| | import gradio as gr |
| | from fastapi import FastAPI, HTTPException, Body |
| | from fastapi.responses import StreamingResponse |
| | from fastapi.staticfiles import StaticFiles |
| | from fastapi.middleware.cors import CORSMiddleware |
| | from pydantic import BaseModel |
| | from selenium import webdriver |
| | from selenium.webdriver.chrome.options import Options |
| | from selenium.webdriver.common.by import By |
| | from selenium.webdriver.support.ui import WebDriverWait |
| | from selenium.webdriver.support import expected_conditions as EC |
| | from PIL import Image |
| | from io import BytesIO |
| | import tempfile |
| | import time |
| | import os |
| | import logging |
| | from huggingface_hub import hf_hub_download |
| |
|
| | |
| | import google.generativeai as genai |
| |
|
| | |
| | logging.basicConfig(level=logging.INFO) |
| | logger = logging.getLogger(__name__) |
| |
|
| | |
| | class GeminiRequest(BaseModel): |
| | """Geminiへのリクエストデータモデル""" |
| | text: str |
| | extension_percentage: float = 10.0 |
| | temperature: float = 0.5 |
| | trim_whitespace: bool = True |
| | style: str = "standard" |
| |
|
| | class ScreenshotRequest(BaseModel): |
| | """スクリーンショットリクエストモデル""" |
| | html_code: str |
| | extension_percentage: float = 10.0 |
| | trim_whitespace: bool = True |
| | style: str = "standard" |
| |
|
| | |
| | def enhance_font_awesome_layout(html_code): |
| | """Font Awesomeレイアウトを改善するCSSを追加""" |
| | |
| | fa_fix_css = """ |
| | <style> |
| | /* Font Awesomeアイコンのレイアウト修正 */ |
| | [class*="fa-"] { |
| | display: inline-block !important; |
| | margin-right: 8px !important; |
| | vertical-align: middle !important; |
| | } |
| | |
| | /* テキスト内のアイコン位置調整 */ |
| | h1 [class*="fa-"], h2 [class*="fa-"], h3 [class*="fa-"], |
| | h4 [class*="fa-"], h5 [class*="fa-"], h6 [class*="fa-"] { |
| | vertical-align: middle !important; |
| | margin-right: 10px !important; |
| | } |
| | |
| | /* 特定パターンの修正 */ |
| | .fa + span, .fas + span, .far + span, .fab + span, |
| | span + .fa, span + .fas, span + .far + span { |
| | display: inline-block !important; |
| | margin-left: 5px !important; |
| | } |
| | |
| | /* カード内アイコン修正 */ |
| | .card [class*="fa-"], .card-body [class*="fa-"] { |
| | float: none !important; |
| | clear: none !important; |
| | position: relative !important; |
| | } |
| | |
| | /* アイコンと文字が重なる場合の調整 */ |
| | li [class*="fa-"], p [class*="fa-"] { |
| | margin-right: 10px !important; |
| | } |
| | |
| | /* インラインアイコンのスペーシング */ |
| | .inline-icon { |
| | display: inline-flex !important; |
| | align-items: center !important; |
| | justify-content: flex-start !important; |
| | } |
| | |
| | /* アイコン後のテキスト */ |
| | [class*="fa-"] + span { |
| | display: inline-block !important; |
| | vertical-align: middle !important; |
| | } |
| | </style> |
| | """ |
| |
|
| | |
| | if '<head>' in html_code: |
| | return html_code.replace('</head>', f'{fa_fix_css}</head>') |
| | |
| | elif '<html' in html_code: |
| | head_end = html_code.find('</head>') |
| | if head_end > 0: |
| | return html_code[:head_end] + fa_fix_css + html_code[head_end:] |
| | else: |
| | body_start = html_code.find('<body') |
| | if body_start > 0: |
| | return html_code[:body_start] + f'<head>{fa_fix_css}</head>' + html_code[body_start:] |
| |
|
| | |
| | return f'<html><head>{fa_fix_css}</head>' + html_code + '</html>' |
| |
|
| | def load_system_instruction(style="standard"): |
| | """ |
| | 指定されたスタイルのシステムインストラクションを読み込む |
| | |
| | Args: |
