Spaces:

factorstudios
/

DWSO

Sleeping

App Files Files Community

factorstudios commited on 25 days ago

Commit

17a3f15

verified ·

1 Parent(s): 4445143

Update app.py

Browse files

Files changed (1) hide show

app.py +80 -16

app.py CHANGED Viewed

@@ -7,17 +7,20 @@ Run with:
 from __future__ import annotations
 import asyncio
 import json
 import re
 import tempfile
 from pathlib import Path
 from typing import Final
-from urllib.parse import urlparse
 from uuid import uuid4
 from fastapi import FastAPI, Form
-from fastapi.responses import FileResponse, HTMLResponse
-from starlette.background import BackgroundTask
 from playwright.async_api import Error as PlaywrightError
 from playwright.async_api import TimeoutError as PlaywrightTimeoutError
 from playwright.async_api import async_playwright
@@ -28,6 +31,10 @@ RESULT_PAGE_TIMEOUT_MS: Final = 90_000
 MEDIA_TIMEOUT_MS: Final = 90_000
 SAVE_TT_ATTEMPTS: Final = 2
 TMP_DIR: Final = Path(tempfile.gettempdir()) / "savett_fastapi_downloads"
 app = FastAPI(title="TikTok MP4 Downloader", docs_url=None, redoc_url=None)
 # The downloader uses a full Chromium browser for each request. Serializing jobs
@@ -35,10 +42,27 @@ app = FastAPI(title="TikTok MP4 Downloader", docs_url=None, redoc_url=None)
 download_lock = asyncio.Lock()
-def render_home(error: str | None = None) -> str:
     error_markup = ""
     if error:
-        error_markup = f'<div class="notice error" role="alert">{error}</div>'
     return f"""<!doctype html>
 <html lang="en">
@@ -86,6 +110,9 @@ def render_home(error: str | None = None) -> str:
     button:active {{ transform: translateY(1px); }}
     .notice {{ margin: 0 38px; padding: 12px 14px; border-radius: 10px; font-size: .92rem; line-height: 1.4; }}
     .error {{ background: #fff0f0; border: 1px solid #f3c2c2; color: #9a2424; }}
     .footer {{ margin: 0; padding: 6px 38px 34px; color: #697488; font-size: .82rem; line-height: 1.55; }}
     .loading {{ display: none; margin-top: 12px; color: #526379; font-size: .9rem; text-align: center; }}
     form.is-submitting button {{ opacity: .75; cursor: wait; }}
@@ -98,16 +125,17 @@ def render_home(error: str | None = None) -> str:
       <header class="header">
         <p class="eyebrow">Local FastAPI utility</p>
         <h1 id="page-title">Download a TikTok MP4</h1>
-        <p class="subhead">Paste a public TikTok video link. The server fetches the MP4 and your browser saves the completed file.</p>
       </header>
       {error_markup}
       <form id="download-form" action="/download" method="post">
         <label for="tiktok-url">TikTok video URL</label>
         <input id="tiktok-url" name="tiktok_url" type="url" required autocomplete="off" placeholder="https://www.tiktok.com/@creator/video/...">
-        <button type="submit">Download MP4</button>
-        <p class="loading" aria-live="polite">Processing the link and preparing your download. This can take up to a minute.</p>
       </form>
-      <p class="footer">Use this utility only to download videos that you own or are otherwise authorized to save. Download links returned by the source service are temporary.</p>
     </section>
   </main>
   <script>
@@ -201,6 +229,35 @@ async def wait_for_mp4_payload(page, tiktok_url: str) -> str:
     )
 async def fetch_mp4(tiktok_url: str) -> tuple[Path, str]:
     """Use Playwright to get SaveTT's MP4 URLs and store the first usable file."""
     TMP_DIR.mkdir(parents=True, exist_ok=True)
@@ -281,15 +338,22 @@ async def download(tiktok_url: str = Form(...)):
     except ValueError as exc:
         return HTMLResponse(render_home(str(exc)), status_code=400)
     try:
         async with download_lock:
-            file_path, download_name = await fetch_mp4(validated_url)
     except Exception as exc:
         return HTMLResponse(render_home(f"Download failed: {exc}"), status_code=502)
-    return FileResponse(
-        file_path,
-        media_type="video/mp4",
-        filename=download_name,
-        background=BackgroundTask(remove_file, file_path),
-    )

