add new ftp host

components/dashboard_page.py

@@ -5,15 +5,14 @@ from sqlalchemy import orm, func # Added func for count
 
 from components.header import Header
 from utils.logger import Logger # Changed from get_logger to Logger
-from utils.gdrive_downloader import PublicFolderAudioLoader
+from utils.cloud_server_audio_loader import CloudServerAudioLoader
 from config import conf
 from utils.database import get_db
 from data.models import Annotation, AudioTrim, TTSData, AnnotationInterval # Added AnnotationInterval
 from data.repository.annotator_workload_repo import AnnotatorWorkloadRepo # For progress
 
 log = Logger() # Changed from get_logger() to Logger()
-LOADER = PublicFolderAudioLoader(conf.GDRIVE_API_KEY)
-GDRIVE_FOLDER = conf.GDRIVE_FOLDER
+LOADER = CloudServerAudioLoader(conf.FTP_URL)
 
 
 class DashboardPage:
@@ -188,12 +187,12 @@ class DashboardPage:
                     log.error(f"Error fetching progress for user {user_id}: {e}") 
                     return "Annotation Progress: Error" # Added label
 
-        def download_voice_fn(folder_link, filename_to_load, autoplay_on_load=True): # Autoplay here is for the btn_load_voice click
+        def download_voice_fn(filename_to_load, autoplay_on_load=True): # Autoplay here is for the btn_load_voice click
             if not filename_to_load:
                 return None, None, gr.update(value=None, autoplay=False)
             try:
                 log.info(f"Downloading voice: {filename_to_load}, Autoplay: {autoplay_on_load}")
-                sr, wav = LOADER.load_audio(folder_link, filename_to_load)
+                sr, wav = LOADER.load_audio(filename_to_load)
                 return (sr, wav), (sr, wav.copy()), gr.update(value=(sr, wav), autoplay=autoplay_on_load)
             except Exception as e:
                 log.error(f"GDrive download failed for {filename_to_load}: {e}") 
@@ -602,7 +601,7 @@ class DashboardPage:
             outputs=self.interactive_ui_elements
         ).then(
             fn=download_voice_fn,
-            inputs=[gr.State(GDRIVE_FOLDER), self.filename, gr.State(True)], # Autoplay TRUE
+            inputs=[self.filename, gr.State(True)], # Autoplay TRUE
             outputs=[self.audio, self.original_audio_state, self.audio],
         ).then( 
             fn=self._apply_multiple_trims_fn,

config.py

@@ -12,8 +12,7 @@ class Config(BaseSettings):
     DB_NAME: str = os.getenv("DB_NAME", "defaultdb")
     HF_TOKEN: str = os.environ.get("HF_TOKEN")
     HF_TTS_DS_REPO: str = os.environ.get("HF_TTS_DS_REPO")
-    GDRIVE_API_KEY: str = os.environ.get("GDRIVE_API_KEY") 
-    GDRIVE_FOLDER: str = os.environ.get("GDRIVE_FOLDER") 
+    FTP_URL: str = os.environ.get("FTP_URL") 
 
     APP_TITLE: str = "Gooya TTS Annotation Tools"
 

test/gdrive_test.py

@@ -1,22 +0,0 @@
-import gradio as gr
-from utils.gdrive_downloader import PublicFolderAudioLoader
-from config import conf
-
-LOADER  = PublicFolderAudioLoader(conf.GDRIVE_API_KEY)
-
-def fetch_audio(folder_link, filename):
-    sr, wav = LOADER.load_audio(folder_link, filename)
-    return (sr, wav)
-
-demo = gr.Interface(
-    fn=fetch_audio,
-    inputs=[
-        gr.Textbox(label="Folder URL or ID",
-                   value=conf.GDRIVE_FOLDER),
-        gr.Textbox(label="Filename (e.g. 0001.wav)")
-    ],
-    outputs=gr.Audio(label="🔊 Audio"),
-)
-
-if __name__ == "__main__":
-    demo.launch()

