app.py

import gradio as gr
import os
import subprocess
import re
from tinydb import TinyDB, Query
import datetime
from datetime import timedelta
from bs4 import BeautifulSoup
import warnings
from article_generator import generate_article
from tavily_search import create_tavily_search_ui, tavily_search_interface
import time

def is_valid_html(html):
    with warnings.catch_warnings(record=True) as w:
        warnings.simplefilter('always')
        BeautifulSoup(html, 'html.parser')
        return len(w) == 0

# データベースの初期化
db = TinyDB("db.json")
current_time = datetime.datetime.now()
one_day_ago = current_time - timedelta(days=1)
db.remove(Query().timestamp.test(lambda x: datetime.datetime.fromisoformat(x) <= one_day_ago))

def process_article(new_keyword, other_keywords, action, edited_article_structure):
    keyword_id = re.sub(r"\W+", "", new_keyword) if new_keyword else None
    response_text = ""
    final_article_content = ""
    final_article_html = ""

    if action == "構成作成":
        all_keywords = new_keyword
        if other_keywords:
            all_keywords += " " + other_keywords
        print(f"Running keywords_processor.py with arguments: {new_keyword}, {other_keywords}")
        result = subprocess.run(["python3", "keywords_processor.py", new_keyword, other_keywords], capture_output=True, text=True)
        print(f"Return code: {result.returncode}")
        if result.returncode != 0:
            response_text = "Keywords Processor Error: " + result.stderr
            print("Keywords Processor Error:", result.stderr)
        else:
            response_text = result.stdout
            print("Keywords Processor Output:", result.stdout)

            # headline_generator.py を実行
            print("Running headline_generator.py")
            headline_result = subprocess.run(["python3", "headline_generator.py", new_keyword], capture_output=True, text=True)
            print(f"Headline Generator Return code: {headline_result.returncode}")
            if headline_result.returncode != 0:
                print("Headline Generator Error:", headline_result.stderr)
            else:
                response_text += headline_result.stdout
                print("Headline Generator Output:", headline_result.stdout)

        # 編集可能な記事構成にのみ出力（HTML内容は空のまま）
        return response_text, "", ""

    elif action == "本文作成":
        response_text = edited_article_structure
        try:
            final_article_content, final_article_html = generate_article(edited_article_structure)

            # output3.txt ファイルが作成されるのを待つ
            while not os.path.exists("output3.txt"):
                time.sleep(0.1)

            with open("output3.txt", "r", encoding="utf-8") as f:
                final_article_content = f.read()
                # final_article_htmlをgr.HTMLオブジェクトとして再初期化
                final_article_html = gr.HTML(value=final_article_content, visible=True)

            # デバッグ出力追加
            print("After updating from output3.txt - final_article_html value:", final_article_html)
            print("After updating from output3.txt - final_article_html type:", type(final_article_html))
        except Exception as e:
            final_article_content = f"予期せぬエラーが発生しました：{str(e)}"
            final_article_html = gr.HTML(visible=False)  # ここでも例外が発生した場合に備えて適切なオブジェクトを設定
            print(f"予期せぬエラーが発生しました：{str(e)}")
            print("final_article_html after error:", final_article_html)
            print("final_article_html type after error:", type(final_article_html))
      
        return response_text, final_article_content, final_article_html

def custom_html_to_markdown(html):
    soup = BeautifulSoup(html, 'html.parser')

    # 不要なタグの除去
    for tag in soup(['script', 'style']):
        tag.decompose()

    # タグごとの処理
    for h in soup.find_all('h1'):
        h.replace_with(f"# {h.get_text().strip()}\n\n")
    for h in soup.find_all('h2'):
        h.replace_with(f"## {h.get_text().strip()}\n\n")
    for h in soup.find_all('h3'):
        h.replace_with(f"### {h.get_text().strip()}\n\n")
    for p in soup.find_all('p'):
        p.replace_with(f"{p.get_text().strip()}\n\n")

    # 最終的なMarkdownテキストの取得
    return soup.get_text()

def display_content(content, formatted_final_article_md, formatted_final_article_html):
    print("Type of formatted_final_article_html before processing in display_content:", type(formatted_final_article_html))
    if type(formatted_final_article_html) is not gr.components.HTML:
        print("Error: formatted_final_article_html is not a gr.HTML object")
        return formatted_final_article_md, formatted_final_article_html  # エラーがあれば早期リターン

    markdown_content = custom_html_to_markdown(content)
    formatted_final_article_md.update(value=markdown_content, visible=True)
    formatted_final_article_html.update(visible=False) # HTMLを非表示にする
    return formatted_final_article_md, formatted_final_article_html

# Gradioアプリの設定
with gr.Blocks(css='''
    .gr-markdown h1, .gr-markdown h2, .gr-markdown h3 {
        word-wrap: break-word;
        overflow-wrap: break-word;
    }
''') as app:
    with gr.Row():
        with gr.Column():
            gr.Markdown("### BabyWriterPRO")
            new_keyword = gr.Textbox(label="キーワード")
            other_keywords = gr.Textbox(label="その他のキーワード", lines=4)
            create_structure_button = gr.Button("構成作成")
            editable_output2 = gr.Textbox(label="編集可能な記事構成", lines=10, placeholder="記事構成がここに表示されます")
            create_article_button = gr.Button("本文作成")
            formatted_final_article_md = gr.Markdown(value="", visible=True, label="最終的な記事本文")
            formatted_final_article_html = gr.HTML(value="", visible=False, label="最終的な記事本文(HTML)")

            create_structure_button.click(
                fn=process_article,
                inputs=[new_keyword, other_keywords, gr.State(value="構成作成"), editable_output2],
                outputs=[editable_output2, formatted_final_article_md]
            )

            create_article_button.click(
                fn=process_article,
                inputs=[new_keyword, other_keywords, gr.State(value="本文作成"), editable_output2],
                outputs=[editable_output2, formatted_final_article_md]
            )

            print("Setup change event for display_format with MD and HTML components")

        with gr.Column():
            gr.Markdown("&nbsp;", elem_id="spacer")  # 空白を作るための見えないMarkdown
            selected_text = gr.Textbox(label="検索キーワード", lines=3, elem_classes="right-align")
            domain_filter = gr.Radio(label="検索フィルタ（all：全検索、official：公式、企業、政府系）", choices=["all", "official"], value="all", elem_classes="right-align")
            search_button = gr.Button("検索", elem_classes="right-align")
            search_display_format = gr.Radio(label="表示形式", choices=["Markdown", "HTML"], value="Markdown", elem_classes="right-align")
            search_results_md = gr.Markdown(elem_classes="right-align")
            search_results_html = gr.HTML(elem_classes="right-align", visible=False)

            def switch_search_display_format(format):
                if format == "Markdown":
                    return gr.update(visible=True), gr.update(visible=False)
                else:
                    return gr.update(visible=False), gr.update(visible=True)

            search_button.click(tavily_search_interface, inputs=[selected_text, domain_filter], outputs=[search_results_md, search_results_html])
            search_display_format.change(switch_search_display_format, inputs=[search_display_format], outputs=[search_results_md, search_results_html])

    port = int(os.getenv('PORT', 7860))
    app.launch(server_name="0.0.0.0", server_port=port)