feat: invert-mask option in UI and core; add docs (no secrets)

API_Usage_Guide.md

@@ -0,0 +1,217 @@
+# API Usage Guide - Photo Object Removal
+
+## Overview
+This guide provides step-by-step instructions for using the Photo Object Removal API to remove objects from images using AI inpainting.
+
+**Base URL:** `https://logicgoinfotechspaces-object-remover.hf.space`  
+**Authentication:** Bearer token (optional)
+
+## Quick Start
+
+### 1. Health Check
+Verify the API is running:
+```bash
+curl -H "Authorization: Bearer <API_TOKEN>" \
+  https://logicgoinfotechspaces-object-remover.hf.space/health
+```
+**Response:** `{"status":"healthy"}`
+
+### 2. Upload Image
+Upload the image you want to edit:
+```bash
+curl -H "Authorization: Bearer <API_TOKEN>" \
+  -F image=@your_image.jpg \
+  https://logicgoinfotechspaces-object-remover.hf.space/upload-image
+```
+**Response:** `{"id":"9cf61445-f83b-4c97-9272-c81647f90d68","filename":"your_image.jpg"}`
+
+### 3. Upload Mask
+Upload a mask showing areas to remove (white/colored areas = remove):
+```bash
+curl -H "Authorization: Bearer <API_TOKEN>" \
+  -F [email protected] \
+  https://logicgoinfotechspaces-object-remover.hf.space/upload-mask
+```
+**Response:** `{"id":"d044a390-dde2-408a-b7cf-d508385e56ed","filename":"mask.png"}`
+
+### 4. Process Image
+Remove objects using the uploaded image and mask:
+```bash
+curl -H "Authorization: Bearer <API_TOKEN>" \
+  -H "Content-Type: application/json" \
+  -d '{"image_id":"9cf61445-f83b-4c97-9272-c81647f90d68","mask_id":"d044a390-dde2-408a-b7cf-d508385e56ed"}' \
+  https://logicgoinfotechspaces-object-remover.hf.space/inpaint
+```
+**Response:** `{"result":"output_b09568698bbd4aa591b1598c01f2f745.png"}`
+
+### 5. View Result
+Open the result image in your browser:
+```
+https://logicgoinfotechspaces-object-remover.hf.space/result/output_b09568698bbd4aa591b1598c01f2f745.png
+```
+
+## Alternative Methods
+
+### Method 1: Get Direct URL
+Use `/inpaint-url` to get a shareable URL:
+```bash
+curl -H "Authorization: Bearer <API_TOKEN>" \
+  -H "Content-Type: application/json" \
+  -d '{"image_id":"<image_id>","mask_id":"<mask_id>"}' \
+  https://logicgoinfotechspaces-object-remover.hf.space/inpaint-url
+```
+**Response:** 
+```json
+{
+  "result":"output_xxx.png",
+  "url":"https://logicgoinfotechspaces-object-remover.hf.space/download/output_xxx.png"
+}
+```
+
+### Method 2: One-Step Processing
+Upload and process in a single request:
+```bash
+curl -H "Authorization: Bearer <API_TOKEN>" \
+  -F [email protected] \
+  -F [email protected] \
+  https://logicgoinfotechspaces-object-remover.hf.space/inpaint-multipart
+```
+
+### Method 3: Download Result
+Download the result file:
+```bash
+curl -L https://logicgoinfotechspaces-object-remover.hf.space/download/output_xxx.png \
+  -o result.png
+```
+
+## Postman Setup
+
+### 1. Upload Image
+- **Method:** POST
+- **URL:** `https://logicgoinfotechspaces-object-remover.hf.space/upload-image`
+- **Headers:** `Authorization: Bearer <API_TOKEN>`
+- **Body:** form-data
+  - Key: `image` (Type: File)
+  - Value: Select your image file
+
+### 2. Upload Mask
+- **Method:** POST
+- **URL:** `https://logicgoinfotechspaces-object-remover.hf.space/upload-mask`
+- **Headers:** `Authorization: Bearer <API_TOKEN>`
+- **Body:** form-data
+  - Key: `mask` (Type: File)
+  - Value: Select your mask file
+
+### 3. Process Image
+- **Method:** POST
+- **URL:** `https://logicgoinfotechspaces-object-remover.hf.space/inpaint`
+- **Headers:** 
+  - `Authorization: Bearer <API_TOKEN>`
+  - `Content-Type: application/json`
+- **Body:** raw JSON
+```json
+{
+  "image_id": "9cf61445-f83b-4c97-9272-c81647f90d68",
+  "mask_id": "d044a390-dde2-408a-b7cf-d508385e56ed"
+}
+```
+
+## Mask Creation Guide
+
+### Mask Formats
+- **RGBA PNG:** Pixels with alpha=0 are treated as areas to remove
+- **RGB/Grayscale:** Pixels with value > 0 are treated as areas to remove
+
+### Creating Masks
+1. **Paint Method:** Use any image editor to paint white/colored areas over objects you want to remove
+2. **Transparency Method:** Create a PNG with transparent areas where you want objects removed
+3. **Grayscale Method:** Create a black image and paint white areas over objects to remove
+
+### Tips
+- Use high contrast (white on black) for best results
+- Make sure mask areas are clearly defined
+- Avoid thin lines or small details in the mask
+
+## Error Handling
+
+### Common Errors
+
+**422 Unprocessable Entity**
+- **Cause:** Missing required fields or invalid file format
+- **Solution:** Check file upload and field names
+
+**404 Not Found**
+- **Cause:** Invalid image_id or mask_id
+- **Solution:** Re-upload files to get fresh IDs
+
+**401 Unauthorized**
+- **Cause:** Missing or invalid API token
+- **Solution:** Add correct Authorization header
+
+**403 Forbidden**
+- **Cause:** Wrong API token
+- **Solution:** Use the correct token you configured
+
+### Troubleshooting
+
+**File Upload Issues in Postman:**
+- Remove warning icons next to filenames
+- Re-select files if warning appears
+- Try different file formats (PNG, JPG)
+- Check file isn't corrupted
+
+**Slow Processing:**
+- Large images take longer to process
+- Typical processing time: 1-3 seconds
+- Wait for completion before checking results
+
+## Complete Workflow Example
+
+```bash
+# 1. Health check
+curl -H "Authorization: Bearer <API_TOKEN>" \
+  https://logicgoinfotechspaces-object-remover.hf.space/health
+
+# 2. Upload image
+curl -H "Authorization: Bearer <API_TOKEN>" \
+  -F [email protected] \
+  https://logicgoinfotechspaces-object-remover.hf.space/upload-image
+
+# 3. Upload mask
+curl -H "Authorization: Bearer <API_TOKEN>" \
+  -F [email protected] \
+  https://logicgoinfotechspaces-object-remover.hf.space/upload-mask
+
+# 4. Process (replace with actual IDs from steps 2&3)
+curl -H "Authorization: Bearer <API_TOKEN>" \
+  -H "Content-Type: application/json" \
+  -d '{"image_id":"IMAGE_ID","mask_id":"MASK_ID"}' \
+  https://logicgoinfotechspaces-object-remover.hf.space/inpaint
+
+# 5. View result (replace with actual filename from step 4)
+# Open in browser: https://logicgoinfotechspaces-object-remover.hf.space/result/OUTPUT_FILENAME
+```
+
+## API Endpoints Summary
+
+| Method | Endpoint | Description | Auth Required |
+|--------|----------|-------------|---------------|
+| GET | `/health` | Health check | No |
+| GET | `/` | API information | No |
+| POST | `/upload-image` | Upload image file | Optional |
+| POST | `/upload-mask` | Upload mask file | Optional |
+| POST | `/inpaint` | Process with IDs | Optional |
+| POST | `/inpaint-url` | Process with URL response | Optional |
+| POST | `/inpaint-multipart` | One-step processing | Optional |
+| GET | `/download/{filename}` | Download result | No |
+| GET | `/result/{filename}` | View result in browser | No |
+| GET | `/logs` | View processing logs | Optional |
+| GET | `/docs` | API documentation | No |
+
+## Support
+
+For issues or questions:
+1. Check the API documentation at `/docs`
+2. Verify your request format matches examples
+3. Ensure files are properly uploaded
+4. Check authentication token is correct

