merge try

app.py

@@ -29,63 +29,48 @@ def zero_padding(y: np.ndarray, end: int):
 
 
 def audio2mel(audio_path: str, seg_len=20):
-    os.makedirs(TEMP_DIR, exist_ok=True)
-    try:
-        y, sr = librosa.load(audio_path, sr=SAMPLE_RATE)
-        y = zero_padding(y, seg_len * sr)
-        mel_spec = librosa.feature.melspectrogram(y=y, sr=sr)
-        log_mel_spec = librosa.power_to_db(mel_spec, ref=np.max)
-        librosa.display.specshow(log_mel_spec)
-        plt.axis("off")
-        plt.savefig(
-            f"{TEMP_DIR}/output.jpg",
-            bbox_inches="tight",
-            pad_inches=0.0,
-        )
-        plt.close()
-
-    except Exception as e:
-        print(f"Error converting {audio_path} : {e}")
+    y, sr = librosa.load(audio_path, sr=SAMPLE_RATE)
+    y = zero_padding(y, seg_len * sr)
+    mel_spec = librosa.feature.melspectrogram(y=y, sr=sr)
+    log_mel_spec = librosa.power_to_db(mel_spec, ref=np.max)
+    librosa.display.specshow(log_mel_spec)
+    plt.axis("off")
+    plt.savefig(
+        f"{TEMP_DIR}/output.jpg",
+        bbox_inches="tight",
+        pad_inches=0.0,
+    )
+    plt.close()
 
 
 def audio2cqt(audio_path: str, seg_len=20):
-    os.makedirs(TEMP_DIR, exist_ok=True)
-    try:
-        y, sr = librosa.load(audio_path, sr=SAMPLE_RATE)
-        y = zero_padding(y, seg_len * sr)
-        cqt_spec = librosa.cqt(y=y, sr=sr)
-        log_cqt_spec = librosa.power_to_db(np.abs(cqt_spec) ** 2, ref=np.max)
-        librosa.display.specshow(log_cqt_spec)
-        plt.axis("off")
-        plt.savefig(
-            f"{TEMP_DIR}/output.jpg",
-            bbox_inches="tight",
-            pad_inches=0.0,
-        )
-        plt.close()
-
-    except Exception as e:
-        print(f"Error converting {audio_path} : {e}")
+    y, sr = librosa.load(audio_path, sr=SAMPLE_RATE)
+    y = zero_padding(y, seg_len * sr)
+    cqt_spec = librosa.cqt(y=y, sr=sr)
+    log_cqt_spec = librosa.power_to_db(np.abs(cqt_spec) ** 2, ref=np.max)
+    librosa.display.specshow(log_cqt_spec)
+    plt.axis("off")
+    plt.savefig(
+        f"{TEMP_DIR}/output.jpg",
+        bbox_inches="tight",
+        pad_inches=0.0,
+    )
+    plt.close()
 
 
 def audio2chroma(audio_path: str, seg_len=20):
-    os.makedirs(TEMP_DIR, exist_ok=True)
-    try:
-        y, sr = librosa.load(audio_path, sr=SAMPLE_RATE)
-        y = zero_padding(y, seg_len * sr)
-        chroma_spec = librosa.feature.chroma_stft(y=y, sr=sr)
-        log_chroma_spec = librosa.power_to_db(np.abs(chroma_spec) ** 2, ref=np.max)
-        librosa.display.specshow(log_chroma_spec)
-        plt.axis("off")
-        plt.savefig(
-            f"{TEMP_DIR}/output.jpg",
-            bbox_inches="tight",
-            pad_inches=0.0,
-        )
-        plt.close()
-
-    except Exception as e:
-        print(f"Error converting {audio_path} : {e}")
+    y, sr = librosa.load(audio_path, sr=SAMPLE_RATE)
+    y = zero_padding(y, seg_len * sr)
+    chroma_spec = librosa.feature.chroma_stft(y=y, sr=sr)
+    log_chroma_spec = librosa.power_to_db(np.abs(chroma_spec) ** 2, ref=np.max)
+    librosa.display.specshow(log_chroma_spec)
+    plt.axis("off")
+    plt.savefig(
+        f"{TEMP_DIR}/output.jpg",
+        bbox_inches="tight",
+        pad_inches=0.0,
+    )
+    plt.close()
 
 
 def infer(wav_path: str, log_name: str, folder_path=TEMP_DIR):
@@ -95,13 +80,15 @@ def infer(wav_path: str, log_name: str, folder_path=TEMP_DIR):
     if not wav_path:
         return None, "Please input an audio!"
 
+    spec = log_name.split("_")[-3]
+    os.makedirs(folder_path, exist_ok=True)
     try:
         model = EvalNet(log_name, len(CLASSES)).model
+        eval("audio2%s" % spec)(wav_path)
+
     except Exception as e:
         return None, f"{e}"
 
-    spec = log_name.split("_")[-3]
-    eval("audio2%s" % spec)(wav_path)
     input = embed_img(f"{folder_path}/output.jpg")
     output: torch.Tensor = model(input)
     pred_id = torch.max(output.data, 1)[1]