| | style: 使用するスタイル名 (standard, cute, resort, cool, dental) |
| | |
| | Returns: |
| | 読み込まれたシステムインストラクション |
| | """ |
| | try: |
| | |
| | valid_styles = ["standard", "cute", "resort", "cool", "dental"] |
| |
|
| | |
| | if style not in valid_styles: |
| | logger.warning(f"無効なスタイル '{style}' が指定されました。デフォルトの 'standard' を使用します。") |
| | style = "standard" |
| |
|
| | logger.info(f"スタイル '{style}' のシステムインストラクションを読み込みます") |
| |
|
| | |
| | local_path = os.path.join(os.path.dirname(__file__), style, "prompt.txt") |
| |
|
| | |
| | if os.path.exists(local_path): |
| | logger.info(f"ローカルファイルを使用: {local_path}") |
| | with open(local_path, 'r', encoding='utf-8') as file: |
| | instruction = file.read() |
| | return instruction |
| |
|
| | |
| | try: |
| | |
| | file_path = hf_hub_download( |
| | repo_id="tomo2chin2/GURAREKOstlyle", |
| | filename=f"{style}/prompt.txt", |
| | repo_type="dataset" |
| | ) |
| |
|
| | logger.info(f"スタイル '{style}' のプロンプトをHuggingFaceから読み込みました: {file_path}") |
| | with open(file_path, 'r', encoding='utf-8') as file: |
| | instruction = file.read() |
| | return instruction |
| |
|
| | except Exception as style_error: |
| | |
| | logger.warning(f"スタイル '{style}' のプロンプト読み込みに失敗: {str(style_error)}") |
| | logger.info("デフォルトのprompt.txtを読み込みます") |
| |
|
| | file_path = hf_hub_download( |
| | repo_id="tomo2chin2/GURAREKOstlyle", |
| | filename="prompt.txt", |
| | repo_type="dataset" |
| | ) |
| |
|
| | with open(file_path, 'r', encoding='utf-8') as file: |
| | instruction = file.read() |
| |
|
| | logger.info("デフォルトのシステムインストラクションを読み込みました") |
| | return instruction |
| |
|
| | except Exception as e: |
| | error_msg = f"システムインストラクションの読み込みに失敗: {str(e)}" |
| | logger.error(error_msg) |
| | raise ValueError(error_msg) |
| |
|
| | def generate_html_from_text(text, temperature=0.3, style="standard"): |
| | """テキストからHTMLを生成する""" |
| | try: |
| | |
| | api_key = os.environ.get("GEMINI_API_KEY") |
| | if not api_key: |
| | logger.error("GEMINI_API_KEY 環境変数が設定されていません") |
| | raise ValueError("GEMINI_API_KEY 環境変数が設定されていません") |
| |
|
| | |
| | model_name = os.environ.get("GEMINI_MODEL", "gemini-1.5-pro") |
| | logger.info(f"使用するGeminiモデル: {model_name}") |
| |
|
| | |
| | genai.configure(api_key=api_key) |
| |
|
| | |
| | system_instruction = load_system_instruction(style) |
| |
|
| | |
| | logger.info(f"Gemini APIにリクエストを送信: テキスト長さ = {len(text)}, 温度 = {temperature}, スタイル = {style}") |
| |
|
| | |
| | model = genai.GenerativeModel(model_name) |
| |
|
| | |
| | generation_config = { |
| | "temperature": temperature, |
| | "top_p": 0.7, |
| | "top_k": 20, |
| | "max_output_tokens": 8192, |
| | "candidate_count": 1 |
| | } |
| |
|
| | |
| | safety_settings = [ |
| | { |
| | "category": "HARM_CATEGORY_HARASSMENT", |
| | "threshold": "BLOCK_MEDIUM_AND_ABOVE" |
| | }, |
| | { |
| | "category": "HARM_CATEGORY_HATE_SPEECH", |
| | "threshold": "BLOCK_MEDIUM_AND_ABOVE" |
| | }, |
| | { |
| | "category": "HARM_CATEGORY_SEXUALLY_EXPLICIT", |
| | "threshold": "BLOCK_MEDIUM_AND_ABOVE" |
| | }, |
| | { |
| | "category": "HARM_CATEGORY_DANGEROUS_CONTENT", |
| | "threshold": "BLOCK_MEDIUM_AND_ABOVE" |
| | } |
| | ] |
| |
|
| | |
| | prompt = f"{system_instruction}\n\n{text}" |
| |
|
| | |
| | response = model.generate_content( |
| | prompt, |
| | generation_config=generation_config, |
| | safety_settings=safety_settings |
| | ) |
| |
|
| | |
| | raw_response = response.text |
| |
|
| | |
| | html_start = raw_response.find("```html") |
