Spaces:
Paused
Paused
Fix Playwright async API compatibility
Browse files
app.py
CHANGED
|
@@ -9,7 +9,7 @@ from fastapi import FastAPI, HTTPException, File, UploadFile
|
|
| 9 |
from fastapi.responses import JSONResponse, RedirectResponse
|
| 10 |
from fastapi.middleware.cors import CORSMiddleware
|
| 11 |
from pydantic import BaseModel
|
| 12 |
-
from playwright.
|
| 13 |
from huggingface_hub import HfApi, upload_file
|
| 14 |
|
| 15 |
|
|
@@ -24,10 +24,10 @@ class PDFResponse(BaseModel):
|
|
| 24 |
repository_url: str
|
| 25 |
|
| 26 |
|
| 27 |
-
def html_to_pdf_api(html_content: str) -> tuple[str, str]:
|
| 28 |
"""
|
| 29 |
HTMLコンテンツをPDFに変換してHugging Faceデータセットリポジトリにアップロード
|
| 30 |
-
knowledge.txtのPlaywright
|
| 31 |
"""
|
| 32 |
try:
|
| 33 |
# 環境変数からリポジトリ情報を取得
|
|
@@ -48,26 +48,26 @@ def html_to_pdf_api(html_content: str) -> tuple[str, str]:
|
|
| 48 |
with tempfile.NamedTemporaryFile(suffix=".pdf", delete=False) as temp_file:
|
| 49 |
temp_path = temp_file.name
|
| 50 |
|
| 51 |
-
# Playwright
|
| 52 |
-
with
|
| 53 |
-
browser = pw.chromium.launch(headless=True)
|
| 54 |
-
page = browser.new_page()
|
| 55 |
|
| 56 |
# HTMLコンテンツを設定(外部リソース読み込み待機)
|
| 57 |
-
page.set_content(html_content, wait_until="networkidle")
|
| 58 |
|
| 59 |
# 印刷メディアを有効にする
|
| 60 |
-
page.emulate_media(media="print")
|
| 61 |
|
| 62 |
# PDFを生成(test.htmlの設定に準拠)
|
| 63 |
-
page.pdf(
|
| 64 |
path=temp_path,
|
| 65 |
format="A4",
|
| 66 |
print_background=True,
|
| 67 |
margin={"top":"15mm","bottom":"15mm","left":"15mm","right":"15mm"}
|
| 68 |
)
|
| 69 |
|
| 70 |
-
browser.close()
|
| 71 |
|
| 72 |
# Hugging Face リポジトリにアップロード
|
| 73 |
api = HfApi(token=hf_token)
|
|
@@ -151,7 +151,7 @@ async def convert_html_to_pdf(request: HTMLRequest):
|
|
| 151 |
raise HTTPException(status_code=400, detail="HTMLコンテンツが空です")
|
| 152 |
|
| 153 |
try:
|
| 154 |
-
filename, pdf_url = html_to_pdf_api(request.html_content)
|
| 155 |
hf_repo_id = os.getenv("HF_DATASET_REPO_ID")
|
| 156 |
repository_url = f"https://huggingface.co/datasets/{hf_repo_id}"
|
| 157 |
|
|
|
|
| 9 |
from fastapi.responses import JSONResponse, RedirectResponse
|
| 10 |
from fastapi.middleware.cors import CORSMiddleware
|
| 11 |
from pydantic import BaseModel
|
| 12 |
+
from playwright.async_api import async_playwright
|
| 13 |
from huggingface_hub import HfApi, upload_file
|
| 14 |
|
| 15 |
|
|
|
|
| 24 |
repository_url: str
|
| 25 |
|
| 26 |
|
| 27 |
+
async def html_to_pdf_api(html_content: str) -> tuple[str, str]:
|
| 28 |
"""
|
| 29 |
HTMLコンテンツをPDFに変換してHugging Faceデータセットリポジトリにアップロード
|
| 30 |
+
knowledge.txtのPlaywright手法を使用(Async版)
|
| 31 |
"""
|
| 32 |
try:
|
| 33 |
# 環境変数からリポジトリ情報を取得
|
|
|
|
| 48 |
with tempfile.NamedTemporaryFile(suffix=".pdf", delete=False) as temp_file:
|
| 49 |
temp_path = temp_file.name
|
| 50 |
|
| 51 |
+
# Playwrightでヘッドレスブラウザを起動(Async版)
|
| 52 |
+
async with async_playwright() as pw:
|
| 53 |
+
browser = await pw.chromium.launch(headless=True)
|
| 54 |
+
page = await browser.new_page()
|
| 55 |
|
| 56 |
# HTMLコンテンツを設定(外部リソース読み込み待機)
|
| 57 |
+
await page.set_content(html_content, wait_until="networkidle")
|
| 58 |
|
| 59 |
# 印刷メディアを有効にする
|
| 60 |
+
await page.emulate_media(media="print")
|
| 61 |
|
| 62 |
# PDFを生成(test.htmlの設定に準拠)
|
| 63 |
+
await page.pdf(
|
| 64 |
path=temp_path,
|
| 65 |
format="A4",
|
| 66 |
print_background=True,
|
| 67 |
margin={"top":"15mm","bottom":"15mm","left":"15mm","right":"15mm"}
|
| 68 |
)
|
| 69 |
|
| 70 |
+
await browser.close()
|
| 71 |
|
| 72 |
# Hugging Face リポジトリにアップロード
|
| 73 |
api = HfApi(token=hf_token)
|
|
|
|
| 151 |
raise HTTPException(status_code=400, detail="HTMLコンテンツが空です")
|
| 152 |
|
| 153 |
try:
|
| 154 |
+
filename, pdf_url = await html_to_pdf_api(request.html_content)
|
| 155 |
hf_repo_id = os.getenv("HF_DATASET_REPO_ID")
|
| 156 |
repository_url = f"https://huggingface.co/datasets/{hf_repo_id}"
|
| 157 |
|