Upload folder using huggingface_hub

app.py

@@ -24,11 +24,15 @@ ALL_DB_COLS = BASE_COLS + EXTRA_COLS
 TABLE_COLS = ["id","filename","category","dataset",
               "tags","keywords","notes","uploaded_at"]
 
-# At top-level config
-HF_INDEX_REPO_ID   = "akazemian/audio-library"  # where index.csv lives *now*
-HF_INDEX_REPO_TYPE = "space"                    # <— it's a Space, not a dataset
-INDEX_FILENAME     = "index.csv"
+# ---------- Load HF index ----------
+# ----------- HF DATASET CONFIG -----------
+HF_DATASET_REPO = "akazemian/audio-html"   # dataset holding the HTMLs + index.csv
+INDEX_FILENAME  = "index.csv"
+# -----------------------------------------
 
+# Always read index.csv from the DATASET (not the Space)
+HF_INDEX_REPO_ID   = HF_DATASET_REPO
+HF_INDEX_REPO_TYPE = "dataset"
 
 # at the top
 import os
@@ -122,27 +126,29 @@ def _df_from_table_value(table_value):
             return pd.DataFrame(table_value, columns=cols)
     return pd.DataFrame(columns=cols)
 
-# ---------- Load HF index ----------
+
 def _load_hf_index() -> pd.DataFrame:
     """
-    Download + read index.csv from the HF dataset repo.
+    Download + read index.csv from the HF *dataset* repo.
     Required columns: id, filename, relpath, category, dataset, tags, keywords, notes, uploaded_at
     """
     local = hf_hub_download(
-        repo_id=HF_INDEX_REPO_ID,
-        repo_type=HF_INDEX_REPO_TYPE,
+        repo_id=HF_INDEX_REPO_ID,      # = HF_DATASET_REPO
+        repo_type=HF_INDEX_REPO_TYPE,  # = "dataset"
         filename=INDEX_FILENAME,
     )
+    # Optional: log where we loaded from (shows in Space logs)
+    print(f"[index] loaded from {HF_INDEX_REPO_TYPE}:{HF_INDEX_REPO_ID}/{INDEX_FILENAME} -> {local}")
 
     df = pd.read_csv(local)
     for c in ["id","filename","relpath","category","dataset","tags","keywords","notes","uploaded_at"]:
         if c not in df.columns:
             df[c] = ""
-    # normalize types
     for c in ["id","filename","relpath","category","dataset","tags","keywords","notes","uploaded_at"]:
         df[c] = df[c].fillna("").astype(str)
     return df
 
+
 # ---------- Sync by model (prefix inside HF dataset) ----------
 from urllib.parse import unquote  # ensure this import exists at top