| | html_end = raw_response.rfind("```") |
| |
|
| | if html_start != -1 and html_end != -1 and html_start < html_end: |
| | html_start += 7 |
| | html_code = raw_response[html_start:html_end].strip() |
| | logger.info(f"HTMLの生成に成功: 長さ = {len(html_code)}") |
| |
|
| | |
| | html_code = enhance_font_awesome_layout(html_code) |
| | logger.info("Font Awesomeレイアウトの最適化を適用しました") |
| |
|
| | return html_code |
| | else: |
| | |
| | logger.warning("レスポンスから ```html ``` タグが見つかりませんでした。全テキストを返します。") |
| | return raw_response |
| |
|
| | except Exception as e: |
| | logger.error(f"HTML生成中にエラーが発生: {e}", exc_info=True) |
| | raise Exception(f"Gemini APIでのHTML生成に失敗しました: {e}") |
| |
|
| | |
| | def trim_image_whitespace(image, threshold=250, padding=10): |
| | """ |
| | 画像から余分な白い空白をトリミングする |
| | |
| | Args: |
| | image: PIL.Image - 入力画像 |
| | threshold: int - どの明るさ以上を空白と判断するか (0-255) |
| | padding: int - トリミング後に残す余白のピクセル数 |
| | |
| | Returns: |
| | トリミングされたPIL.Image |
| | """ |
| | |
| | gray = image.convert('L') |
| |
|
| | |
| | data = gray.getdata() |
| | width, height = gray.size |
| |
|
| | |
| | min_x, min_y = width, height |
| | max_x = max_y = 0 |
| |
|
| | |
| | pixels = list(data) |
| | pixels = [pixels[i * width:(i + 1) * width] for i in range(height)] |
| |
|
| | |
| | for y in range(height): |
| | for x in range(width): |
| | if pixels[y][x] < threshold: |
| | min_x = min(min_x, x) |
| | min_y = min(min_y, y) |
| | max_x = max(max_x, x) |
| | max_y = max(max_y, y) |
| |
|
| | |
| | if min_x > max_x or min_y > max_y: |
| | logger.warning("トリミング領域が見つかりません。元の画像を返します。") |
| | return image |
| |
|
| | |
| | min_x = max(0, min_x - padding) |
| | min_y = max(0, min_y - padding) |
| | max_x = min(width - 1, max_x + padding) |
| | max_y = min(height - 1, max_y + padding) |
| |
|
| | |
| | trimmed = image.crop((min_x, min_y, max_x + 1, max_y + 1)) |
| |
|
| | logger.info(f"画像をトリミングしました: 元サイズ {width}x{height} → トリミング後 {trimmed.width}x{trimmed.height}") |
| | return trimmed |
| |
|
| | |
| |
|
| | def render_fullpage_screenshot(html_code: str, extension_percentage: float = 6.0, |
| | trim_whitespace: bool = True) -> Image.Image: |
| | """ |
| | Renders HTML code to a full-page screenshot using Selenium. |
| | |
| | Args: |
| | html_code: The HTML source code string. |
| | extension_percentage: Percentage of extra space to add vertically. |
| | trim_whitespace: Whether to trim excess whitespace from the image. |
| | |
| | Returns: |
| | A PIL Image object of the screenshot. |
| | """ |
| | tmp_path = None |
| | driver = None |
| |
|
| | |
| | try: |
| | with tempfile.NamedTemporaryFile(suffix=".html", delete=False, mode='w', encoding='utf-8') as tmp_file: |
| | tmp_path = tmp_file.name |
| | tmp_file.write(html_code) |
| | logger.info(f"HTML saved to temporary file: {tmp_path}") |
| | except Exception as e: |
| | logger.error(f"Error writing temporary HTML file: {e}") |
| | return Image.new('RGB', (1, 1), color=(0, 0, 0)) |
| |
|
| | |
| | options = Options() |
| | options.add_argument("--headless") |
| | options.add_argument("--no-sandbox") |
| | options.add_argument("--disable-dev-shm-usage") |
| | options.add_argument("--force-device-scale-factor=1") |
| | options.add_argument("--disable-features=NetworkService") |
| | options.add_argument("--dns-prefetch-disable") |
| | |
| | webdriver_path = os.environ.get("CHROMEDRIVER_PATH") |
| | if webdriver_path and os.path.exists(webdriver_path): |