 from __future__ import annotations
 import asyncio
+from datetime import datetime, timezone
+from html import escape
 import json
+import os
 import re
 import tempfile
 from pathlib import Path
 from typing import Final
+from urllib.parse import quote, urlparse
 from uuid import uuid4
 from fastapi import FastAPI, Form
+from fastapi.responses import HTMLResponse
+from huggingface_hub import HfApi
 from playwright.async_api import Error as PlaywrightError
 from playwright.async_api import TimeoutError as PlaywrightTimeoutError
 from playwright.async_api import async_playwright
 MEDIA_TIMEOUT_MS: Final = 90_000
 SAVE_TT_ATTEMPTS: Final = 2
 TMP_DIR: Final = Path(tempfile.gettempdir()) / "savett_fastapi_downloads"
+HF_TOKEN_ENV: Final = "HF_TOKEN"
+HF_DATASET_REPO_ID: Final = "Elias2211/cont"
+HF_DATASET_DIRECTORY: Final = "v1"
+HF_DATASET_FILE_BASE_URL: Final = f"https://huggingface.co/datasets/{HF_DATASET_REPO_ID}/blob/main"
 app = FastAPI(title="TikTok MP4 Downloader", docs_url=None, redoc_url=None)
 # The downloader uses a full Chromium browser for each request. Serializing jobs
 download_lock = asyncio.Lock()
+class HuggingFaceUploadError(RuntimeError):
+    """Raised for non-secret-bearing Hugging Face upload failures."""
+class HuggingFaceConfigurationError(HuggingFaceUploadError):
+    """Raised when the server was deployed without its required HF_TOKEN secret."""
+def render_home(error: str | None = None, success: tuple[str, str] | None = None) -> str:
     error_markup = ""
+    success_markup = ""
     if error:
+        error_markup = f'<div class="notice error" role="alert">{escape(error)}</div>'
+    if success:
+        remote_path, dataset_url = success
+        success_markup = (
+            '<div class="notice success" role="status">'
+            f'Upload complete: <code>{escape(remote_path)}</code>. '
+            f'<a href="{escape(dataset_url, quote=True)}" target="_blank" rel="noreferrer">Open in dataset</a>.'
+            '</div>'
+        )
     return f"""<!doctype html>
 <html lang="en">
     button:active {{ transform: translateY(1px); }}
     .notice {{ margin: 0 38px; padding: 12px 14px; border-radius: 10px; font-size: .92rem; line-height: 1.4; }}
     .error {{ background: #fff0f0; border: 1px solid #f3c2c2; color: #9a2424; }}
+    .success {{ background: #edf9f1; border: 1px solid #b9e2c5; color: #17683a; }}
+    .success a {{ color: inherit; font-weight: 700; }}
+    code {{ overflow-wrap: anywhere; }}
     .footer {{ margin: 0; padding: 6px 38px 34px; color: #697488; font-size: .82rem; line-height: 1.55; }}
     .loading {{ display: none; margin-top: 12px; color: #526379; font-size: .9rem; text-align: center; }}
     form.is-submitting button {{ opacity: .75; cursor: wait; }}
       <header class="header">
         <p class="eyebrow">Local FastAPI utility</p>
         <h1 id="page-title">Download a TikTok MP4</h1>
+        <p class="subhead">Paste a public TikTok video link. The server retrieves the MP4 and uploads it to the configured Hugging Face dataset.</p>
       </header>
       {error_markup}
+      {success_markup}
       <form id="download-form" action="/download" method="post">
         <label for="tiktok-url">TikTok video URL</label>
         <input id="tiktok-url" name="tiktok_url" type="url" required autocomplete="off" placeholder="https://www.tiktok.com/@creator/video/...">
+        <button type="submit">Upload MP4 to Dataset</button>
+        <p class="loading" aria-live="polite">Downloading the video and uploading it to the dataset. This can take a few minutes.</p>
       </form>
+      <p class="footer">Use this utility only to archive videos that you own or are otherwise authorized to save. Temporary local files are removed after the dataset upload completes.</p>
     </section>
   </main>
   <script>
     )
+def upload_mp4_to_dataset(local_path: Path, video_id: str) -> tuple[str, str]:
+    """Upload a local MP4 to the configured dataset without persisting its token."""
+    token = os.getenv(HF_TOKEN_ENV)
+    if not token:
+        raise HuggingFaceConfigurationError(
+            f"{HF_TOKEN_ENV} is not configured in the server environment. Add a write token as a deployment secret."
+        )
+    timestamp = datetime.now(timezone.utc).strftime("%Y%m%dT%H%M%SZ")
+    remote_path = f"{HF_DATASET_DIRECTORY}/{timestamp}_{video_id}_{uuid4().hex[:8]}.mp4"
+    try:
+        api = HfApi(token=token, library_name="savett_fastapi_uploader")
+        api.upload_file(
+            path_or_fileobj=str(local_path),
+            path_in_repo=remote_path,
+            repo_id=HF_DATASET_REPO_ID,
+            repo_type="dataset",
+            commit_message=f"Add TikTok video {video_id}",
+        )
+    except Exception as exc:
+        raise HuggingFaceUploadError(
+            "Hugging Face rejected or interrupted the dataset upload. "
+            "Check that HF_TOKEN has write access to Elias2211/cont."
+        ) from exc
+    dataset_url = f"{HF_DATASET_FILE_BASE_URL}/{quote(remote_path, safe='/')}"
+    return remote_path, dataset_url
 async def fetch_mp4(tiktok_url: str) -> tuple[Path, str]:
     """Use Playwright to get SaveTT's MP4 URLs and store the first usable file."""
     TMP_DIR.mkdir(parents=True, exist_ok=True)
     except ValueError as exc:
         return HTMLResponse(render_home(str(exc)), status_code=400)
+    file_path: Path | None = None
     try:
         async with download_lock:
+            file_path, _download_name = await fetch_mp4(validated_url)
+            video_id = video_id_from_url(validated_url)
+            remote_path, dataset_url = await asyncio.to_thread(
+                upload_mp4_to_dataset, file_path, video_id
+            )
+    except HuggingFaceConfigurationError as exc:
+        return HTMLResponse(render_home(f"Server configuration error: {exc}"), status_code=503)
+    except HuggingFaceUploadError as exc:
+        return HTMLResponse(render_home(f"Upload failed: {exc}"), status_code=502)
     except Exception as exc:
         return HTMLResponse(render_home(f"Download failed: {exc}"), status_code=502)
+    finally:
+        if file_path is not None:
+            remove_file(file_path)
+    return HTMLResponse(render_home(success=(remote_path, dataset_url)))