initialization

.env

@@ -0,0 +1,3 @@
+API_KEY=123456
+TYPHOON_OCR_API_KEY=sk-D1sMp4fivWSKNhnydFOvqFwzdxqK7OIsVoLHn6rN3komSG3L
+TYPHOON_BASE_URL=https://api.opentyphoon.ai/v1

app.py

@@ -0,0 +1,120 @@
+import os
+import tempfile
+from fastapi import FastAPI, UploadFile, File, HTTPException, Header
+from fastapi.middleware.cors import CORSMiddleware
+from typhoon_ocr import ocr_document
+from pdf2image import convert_from_bytes
+from PIL import Image
+from docx import Document
+from dotenv import load_dotenv
+import gradio as gr
+
+# --- Load env ---
+load_dotenv()
+API_KEY = os.getenv("API_KEY")
+TYPHOON_API_KEY = os.getenv("TYPHOON_OCR_API_KEY")
+TYPHOON_BASE_URL = os.getenv("TYPHOON_BASE_URL", "https://api.opentyphoon.ai/v1")
+
+# --- FastAPI Init ---
+app = FastAPI()
+
+app.add_middleware(
+    CORSMiddleware,
+    allow_origins=["*"],
+    allow_methods=["*"],
+    allow_headers=["*"],
+)
+
+# --- PDF to PIL images ---
+def pdf_to_images(pdf_bytes: bytes) -> list[Image.Image]:
+    return convert_from_bytes(pdf_bytes)
+
+# --- Core OCR Handler ---
+def process_ocr(file_path: str, task_type: str, page_num: int):
+    md_result = ""
+    doc = Document()
+
+    if file_path.lower().endswith(".pdf"):
+        with open(file_path, "rb") as f:
+            images = pdf_to_images(f.read())
+
+        pages = range(len(images)) if page_num <= 0 else [page_num - 1]
+
+        for i in pages:
+            img = images[i]
+            result = ocr_document(img, task_type=task_type)
+            text = result if isinstance(result, str) else result.get("text", "")
+            md_result += f"\n## หน้า {i + 1}\n{text.strip()}\n"
+            doc.add_heading(f"หน้า {i + 1}", level=1)
+            doc.add_paragraph(text.strip())
+    else:
+        result = ocr_document(file_path, task_type=task_type)
+        text = result if isinstance(result, str) else result.get("text", "")
+        md_result = text.strip()
+        doc.add_paragraph(text.strip())
+
+    # Save .md
+    md_file = tempfile.NamedTemporaryFile(delete=False, suffix=".md", mode="w", encoding="utf-8")
+    md_file.write(md_result)
+    md_file.close()
+
+    # Save .docx
+    docx_file = tempfile.NamedTemporaryFile(delete=False, suffix=".docx")
+    doc.save(docx_file.name)
+
+    return md_result, md_file.name, docx_file.name
+
+# --- API Endpoint ---
+@app.post("/api/ocr_document")
+async def ocr_endpoint(
+    file: UploadFile = File(...),
+    task_type: str = "default",
+    page_num: int = -1,
+    x_api_key: str | None = Header(None)
+):
+    if API_KEY and x_api_key != API_KEY:
+        raise HTTPException(status_code=401, detail="API key ผิดพ่อง")
+
+    try:
+        with tempfile.NamedTemporaryFile(delete=False, suffix=os.path.splitext(file.filename)[-1]) as tmp:
+            tmp.write(await file.read())
+            tmp_path = tmp.name
+
+        md, md_path, docx_path = process_ocr(tmp_path, task_type, page_num)
+
+        return {
+            "markdown": md,
+            "md_file": md_path,
+            "docx_file": docx_path,
+            "task_type": task_type,
+        }
+
+    except Exception as e:
+        raise HTTPException(status_code=500, detail=str(e))
+
+# --- Gradio UI ---
+def gradio_handler(file, task_type, page_num):
+    if file is None:
+        return "❌ อัปโหลดไฟล์ก่อน ไอ้สัส", None, None
+    try:
+        md, md_path, docx_path = process_ocr(file.name, task_type, page_num)
+        return md, md_path, docx_path
+    except Exception as e:
+        return f"⚠️ Error: {str(e)}", None, None
+
+with gr.Blocks() as demo:
+    gr.Markdown("## 📄 Typhoon OCR (ทุกหน้า) | Text Extractor | PDF / JPG")
+    with gr.Row():
+        file_input = gr.File(label="📤 อัปโหลด PDF หรือ รูป", file_types=[".pdf", ".jpg", ".jpeg", ".png"])
+        page_input = gr.Number(value=-1, label="📄 เลขหน้า (ใส่ -1 เพื่อแปลงทุกหน้า)")
+        task_type = gr.Radio(["default", "structure"], label="📌 OCR Mode", value="default")
+    ocr_output = gr.Textbox(label="📋 ผล Markdown", lines=20)
+    with gr.Row():
+        btn = gr.Button("🚀 OCR Now!")
+        md_out = gr.File(label="📥 .md ไฟล์")
+        docx_out = gr.File(label="📥 .docx ไฟล์")
+    btn.click(fn=gradio_handler, inputs=[file_input, task_type, page_input], outputs=[ocr_output, md_out, docx_out])
+
+demo.launch()
+# หรือถ้าจะ mount เข้า FastAPI route
+# app = gr.mount_gradio_app(app, demo, path="/ui")

packages.txt

@@ -0,0 +1 @@
+poppler-utils

requirements.txt

@@ -0,0 +1,9 @@
+fastapi
+gradio
+python-multipart
+python-docx
+typhoon-ocr
+pdf2image
+Pillow
+python-dotenv
+uvicorn