| | logger.info(f"Using CHROMEDRIVER_PATH: {webdriver_path}") |
| | service = webdriver.ChromeService(executable_path=webdriver_path) |
| | else: |
| | logger.info("CHROMEDRIVER_PATH not set or invalid, using default PATH lookup.") |
| | service = None |
| |
|
| | try: |
| | logger.info("Initializing WebDriver...") |
| | if service: |
| | driver = webdriver.Chrome(service=service, options=options) |
| | else: |
| | driver = webdriver.Chrome(options=options) |
| | logger.info("WebDriver initialized.") |
| |
|
| | |
| | initial_width = 1200 |
| | initial_height = 1000 |
| | driver.set_window_size(initial_width, initial_height) |
| | file_url = "file://" + tmp_path |
| | logger.info(f"Navigating to {file_url}") |
| | driver.get(file_url) |
| |
|
| | |
| | logger.info("Waiting for body element...") |
| | WebDriverWait(driver, 15).until( |
| | EC.presence_of_element_located((By.TAG_NAME, "body")) |
| | ) |
| | logger.info("Body element found. Waiting for resource loading...") |
| |
|
| | |
| | time.sleep(3) |
| |
|
| | |
| | logger.info("Checking for Font Awesome resources...") |
| | fa_count = driver.execute_script(""" |
| | var icons = document.querySelectorAll('.fa, .fas, .far, .fab, [class*="fa-"]'); |
| | return icons.length; |
| | """) |
| | logger.info(f"Found {fa_count} Font Awesome elements") |
| |
|
| | |
| | doc_ready = driver.execute_script("return document.readyState;") |
| | logger.info(f"Document ready state: {doc_ready}") |
| |
|
| | |
| | if fa_count > 50: |
| | logger.info("Many Font Awesome icons detected, waiting additional time") |
| | time.sleep(2) |
| |
|
| | |
| | logger.info("Performing content rendering scroll...") |
| | total_height = driver.execute_script("return Math.max(document.body.scrollHeight, document.documentElement.scrollHeight);") |
| | viewport_height = driver.execute_script("return window.innerHeight;") |
| | scrolls_needed = max(1, total_height // viewport_height) |
| |
|
| | for i in range(scrolls_needed + 1): |
| | scroll_pos = i * (viewport_height - 200) |
| | driver.execute_script(f"window.scrollTo(0, {scroll_pos});") |
| | time.sleep(0.2) |
| |
|
| | |
| | driver.execute_script("window.scrollTo(0, 0);") |
| | time.sleep(0.5) |
| | logger.info("Scroll rendering completed") |
| |
|
| | |
| | driver.execute_script(""" |
| | document.documentElement.style.overflow = 'hidden'; |
| | document.body.style.overflow = 'hidden'; |
| | """) |
| | logger.info("Scrollbars hidden") |
| |
|
| | |
| | dimensions = driver.execute_script(""" |
| | return { |
| | width: Math.max( |
| | document.documentElement.scrollWidth, |
| | document.documentElement.offsetWidth, |
| | document.documentElement.clientWidth, |
| | document.body ? document.body.scrollWidth : 0, |
| | document.body ? document.body.offsetWidth : 0, |
| | document.body ? document.body.clientWidth : 0 |
| | ), |
| | height: Math.max( |
| | document.documentElement.scrollHeight, |
| | document.documentElement.offsetHeight, |
| | document.documentElement.clientHeight, |
| | document.body ? document.body.scrollHeight : 0, |
| | document.body ? document.body.offsetHeight : 0, |
| | document.body ? document.body.clientHeight : 0 |
| | ) |
| | }; |
| | """) |
| | scroll_width = dimensions['width'] |
| | scroll_height = dimensions['height'] |
| | logger.info(f"Detected dimensions: width={scroll_width}, height={scroll_height}") |
| |
|
| | |
| | driver.execute_script("window.scrollTo(0, document.body.scrollHeight);") |
| | time.sleep(0.5) |
