fixed errors

app.py

@@ -1,5 +1,5 @@
 # =========================
-# 1. DOWNLOAD BOTH DATASETS
+# 1. DOWNLOAD + SELECTIVE EXTRACT
 # =========================
 import os, requests, zipfile
 
@@ -10,24 +10,65 @@ urls = [
 
 os.makedirs("data", exist_ok=True)
 
+# =========================
+# 2. LOAD LABELS FIRST
+# =========================
+import pandas as pd
+
+def load_split(csv_file):
+    df = pd.read_csv(csv_file)
+    df.columns = df.columns.str.strip()
+
+    split = {}
+    for _, row in df.iterrows():
+        try:
+            pid = str(int(row["Participant_ID"]))
+            split[pid] = int(row["PHQ8_Binary"])
+        except:
+            continue
+    return split
+
+train_labels = load_split("train_split_Depression_AVEC2017.csv")
+dev_labels = load_split("dev_split_Depression_AVEC2017.csv")
+
+ALL_REQUIRED_IDS = set(list(train_labels.keys()) + list(dev_labels.keys()))
+
+print("Total required participants:", len(ALL_REQUIRED_IDS))
+
+# =========================
+# 3. SELECTIVE EXTRACTION
+# =========================
+def extract_needed(zip_path):
+    with zipfile.ZipFile(zip_path, "r") as zip_ref:
+        for file in zip_ref.namelist():
+
+            # extract only participant folders we need
+            for pid in ALL_REQUIRED_IDS:
+                if f"{pid}_" in file:
+                    zip_ref.extract(file, "data")
+                    break
+
+# =========================
+# 4. DOWNLOAD + EXTRACT
+# =========================
 for i, url in enumerate(urls):
-    zip_path = f"data_{i}.zip"
+    zip_path = f"temp_{i}.zip"
+
+    print(f"Downloading dataset {i+1}...")
+    r = requests.get(url, stream=True)
+    with open(zip_path, "wb") as f:
+        for chunk in r.iter_content(8192):
+            f.write(chunk)
 
-    if not os.path.exists(zip_path):
-        print(f"Downloading dataset {i+1}...")
-        r = requests.get(url, stream=True)
-        with open(zip_path, "wb") as f:
-            for chunk in r.iter_content(8192):
-                f.write(chunk)
+    print(f"Extracting required files from dataset {i+1}...")
+    extract_needed(zip_path)
 
-        print(f"Extracting dataset {i+1}...")
-        with zipfile.ZipFile(zip_path, "r") as zip_ref:
-            zip_ref.extractall(f"data/set_{i}")
+    os.remove(zip_path)  # 🔥 VERY IMPORTANT
 
 # =========================
-# 2. GET ALL PARTICIPANTS
+# 5. GET PARTICIPANTS
 # =========================
-def get_all_participant_paths():
+def get_all_paths():
     paths = []
     for root, dirs, _ in os.walk("data"):
         for d in dirs:
@@ -35,14 +76,13 @@ def get_all_participant_paths():
                 paths.append(os.path.join(root, d))
     return paths
 
-ALL_PATHS = get_all_participant_paths()
-print("Total participant folders:", len(ALL_PATHS))
+ALL_PATHS = get_all_paths()
+print("Extracted participant folders:", len(ALL_PATHS))
 
 # =========================
-# 3. IMPORT LIBRARIES
+# 6. IMPORT LIBRARIES
 # =========================
 import numpy as np
-import pandas as pd
 import librosa
 import torch
 import torch.nn as nn
@@ -53,29 +93,7 @@ from sklearn.metrics import accuracy_score, f1_score
 from sklearn.preprocessing import StandardScaler
 
 # =========================
-# 4. LOAD SPLITS
-# =========================
-def load_split(csv_file):
-    df = pd.read_csv(csv_file)
-    df.columns = df.columns.str.strip()
-
-    split = {}
-    for _, row in df.iterrows():
-        try:
-            pid = str(int(row["Participant_ID"]))
-            split[pid] = int(row["PHQ8_Binary"])
-        except:
-            continue
-    return split
-
-train_labels = load_split("train_split_Depression_AVEC2017.csv")
-dev_labels = load_split("dev_split_Depression_AVEC2017.csv")
-
-print("Train labels:", len(train_labels))
-print("Dev labels:", len(dev_labels))
-
-# =========================
-# 5. LOAD TEXT MODEL
+# 7. TEXT MODEL
 # =========================
 device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
 
