update in index

mir-dataset-loaders · Nov 7, 2024 · 809d3a9 · 809d3a9
1 parent d264549
commit 809d3a9
Show file tree

Hide file tree

Showing 3 changed files with 33 additions and 29 deletions.
diff --git a/docs/source/table.rst b/docs/source/table.rst
@@ -64,11 +64,11 @@
           :target: http://creativecommons.org/publicdomain/zero/1.0/
 
    * - BRID
-     - - audio: ❌
-       - annotations: ❌
+     - - audio: ✅
+       - annotations: ✅
      - - :ref:`beats`
        - :ref:`tempo`
-     - 35
+     - 367
      - .. image:: https://licensebuttons.net/l/by-nc-sa/4.0/80x15.png
           :target: https://creativecommons.org/licenses/by-nc-sa/4.0
 

diff --git a/mirdata/datasets/brid.py b/mirdata/datasets/brid.py
@@ -7,7 +7,7 @@
 
     **Dataset Overview:**
 
-    BRID comprises a total of 35 tracks, averaging about 30 seconds each, amounting to approximately 2 hours and 57 minutes of music. These tracks include recordings of various Brazilian instruments, played in different Brazilian rhythmic styles.
+    BRID comprises a total of 367 tracks, averaging about 30 seconds each, amounting to approximately 2 hours and 57 minutes of music. These tracks include recordings of various Brazilian instruments, played in different Brazilian rhythmic styles.
 
     **Instruments and Rhythms:**
 
@@ -41,7 +41,7 @@
 import numpy as np
 from typing import BinaryIO, Optional, TextIO, Tuple
 
-from mirdata import annotations, core, io, jams_utils
+from mirdata import annotations, core, download_utils, io, jams_utils
 
 
 BIBTEX = """
@@ -58,30 +58,33 @@
     "test": "sample",
     "1.0": core.Index(
         filename="brid_full_index_1.0.json",
-        url="https://zenodo.org/records/14048016/files/brid_full_index_1.0.json?download=1",
-        checksum="18701a7b6c20e549c0394bbd80c01699",
+        url="https://zenodo.org/records/14052434/files/brid_full_index_1.0.json?download=1",
+        checksum="6292a6d36d6ae267534107f4e5f6bcca",
     ),
     "sample": core.Index(filename="brid_full_index_1.0_sample.json"),
 }
 
 
-REMOTES = None
+REMOTES = {
+    "annotations": download_utils.RemoteFileMetadata(
+        filename="annotations.zip",
+        url="https://zenodo.org/records/14051323/files/annotations.zip?download=1",
+        checksum="678b2fa99c8d220cddd9f5e20d55d0c1",
+        destination_dir="BRID_1.0",
+    ),
+    "audio": download_utils.RemoteFileMetadata(
+        filename="audio.zip",
+        url="https://zenodo.org/records/14051323/files/audio.zip?download=1",
+        checksum="3514b53d66515181f95619adb71a59b4",
+        destination_dir="BRID_1.0",
+    ),
+}
+
 
 LICENSE_INFO = (
     "Creative Commons Attribution Non Commercial Share Alike 4.0 International."
 )
 
-DOWNLOAD_INFO = """
-    Unfortunately the BRID dataset is not available for download.
-    If you have the BRID dataset, place the contents into a folder called
-    BRID with the following structure:
-        > BRID_1.0/
-            > audio/
-            > annotations/beats
-            > annotations/tempo
-    and copy the BRID folder to {}
-    """
-
 
 class Track(core.Track):
     """BRID Rhythm class
@@ -230,5 +233,4 @@ def __init__(self, data_home=None, version="default"):
             indexes=INDEXES,
             remotes=REMOTES,
             license_info=LICENSE_INFO,
-            download_info=DOWNLOAD_INFO,
         )
diff --git a/scripts/make_brid_index.py b/scripts/make_brid_index.py
@@ -29,26 +29,28 @@ def make_brid_rhythm_index(dataset_data_path):
                 beat_file = f"{idx}.beats"
                 tempo_file = f"{idx}.bpm"
 
-                relative_beat_path = os.path.join("Annotations", "beats", beat_file)
-                relative_tempo_path = os.path.join("Annotations", "tempo", tempo_file)
+                # Set relative paths to None if files do not exist
+                relative_beat_path = os.path.join("Annotations", "beats", beat_file) if os.path.exists(os.path.join(beats_path, beat_file)) else None
+                relative_tempo_path = os.path.join("Annotations", "tempo", tempo_file) if os.path.exists(os.path.join(tempo_path, tempo_file)) else None
+
 
                 # Check if annotation files exist
-                beat_file_path = os.path.join(beats_path, beat_file)
-                tempo_file_path = os.path.join(tempo_path, tempo_file)
-                
+                beat_file_path = os.path.join(beats_path, beat_file) if relative_beat_path else None
+                tempo_file_path = os.path.join(tempo_path, tempo_file) if relative_tempo_path else None
+
                 # Add track information to index
                 cmr_index["tracks"][idx] = {
                     "audio": (
                         os.path.join(dataset_folder_name, relative_audio_path),
                         md5(os.path.join(root, filename)),
                     ),
                     "beats": (
-                        os.path.join(dataset_folder_name, relative_beat_path),
-                        md5(beat_file_path) if os.path.exists(beat_file_path) else None,
+                        os.path.join(dataset_folder_name, relative_beat_path) if relative_beat_path else None,
+                        md5(beat_file_path) if beat_file_path else None,
                     ),
                     "tempo": (
-                        os.path.join(dataset_folder_name, relative_tempo_path),
-                        md5(tempo_file_path) if os.path.exists(tempo_file_path) else None,
+                        os.path.join(dataset_folder_name, relative_tempo_path) if relative_tempo_path else None,
+                        md5(tempo_file_path) if tempo_file_path else None,
                     )
                 }