| | driver.execute_script("window.scrollTo(0, 0);") |
| | time.sleep(0.5) |
| |
|
| | dimensions_after = driver.execute_script("return {height: Math.max(document.documentElement.scrollHeight, document.body.scrollHeight)};") |
| | scroll_height = max(scroll_height, dimensions_after['height']) |
| | logger.info(f"After scroll check, height={scroll_height}") |
| |
|
| | |
| | scroll_width = max(scroll_width, 100) |
| | scroll_height = max(scroll_height, 100) |
| | scroll_width = min(scroll_width, 2000) |
| | scroll_height = min(scroll_height, 4000) |
| |
|
| | |
| | logger.info("Waiting for layout stabilization...") |
| | time.sleep(2) |
| |
|
| | |
| | adjusted_height = int(scroll_height * (1 + extension_percentage / 100.0)) |
| | adjusted_height = max(adjusted_height, scroll_height, 100) |
| | logger.info(f"Adjusted height calculated: {adjusted_height} (extension: {extension_percentage}%)") |
| |
|
| | |
| | adjusted_width = scroll_width |
| | logger.info(f"Resizing window to: width={adjusted_width}, height={adjusted_height}") |
| | driver.set_window_size(adjusted_width, adjusted_height) |
| | time.sleep(1) |
| |
|
| | |
| | resource_state = driver.execute_script(""" |
| | return { |
| | readyState: document.readyState, |
| | resourcesComplete: !document.querySelector('img:not([complete])') && |
| | !document.querySelector('link[rel="stylesheet"]:not([loaded])') |
| | }; |
| | """) |
| | logger.info(f"Resource state: {resource_state}") |
| |
|
| | if resource_state['readyState'] != 'complete': |
| | logger.info("Document still loading, waiting additional time...") |
| | time.sleep(1) |
| |
|
| | |
| | driver.execute_script("window.scrollTo(0, 0);") |
| | time.sleep(0.5) |
| | logger.info("Scrolled to top.") |
| |
|
| | |
| | logger.info("Taking screenshot...") |
| | png = driver.get_screenshot_as_png() |
| | logger.info("Screenshot taken successfully.") |
| |
|
| | |
| | img = Image.open(BytesIO(png)) |
| | logger.info(f"Screenshot dimensions: {img.width}x{img.height}") |
| |
|
| | |
| | if trim_whitespace: |
| | img = trim_image_whitespace(img, threshold=248, padding=20) |
| | logger.info(f"Trimmed dimensions: {img.width}x{img.height}") |
| |
|
| | return img |
| |
|
| | except Exception as e: |
| | logger.error(f"Error during screenshot generation: {e}", exc_info=True) |
| | |
| | return Image.new('RGB', (1, 1), color=(0, 0, 0)) |
| | finally: |
| | logger.info("Cleaning up...") |
| | if driver: |
| | try: |
| | driver.quit() |
| | logger.info("WebDriver quit successfully.") |
| | except Exception as e: |
| | logger.error(f"Error quitting WebDriver: {e}", exc_info=True) |
| | if tmp_path and os.path.exists(tmp_path): |
| | try: |
| | os.remove(tmp_path) |
| | logger.info(f"Temporary file {tmp_path} removed.") |
| | except Exception as e: |
| | logger.error(f"Error removing temporary file {tmp_path}: {e}", exc_info=True) |
| |
|
| | |
| | def text_to_screenshot(text: str, extension_percentage: float, temperature: float = 0.3, |
| | trim_whitespace: bool = True, style: str = "standard") -> Image.Image: |
| | """テキストをGemini APIでHTMLに変換し、スクリーンショットを生成する統合関数""" |
| | try: |
| | |
| | html_code = generate_html_from_text(text, temperature, style) |
| |
|
| | |
| | return render_fullpage_screenshot(html_code, extension_percentage, trim_whitespace) |
| | except Exception as e: |
| | logger.error(f"テキストからスクリーンショット生成中にエラーが発生: {e}", exc_info=True) |
| | return Image.new('RGB', (1, 1), color=(0, 0, 0)) |
| |
|
| | |
| | app = FastAPI() |
| |