@@ -84,7 +102,7 @@ bert = AutoModel.from_pretrained("distilbert-base-uncased").to(device)
 bert.eval()
 
 # =========================
-# 6. FEATURE FUNCTIONS
+# 8. FEATURE FUNCTIONS
 # =========================
 def load_text(folder):
     try:
@@ -132,10 +150,10 @@ def get_visual_features(folder):
     return np.concatenate(feats)
 
 # =========================
-# 7. BUILD DATASETS
+# 9. BUILD DATA
 # =========================
-def build_dataset(label_dict):
-    X_t, X_a, X_v, y = [], [], [], []
+def build(label_dict):
+    Xt, Xa, Xv, y = [], [], [], []
 
     for path in tqdm(ALL_PATHS):
         pid = os.path.basename(path).split("_")[0]
@@ -143,27 +161,21 @@ def build_dataset(label_dict):
         if pid not in label_dict:
             continue
 
-        X_t.append(get_text_embedding(load_text(path)))
-        X_a.append(get_audio_features(path))
-        X_v.append(get_visual_features(path))
+        Xt.append(get_text_embedding(load_text(path)))
+        Xa.append(get_audio_features(path))
+        Xv.append(get_visual_features(path))
         y.append(label_dict[pid])
 
-    return np.array(X_t), np.array(X_a), np.array(X_v), np.array(y)
-
-print("\nBuilding train...")
-Xt, Xa, Xv, y_train = build_dataset(train_labels)
+    return np.array(Xt), np.array(Xa), np.array(Xv), np.array(y)
 
-print("Building dev...")
-Xt_test, Xa_test, Xv_test, y_test = build_dataset(dev_labels)
+Xt, Xa, Xv, y_train = build(train_labels)
+Xt_test, Xa_test, Xv_test, y_test = build(dev_labels)
 
 print("Train size:", len(y_train))
 print("Test size:", len(y_test))
 
-if len(y_train) == 0 or len(y_test) == 0:
-    raise Exception("Dataset still empty → check upload")
-
 # =========================
-# 8. NORMALIZE
+# 10. NORMALIZE
 # =========================
 sc_t, sc_a, sc_v = StandardScaler(), StandardScaler(), StandardScaler()
 
@@ -176,16 +188,14 @@ Xa_test = sc_a.transform(Xa_test)
 Xv_test = sc_v.transform(Xv_test)
 
 # =========================
-# 9. MODEL
+# 11. MODEL
 # =========================
 class Model(nn.Module):
     def __init__(self, vdim):
         super().__init__()
-
         self.t = nn.Sequential(nn.Linear(768,128), nn.ReLU())
         self.a = nn.Sequential(nn.Linear(40,32), nn.ReLU())
         self.v = nn.Sequential(nn.Linear(vdim,64), nn.ReLU())
-
         self.f = nn.Sequential(
             nn.Linear(224,64),
             nn.ReLU(),
@@ -194,10 +204,7 @@ class Model(nn.Module):
         )
 
     def forward(self, t,a,v):
-        t = self.t(t)
-        a = self.a(a)
-        v = self.v(v)
-        return self.f(torch.cat([t,a,v],1))
+        return self.f(torch.cat([self.t(t), self.a(a), self.v(v)],1))
 
 model = Model(Xv.shape[1]).to(device)
 
@@ -214,30 +221,22 @@ Xa_test = torch.tensor(Xa_test, dtype=torch.float32).to(device)
 Xv_test = torch.tensor(Xv_test, dtype=torch.float32).to(device)
 
 # =========================
-# 10. TRAIN
+# 12. TRAIN
 # =========================
-print("\nTraining...")
-
 for e in range(10):
     model.train()
     opt.zero_grad()
-
-    out = model(Xt,Xa,Xv).squeeze()
-    loss = loss_fn(out, yt)
-
+    loss = loss_fn(model(Xt,Xa,Xv).squeeze(), yt)
     loss.backward()
     opt.step()
-
     print(f"Epoch {e+1}: {loss.item():.4f}")
 
 # =========================
-# 11. EVAL
+# 13. EVAL
 # =========================
 model.eval()
-
 with torch.no_grad():
-    out = model(Xt_test,Xa_test,Xv_test).squeeze()
-    pred = (torch.sigmoid(out)>0.5).int().cpu().numpy()
+    pred = (torch.sigmoid(model(Xt_test,Xa_test,Xv_test).squeeze())>0.5).int().cpu().numpy()
 
 print("\nRESULTS")
 print("Accuracy:", accuracy_score(y_test, pred))