/soundfonts>

soundfonts.py

@@ -0,0 +1,75 @@
+import os
+import random
+import requests
+import datasets
+from tqdm import tqdm
+
+_DOMAIN = "https://www.modelscope.cn"
+
+_URL = f"{_DOMAIN}/datasets/Genius-Society/{os.path.basename(__file__)[:-3]}"
+
+
+class soundfonts(datasets.GeneratorBasedBuilder):
+    def _info(self):
+        return datasets.DatasetInfo(
+            features=datasets.Features(
+                {
+                    "soundfont": datasets.Value("string"),
+                    "name": datasets.Value("string"),
+                    "format": datasets.Value("string"),
+                }
+            ),
+            supervised_keys=("soundfont", "format"),
+            homepage=_URL,
+            license="mit",
+            version="0.0.1",
+        )
+
+    def _get_files(
+        self,
+        root: str,
+        url=f"{_DOMAIN}/api/v1/datasets/167741/repo/tree",
+        page_size=50,
+    ):
+        try:
+            response = requests.get(
+                url,
+                params={
+                    "Revision": "master",
+                    "Root": root,
+                    "PageNumber": 1,
+                    "PageSize": page_size,
+                },
+            )
+            response.raise_for_status()
+            return response.json()["Data"]["Files"]
+
+        except Exception as e:
+            print(f"{e}, retrying...")
+            return self._get_files(root)
+
+    def _split_generators(self, _):
+        dataset = []
+        files = self._get_files("data")
+        for file in tqdm(files, desc="Parsing soundfonts"):
+            fname, ext = os.path.splitext(file["Name"])
+            dataset.append(
+                {
+                    "soundfont": f"{_URL}/resolve/master/" + file["Path"],
+                    "name": str(fname).strip(),
+                    "format": str(ext).strip()[1:].upper(),
+                }
+            )
+
+        random.shuffle(dataset)
+
+        return [
+            datasets.SplitGenerator(
+                name=datasets.Split.TRAIN,
+                gen_kwargs={"files": dataset},
+            )
+        ]
+
+    def _generate_examples(self, files):
+        for i, path in enumerate(files):
+            yield i, path