|
| | |
| | app.add_middleware( |
| | CORSMiddleware, |
| | allow_origins=["*"], |
| | allow_credentials=True, |
| | allow_methods=["*"], |
| | allow_headers=["*"], |
| | ) |
| |
|
| | |
| | |
| | gradio_dir = os.path.dirname(gr.__file__) |
| | logger.info(f"Gradio version: {gr.__version__}") |
| | logger.info(f"Gradio directory: {gradio_dir}") |
| |
|
| | |
| | static_dir = os.path.join(gradio_dir, "templates", "frontend", "static") |
| | if os.path.exists(static_dir): |
| | logger.info(f"Mounting static directory: {static_dir}") |
| | app.mount("/static", StaticFiles(directory=static_dir), name="static") |
| |
|
| | |
| | app_dir = os.path.join(gradio_dir, "templates", "frontend", "_app") |
| | if os.path.exists(app_dir): |
| | logger.info(f"Mounting _app directory: {app_dir}") |
| | app.mount("/_app", StaticFiles(directory=app_dir), name="_app") |
| |
|
| | |
| | assets_dir = os.path.join(gradio_dir, "templates", "frontend", "assets") |
| | if os.path.exists(assets_dir): |
| | logger.info(f"Mounting assets directory: {assets_dir}") |
| | app.mount("/assets", StaticFiles(directory=assets_dir), name="assets") |
| |
|
| | |
| | cdn_dir = os.path.join(gradio_dir, "templates", "cdn") |
| | if os.path.exists(cdn_dir): |
| | logger.info(f"Mounting cdn directory: {cdn_dir}") |
| | app.mount("/cdn", StaticFiles(directory=cdn_dir), name="cdn") |
| |
|
| |
|
| | |
| | @app.post("/api/screenshot", |
| | response_class=StreamingResponse, |
| | tags=["Screenshot"], |
| | summary="Render HTML to Full Page Screenshot", |
| | description="Takes HTML code and an optional vertical extension percentage, renders it using a headless browser, and returns the full-page screenshot as a PNG image.") |
| | async def api_render_screenshot(request: ScreenshotRequest): |
| | """ |
| | API endpoint to render HTML and return a screenshot. |
| | """ |
| | try: |
| | logger.info(f"API request received. Extension: {request.extension_percentage}%") |
| | |
| | pil_image = render_fullpage_screenshot( |
| | request.html_code, |
| | request.extension_percentage, |
| | request.trim_whitespace |
| | ) |
| |
|
| | if pil_image.size == (1, 1): |
| | logger.error("Screenshot generation failed, returning 1x1 error image.") |
| | |
| | |
| |
|
| | |
| | img_byte_arr = BytesIO() |
| | pil_image.save(img_byte_arr, format='PNG') |
| | img_byte_arr.seek(0) |
| |
|
| | logger.info("Returning screenshot as PNG stream.") |
| | return StreamingResponse(img_byte_arr, media_type="image/png") |
| |
|
| | except Exception as e: |
| | logger.error(f"API Error: {e}", exc_info=True) |
| | raise HTTPException(status_code=500, detail=f"Internal Server Error: {e}") |
| |
|
| | |
| | @app.post("/api/text-to-screenshot", |
| | response_class=StreamingResponse, |
| | tags=["Screenshot", "Gemini"], |
| | summary="テキストからインフォグラフィックを生成", |
| | description="テキストをGemini APIを使ってHTMLインフォグラフィックに変換し、スクリーンショットとして返します。") |
| | async def api_text_to_screenshot(request: GeminiRequest): |
| | """ |
| | テキストからHTMLインフォグラフィックを生成してスクリーンショットを返すAPIエンドポイント |
| | """ |
| | try: |
| | logger.info(f"テキスト→スクリーンショットAPIリクエスト受信。テキスト長さ: {len(request.text)}, " |
| | f"拡張率: {request.extension_percentage}%, 温度: {request.temperature}, " |
| | f"スタイル: {request.style}") |
| |
|
| | |
| | pil_image = text_to_screenshot( |
| | request.text, |
| | request.extension_percentage, |
| | request.temperature, |
| | request.trim_whitespace, |
| | request.style |
| | ) |
| |
|
| | if pil_image.size == (1, 1): |