app.py

@@ -82,6 +82,7 @@ def run_streamlit_ui():
         img_input = img_input.resize((new_width, new_height))
     
     stroke_width = st.slider("Brush size", 1, 100, 50)
+    invert_mask = st.checkbox("Painted area = remove", value=True)
 
     st.write("**Now draw (brush) the part of image that you want to remove.**")
     
@@ -121,11 +122,17 @@ def run_streamlit_ui():
         
         if st.button('Submit'):
             with st.spinner("AI is doing the magic!"):
-                output = process_inpaint(np.array(img_input), np.array(im))
+                output = process_inpaint(np.array(img_input), np.array(im), invert_mask=invert_mask)
                 img_output = Image.fromarray(output).convert("RGB")
             
             st.write("AI has finished the job!")
-            st.image(img_output)
+            col1, col2 = st.columns(2)
+            with col1:
+                st.caption("Before")
+                st.image(img_input.convert("RGB"))
+            with col2:
+                st.caption("After")
+                st.image(img_output)
             
             uploaded_name = os.path.splitext(uploaded_file.name)[0]
             image_download_button(

src/core.py

@@ -442,7 +442,7 @@ def get_args_parser():
     return parser.parse_args()
 
 
-def process_inpaint(image, mask):
+def process_inpaint(image, mask, invert_mask=True):
     image = cv2.cvtColor(image, cv2.COLOR_RGBA2RGB)
     original_shape = image.shape
     interpolation = cv2.INTER_CUBIC
@@ -458,7 +458,10 @@ def process_inpaint(image, mask):
     print(f"Resized image shape: {image.shape}")
     image = norm_img(image)
 
-    mask = 255-mask[:,:,3]
+    # Convert RGBA mask to single-channel mask.
+    # When invert_mask=True (default), areas the user paints (alpha=0) become 255 (to remove).
+    alpha_channel = mask[:,:,3]
+    mask = (255 - alpha_channel) if invert_mask else alpha_channel
     mask = resize_max_size(mask, size_limit=size_limit, interpolation=interpolation)
     mask = norm_img(mask)