utils/cloud_server_audio_loader.py

@@ -0,0 +1,42 @@
+# cloud_server_audio_loader.py
+
+import io
+import requests
+import numpy as np
+from pydub import AudioSegment
+
+class CloudServerAudioLoader:
+    def __init__(self, base_url: str) -> None:
+        if not base_url.endswith("/"):
+            base_url += "/"
+        self.base_url = base_url
+
+    def _download_to_buf(self, filename: str) -> io.BytesIO:
+        url = self.base_url + filename
+        resp = requests.get(url, stream=True)
+        if resp.status_code != 200:
+            raise FileNotFoundError(f"'{filename}' not found. HTTP {resp.status_code}")
+        buf = io.BytesIO(resp.content)
+        buf.seek(0)
+        return buf
+
+    def load_audio(self, filename: str) -> tuple[int, np.ndarray]:
+        buf = self._download_to_buf(filename)
+        seg = AudioSegment.from_file(buf)
+        samples = np.array(seg.get_array_of_samples())
+
+        if seg.channels > 1:
+            samples = samples.reshape(-1, seg.channels)
+
+        if np.issubdtype(samples.dtype, np.integer):
+            max_int = np.iinfo(samples.dtype).max 
+            samples = samples.astype(np.float32)
+            samples /= max_int  
+        else:
+            max_val = np.abs(samples).max()
+            if max_val > 1:
+                samples = samples / max_val
+            samples = samples.astype(np.float32)
+        # --------------------------------------------------------
+
+        return seg.frame_rate, samples

utils/gdrive_downloader.py

@@ -1,98 +0,0 @@
-# gdrive_downloader.py
-
-from __future__ import annotations
-import io
-import re
-import numpy as np
-from pydub import AudioSegment
-from googleapiclient.discovery import build
-from googleapiclient.http import MediaIoBaseDownload
-
-
-def extract_folder_id(url_or_id: str) -> str:
-    """
-    اگر کاربر لینک فولدر بدهد ← ID را برمی‌گرداند.
-    اگر خودش ID باشد همان را برمی‌گرداند.
-    """
-    s = url_or_id.strip()
-    if "/" not in s and "?" not in s:
-        return s  # احتمالاً خودش ID است
-    m = re.search(r"/folders/([a-zA-Z0-9_-]{10,})", s)
-    if not m:
-        raise ValueError("Cannot extract folder id from url")
-    return m.group(1)
-
-
-class PublicFolderAudioLoader:
-    """
-    دانلودر فایل صوتی از فولدر عمومی گوگل‌درایو بدون ذخیره روی دیسک.
-
-    Parameters
-    ----------
-    api_key : str
-        Google API Key (کیِ عمومی؛ نه OAuth, نه سرویس‌اکانت).
-    """
-
-    def __init__(self, api_key: str) -> None:
-        self.svc = build("drive", "v3", developerKey=api_key, cache_discovery=False)
-
-    # ---------- helpers ---------- #
-    def _file_id_by_name(self, folder_id: str, filename: str) -> str:
-        q = (
-            f"'{folder_id}' in parents "
-            f"and name = '{filename}' "
-            f"and trashed = false"
-        )
-        rsp = (
-            self.svc.files()
-            .list(q=q, fields="files(id,name)", pageSize=5, supportsAllDrives=True)
-            .execute()
-        )
-        files = rsp.get("files", [])
-        if not files:
-            raise FileNotFoundError(f"'{filename}' not found in folder {folder_id}")
-        return files[0]["id"]
-
-    def _download_to_buf(self, file_id: str) -> io.BytesIO:
-        request = self.svc.files().get_media(fileId=file_id, supportsAllDrives=True)
-        buf = io.BytesIO()
-        downloader = MediaIoBaseDownload(buf, request)
-        done = False
-        while not done:
-            _, done = downloader.next_chunk()
-        buf.seek(0)
-        return buf
-
-    # ---------- public ---------- #
-    def load_audio(
-        self,
-        folder_url_or_id: str,
-        filename: str,
-    ) -> tuple[int, np.ndarray]:
-        # """
-        # فایل را به `(sample_rate, np.ndarray)` نرمال‌شده در بازه‌ی [-1,1] تبدیل می‌کند.
-        # """
-        folder_id = extract_folder_id(folder_url_or_id)
-        file_id = self._file_id_by_name(folder_id, filename)
-        buf = self._download_to_buf(file_id)
-        seg = AudioSegment.from_file(buf)
-        samples = np.array(seg.get_array_of_samples())
-
-        # اگر چندکاناله بود، شکل دهیم
-        if seg.channels > 1:
-            samples = samples.reshape(-1, seg.channels)
-
-        # ---------------------- نرمال‌سازی ----------------------
-        if np.issubdtype(samples.dtype, np.integer):
-            max_int = np.iinfo(samples.dtype).max  # ← قبل از cast
-            samples = samples.astype(np.float32)
-            samples /= max_int  # ← از max_int استفاده می‌کنیم
-        else:
-            # در حالت float
-            max_val = np.abs(samples).max()
-            if max_val > 1:
-                samples = samples / max_val
-            samples = samples.astype(np.float32)
-        # --------------------------------------------------------
-
-        return seg.frame_rate, samples