| | logger.error("スクリーンショット生成に失敗しました。1x1エラー画像を返します。") |
| | |
| |
|
| |
|
| | |
| | img_byte_arr = BytesIO() |
| | pil_image.save(img_byte_arr, format='PNG') |
| | img_byte_arr.seek(0) |
| |
|
| | logger.info("スクリーンショットをPNGストリームとして返します。") |
| | return StreamingResponse(img_byte_arr, media_type="image/png") |
| |
|
| | except Exception as e: |
| | logger.error(f"API Error: {e}", exc_info=True) |
| | raise HTTPException(status_code=500, detail=f"Internal Server Error: {e}") |
| |
|
| | |
| | |
| | def process_input(input_mode, input_text, extension_percentage, temperature, trim_whitespace, style): |
| | """入力モードに応じて適切な処理を行う""" |
| | if input_mode == "HTML入力": |
| | |
| | return render_fullpage_screenshot(input_text, extension_percentage, trim_whitespace) |
| | else: |
| | |
| | return text_to_screenshot(input_text, extension_percentage, temperature, trim_whitespace, style) |
| |
|
| | |
| | with gr.Blocks(title="Full Page Screenshot (テキスト変換対応)", theme=gr.themes.Base()) as iface: |
| | gr.Markdown("# HTMLビューア & テキスト→インフォグラフィック変換") |
| | gr.Markdown("HTMLコードをレンダリングするか、テキストをGemini APIでインフォグラフィックに変換して画像として取得します。") |
| |
|
| | with gr.Row(): |
| | input_mode = gr.Radio( |
| | ["HTML入力", "テキスト入力"], |
| | label="入力モード", |
| | value="HTML入力" |
| | ) |
| |
|
| | |
| | input_text = gr.Textbox( |
| | lines=15, |
| | label="入力", |
| | placeholder="HTMLコードまたはテキストを入力してください。入力モードに応じて処理されます。" |
| | ) |
| |
|
| | with gr.Row(): |
| | with gr.Column(scale=1): |
| | |
| | style_dropdown = gr.Dropdown( |
| | choices=["standard", "cute", "resort", "cool", "dental"], |
| | value="standard", |
| | label="デザインスタイル", |
| | info="テキスト→HTML変換時のデザインテーマを選択します", |
| | visible=False |
| | ) |
| |
|
| | with gr.Column(scale=2): |
| | extension_percentage = gr.Slider( |
| | minimum=0, |
| | maximum=30, |
| | step=1.0, |
| | value=10, |
| | label="上下高さ拡張率(%)" |
| | ) |
| |
|
| | |
| | temperature = gr.Slider( |
| | minimum=0.0, |
| | maximum=1.0, |
| | step=0.1, |
| | value=0.5, |
| | label="生成時の温度(低い=一貫性高、高い=創造性高)", |
| | visible=False |
| | ) |
| |
|
| | |
| | trim_whitespace = gr.Checkbox( |
| | label="余白を自動トリミング", |
| | value=True, |
| | info="生成される画像から余分な空白領域を自動的に削除します" |
| | ) |
| |
|
| | submit_btn = gr.Button("生成") |
| | output_image = gr.Image(type="pil", label="ページ全体のスクリーンショット") |
| |
|
| | |
| | def update_controls_visibility(mode): |
| | |
| | is_text_mode = mode == "テキスト入力" |
| | return [ |
| | gr.update(visible=is_text_mode), |
| | gr.update(visible=is_text_mode), |
| | ] |
| |
|
| | input_mode.change( |
| | fn=update_controls_visibility, |
| | inputs=input_mode, |
| | outputs=[temperature, style_dropdown] |
| | ) |
| |
|
| | |
| | submit_btn.click( |
| | fn=process_input, |
| | inputs=[input_mode, input_text, extension_percentage, temperature, trim_whitespace, style_dropdown], |
| | outputs=output_image |
| | ) |
| |
|
| | |
| | gemini_model = os.environ.get("GEMINI_MODEL", "gemini-1.5-pro") |
| | gr.Markdown(f""" |
| | ## APIエンドポイント |
| | - `/api/screenshot` - HTMLコードからスクリーンショットを生成 |
| | - `/api/text-to-screenshot` - テキストからインフォグラフィックスクリーンショットを生成 |
| | |
| | ## 設定情報 |
| | - 使用モデル: {gemini_model} (環境変数 GEMINI_MODEL で変更可能) |
| | - 対応スタイル: standard, cute, resort, cool, dental |
| | """) |
| |
|
| | |
| | app = gr.mount_gradio_app(app, iface, path="/") |
| |
|
| | |
| | if __name__ == "__main__": |
| | import uvicorn |
| | logger.info("Starting Uvicorn server for local development...") |
| | uvicorn.run(app, host="0.0.0.0", port=7860) |