Scale app: Added .docx support, download feature, and cleaned up DEPLOY.md

DEPLOY.md

@@ -1,35 +0,0 @@
-# 🌍 Deploying to a Public URL
-
-To share your Clinical De-identification Dashboard with others, follow these steps to deploy it to **Hugging Face Spaces**.
-
-## Prerequisites
-1. A [Hugging Face](https://huggingface.co/join) account.
-2. Your project already pushed to your [GitHub Repository](https://github.com/sarvanithin/clinical-deidentify).
-
-## Step-by-Step Deployment
-
-### 1. Create a New Space
-- Go to [huggingface.co/spaces](https://huggingface.co/spaces) and click **"Create new Space"**.
-- **Space Name**: Give it a name (e.g., `clinical-deidentify`).
-- **SDK**: Select **Docker**.
-- **Docker Template**: Choose **Blank**.
-- **Pricing**: The **Free (16GB RAM)** tier is perfect for this models.
-
-### 2. Connect to GitHub
-- Instead of uploading files manually, click on the **"Settings"** tab of your new Space.
-- Scroll down to **"GitHub Repo"**.
-- Click **"Connect"** and select your `clinical-deidentify` repository.
-- Ensure the branch is set to `main`.
-
-### 3. Automatic Deployment
-- Once connected, Hugging Face will detect your `Dockerfile`.
-- It will automatically start building the image and downloading the transformer model.
-- After a few minutes, your dashboard will be live at `https://huggingface.co/spaces/YOUR_USER/clinical-deidentify`!
-
-## Why Hugging Face Spaces?
-- **High RAM**: Clinical-NER models require several GBs of RAM; HF Spaces provides 16GB for free.
-- **CI/CD**: Every time you `git push` to GitHub, your public URL will automatically update.
-- **Privacy**: You can set the Space to **Private** if you only want specific people to access it.
-
----
-*Note: Ensure you have performed the final `git push` from your local machine before connecting to GitHub.*

app/main.py

@@ -3,6 +3,7 @@ from fastapi.staticfiles import StaticFiles
 from fastapi.responses import FileResponse
 from fastapi.middleware.cors import CORSMiddleware
 import fitz  # PyMuPDF
+import docx  # python-docx
 from .models import DeidRequest, DeidResponse, BatchDeidRequest, BatchDeidResponse, FeedbackRequest
 from .pipeline.hybrid import DeidPipeline
 import json
@@ -87,8 +88,13 @@ async def deidentify_file(file: UploadFile = File(...)):
             doc.close()
         elif filename.endswith(".txt"):
             text = content.decode("utf-8")
+        elif filename.endswith(".docx"):
+            # Extract text from Word document
+            from io import BytesIO
+            doc = docx.Document(BytesIO(content))
+            text = "\n".join([para.text for para in doc.paragraphs])
         else:
-            raise HTTPException(status_code=400, detail="Unsupported file type. Please upload .pdf or .txt")
+            raise HTTPException(status_code=400, detail="Unsupported file type. Please upload .pdf, .txt, or .docx")
 
         if not text.strip():
             raise HTTPException(status_code=400, detail="File is empty or no text could be extracted.")

app/static/main.js

@@ -172,5 +172,18 @@ document.addEventListener('DOMContentLoaded', () => {
         const originalIcon = copyBtn.innerHTML;
         copyBtn.innerHTML = '<svg viewBox="0 0 24 24" width="18" height="18" style="color:var(--success)"><path fill="currentColor" d="M9 16.17L4.83 12l-1.42 1.41L9 19 21 7l-1.41-1.41z"/></svg>';
         setTimeout(() => copyBtn.innerHTML = originalIcon, 2000);
+        // Download logic
+        downloadBtn.addEventListener('click', () => {
+            if (!currentDeidentifiedText) return;
+
+            const blob = new Blob([currentDeidentifiedText], { type: 'text/plain' });
+            const url = URL.createObjectURL(blob);
+            const a = document.createElement('a');
+            a.href = url;
+            a.download = `deidentified_${new Date().getTime()}.txt`;
+            document.body.appendChild(a);
+            a.click();
+            document.body.removeChild(a);
+            URL.revokeObjectURL(url);
+        });
     });
-});

requirements.txt

@@ -7,3 +7,4 @@ spacy>=3.7.2
 pytest>=7.4.3
 pymupdf
 python-multipart
+python-docx