Add Streamlit web UI and Docker support

- app.py: Streamlit-based web UI for file-to-markdown conversion - Dockerfile.ui: Docker image for UI (ffmpeg, ExifTool, port 8501) - .dockerignore: whitelist app.py for Docker build - CLAUDE.md: project instructions for Claude Code Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-04-13 15:10:41 +09:00
parent 63cbbd9de6
commit 7b040a4445
4 changed files with 166 additions and 0 deletions
@@ -0,0 +1,117 @@
+import io
+import tempfile
+import os
+import streamlit as st
+from markitdown import MarkItDown
+
+st.set_page_config(
+    page_title="MarkItDown",
+    page_icon="📄",
+    layout="wide",
+)
+
+st.title("📄 MarkItDown")
+st.caption("파일을 Markdown으로 변환합니다")
+
+SUPPORTED_EXTENSIONS = [
+    "pdf", "docx", "pptx", "xlsx", "xls",
+    "jpg", "jpeg", "png",
+    "mp3", "wav",
+    "html", "htm",
+    "csv", "json", "xml",
+    "ipynb", "epub", "zip", "msg",
+]
+
+# Sidebar
+with st.sidebar:
+    st.header("설정")
+    show_preview = st.toggle("Markdown 렌더링 미리보기", value=True)
+    st.divider()
+    st.markdown("**지원 포맷**")
+    st.markdown(
+        "PDF · DOCX · PPTX · XLSX · XLS\n\n"
+        "JPG · PNG · MP3 · WAV\n\n"
+        "HTML · CSV · JSON · XML\n\n"
+        "IPYNB · EPUB · ZIP · MSG"
+    )
+
+# URL 변환
+url_tab, file_tab = st.tabs(["URL 변환", "파일 업로드"])
+
+md = MarkItDown()
+
+with url_tab:
+    url = st.text_input("URL 입력", placeholder="https://example.com 또는 YouTube URL")
+    if st.button("변환", key="url_btn", disabled=not url):
+        with st.spinner("변환 중..."):
+            try:
+                result = md.convert(url)
+                st.session_state["url_result"] = result.text_content
+                st.session_state["url_filename"] = "output.md"
+            except Exception as e:
+                st.error(f"변환 실패: {e}")
+
+    if "url_result" in st.session_state:
+        _content = st.session_state["url_result"]
+        col1, col2 = st.columns([1, 1]) if show_preview else (st.container(), None)
+
+        with col1:
+            st.subheader("Markdown 원문")
+            st.code(_content, language="markdown")
+
+        if show_preview and col2:
+            with col2:
+                st.subheader("미리보기")
+                st.markdown(_content)
+
+        st.download_button(
+            "⬇️ .md 파일 다운로드",
+            data=_content,
+            file_name=st.session_state["url_filename"],
+            mime="text/markdown",
+        )
+
+with file_tab:
+    uploaded = st.file_uploader(
+        "파일을 끌어다 놓거나 클릭해서 선택하세요",
+        type=SUPPORTED_EXTENSIONS,
+    )
+
+    if uploaded is not None:
+        if st.button("변환", key="file_btn"):
+            with st.spinner("변환 중..."):
+                try:
+                    suffix = os.path.splitext(uploaded.name)[1]
+                    with tempfile.NamedTemporaryFile(delete=False, suffix=suffix) as tmp:
+                        tmp.write(uploaded.getvalue())
+                        tmp_path = tmp.name
+
+                    result = md.convert(tmp_path)
+                    os.unlink(tmp_path)
+
+                    st.session_state["file_result"] = result.text_content
+                    st.session_state["file_filename"] = os.path.splitext(uploaded.name)[0] + ".md"
+                except Exception as e:
+                    st.error(f"변환 실패: {e}")
+
+    if "file_result" in st.session_state:
+        _content = st.session_state["file_result"]
+
+        if show_preview:
+            col1, col2 = st.columns([1, 1])
+            with col1:
+                st.subheader("Markdown 원문")
+                st.code(_content, language="markdown")
+            with col2:
+                st.subheader("미리보기")
+                st.markdown(_content)
+        else:
+            st.subheader("Markdown 원문")
+            st.code(_content, language="markdown")
+
+        st.download_button(
+            "⬇️ .md 파일 다운로드",
+            data=_content,
+            file_name=st.session_state["file_filename"],
+            mime="text/markdown",
+        )