Upload folder using huggingface_hub

.idea/workspace.xml

@@ -59,7 +59,7 @@
       <option name="number" value="Default" />
       <option name="presentableId" value="Default" />
       <updated>1748334702944</updated>
-      <workItem from="1748334704041" duration="2124000" />
+      <workItem from="1748334704041" duration="4123000" />
     </task>
     <servers />
   </component>

README.md

@@ -1,3 +1,20 @@
----
-license: unknown
----
+# Polish Politicians Speech
+
+**Dataset in the public domain (2025)**
+
+## Description
+
+Studio-recorded short utterances by 16 Polish political figures (e.g. “Bartoszewicz Artur”, “Zandberg Adrian”), each reading ~15 sentences. Audio in `.wav` (16 kHz). 
+
+## Files
+
+- `audio/` – per-speaker WAV files  
+- `transcript/pl/test.tsv` – tabular metadata & transcripts  
+
+## Usage
+
+```py
+from datasets import load_dataset
+ds = load_dataset("USERNAME/pl-politicians-speech", "plain_audio")
+print(ds["train"][0])
+# {'path': ..., 'audio': {...}, 'sentence': ..., 'age': ..., 'gender': ...}

pl_debates_test.py

@@ -13,7 +13,7 @@ _CITATION = """
 }
 """
 
-# Replace YOUR-USER/YOUR-REPO with your actual namespace/repo
+# Base URLs on the Hub — replace YOUR-USER/YOUR-REPO accordingly
 _BASE_URL       = "https://huggingface.co/datasets/directtt/pl_debates_test/resolve/main/"
 _TRANSCRIPT_URL = _BASE_URL + "transcript/pl/test.tsv"
 _AUDIO_TAR_URL  = _BASE_URL + "audio_pl.tar"
@@ -38,7 +38,7 @@ class PlDebatesTest(datasets.GeneratorBasedBuilder):
                 "audio":       datasets.Audio(sampling_rate=16_000),
                 "sentence":    datasets.Value("string"),
                 "age":         datasets.Value("int32"),
-                "gender":      datasets.ClassLabel(names=["male", "female", "unknown"]),
+                "gender":      datasets.Value("string"),
                 "speech_type": datasets.Value("string"),
                 "source":      datasets.Value("string"),
             }),
@@ -48,51 +48,94 @@ class PlDebatesTest(datasets.GeneratorBasedBuilder):
         )
 
     def _split_generators(self, dl_manager):
-        # Download & extract both the TSV and the audio tarball
-        downloaded = dl_manager.download_and_extract({
-            "tsv": _TRANSCRIPT_URL,
-            "tar": _AUDIO_TAR_URL,
-        })
-        return [
-            datasets.SplitGenerator(
-                name=datasets.Split.TEST,
-                gen_kwargs={
-                    "tsv_path":   downloaded["tsv"],
-                    "audio_root": downloaded["tar"],  # now contains audio/pl/…
-                },
-            ),
-        ]
+        if dl_manager.is_streaming:
+            # streaming: download the tar and the TSV, but don't extract to disk
+            tsv_path = dl_manager.download(_TRANSCRIPT_URL)
+            tar_path = dl_manager.download(_AUDIO_TAR_URL)
+            archives = [dl_manager.iter_archive(tar_path)]
+            return [
+                datasets.SplitGenerator(
+                    name=datasets.Split.TEST,
+                    gen_kwargs={
+                        "tsv_path":   tsv_path,
+                        "archives":   archives,
+                        "streaming":  True,
+                    },
+                ),
+            ]
+        else:
+            # eager: download & extract everything
+            downloaded = dl_manager.download_and_extract({
+                "tsv": _TRANSCRIPT_URL,
+                "tar": _AUDIO_TAR_URL,
+            })
+            return [
+                datasets.SplitGenerator(
+                    name=datasets.Split.TEST,
+                    gen_kwargs={
+                        "tsv_path":   downloaded["tsv"],
+                        "audio_root": downloaded["tar"],
+                        "streaming":  False,
+                    },
+                ),
+            ]
 
-    def _generate_examples(self, tsv_path, audio_root):
-        # 1) Load the metadata from the TSV
+    def _generate_examples(self, tsv_path, streaming, archives=None, audio_root=None):
+        # 1) load metadata
         meta = {}
         with open(tsv_path, encoding="utf-8") as f:
             reader = csv.DictReader(f, delimiter="\t")
             for row in reader:
                 meta[row["path"]] = row
 
-        # 2) Walk the unpacked audio directory
         key = 0
-        pl_root = os.path.join(audio_root, "audio", "pl")
-        for speaker in sorted(os.listdir(pl_root)):
-            sp_dir = os.path.join(pl_root, speaker)
-            if not os.path.isdir(sp_dir):
-                continue
-            for fname in sorted(os.listdir(sp_dir)):
-                if not fname.endswith(".wav"):
-                    continue
-                rel_path = os.path.join("audio", "pl", speaker, fname)
-                row = meta.get(rel_path)
-                if row is None:
-                    # no transcript entry for this file
+
+        if streaming:
+            # iterate inside the tar archive without extracting
+            for archive in archives:
+                for path_in_tar, fileobj in archive:
+                    # path_in_tar looks like "audio/pl/<speaker>/<fname>.wav"
+                    if not path_in_tar.endswith(".wav"):
+                        continue
+                    # build the rel_path for lookup
+                    rel_path = path_in_tar.replace("\\", "/")  # normalize on Windows
+                    row = meta.get(rel_path)
+                    if row is None:
+                        continue
+                    # read the bytes and yield them
+                    audio_bytes = fileobj.read()
+                    yield key, {
+                        "path":        rel_path,
+                        "audio":       {"path": None, "bytes": audio_bytes},
+                        "sentence":    row["sentence"],
+                        "age":         int(row.get("age", -1)),
+                        "gender":      row["gender"],
+                        "speech_type": row["speech_type"],
+                        "source":      row["source"],
+                    }
+                    key += 1
+
+        else:
+            # eager: walk the extracted folder on disk
+            pl_root = os.path.join(audio_root, "audio", "pl")
+            for speaker in sorted(os.listdir(pl_root)):
+                sp_dir = os.path.join(pl_root, speaker)
+                if not os.path.isdir(sp_dir):
                     continue
-                yield key, {
-                    "path":        rel_path,
-                    "audio":       os.path.join(sp_dir, fname),
-                    "sentence":    row["sentence"],
-                    "age":         int(row.get("age", -1)),
-                    "gender":      row.get("gender", "unknown") or "unknown",
-                    "speech_type": row["speech_type"],
-                    "source":      row["source"],
-                }
-                key += 1
+                for fname in sorted(os.listdir(sp_dir)):
+                    if not fname.endswith(".wav"):
+                        continue
+                    rel_path = os.path.join("audio", "pl", speaker, fname)
+                    row = meta.get(rel_path)
+                    if row is None:
+                        continue
+                    yield key, {
+                        "path":        rel_path,
+                        "audio":       os.path.join(sp_dir, fname),
+                        "sentence":    row["sentence"],
+                        "age":         int(row.get("age", -1)),
+                        "gender":      row.get("gender", "unknown") or "unknown",
+                        "speech_type": row["speech_type"],
+                        "source":      row["source"],
+                    }
+                    key += 1