Fix HuggingFace folder structure

README.md

@@ -1,10 +1,24 @@
----
-title: Civicconnect
-emoji: 🔥
-colorFrom: purple
-colorTo: green
-sdk: docker
-pinned: false
----
-
-Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference
+Civic ML Backend (Full: pHash + Location)
+
+Overview:
+
+- FastAPI backend that accepts citizen reports and performs:
+  - rule-based abuse detection
+  - text duplicate detection
+  - image duplicate detection using pHash (imagehash)
+  - location-based duplicate detection using Haversine formula
+  - image classification using CLIP (if available) with URL fallback
+  - priority detection (keyword-based)
+
+Run:
+
+1. python -m venv .venv
+2. source .venv/bin/activate # Windows: .venv\Scripts\Activate
+3. pip install -r requirements.txt
+4. uvicorn app.main:app --reload --port 8000
+
+Notes:
+
+- The in-memory stores (seen_reports, seen_image_hashes, seen_locations) are ephemeral and reset on server restart.
+- CLIP model download requires internet and may take time; if unavailable, the system uses URL keyword fallback for image labels.
+- data/dataset.jsonl collects all incoming reports and results for later training/audit.

ml-backend-with-image/README.md

@@ -1,24 +0,0 @@
-Civic ML Backend (Full: pHash + Location)
-
-Overview:
-
-- FastAPI backend that accepts citizen reports and performs:
-  - rule-based abuse detection
-  - text duplicate detection
-  - image duplicate detection using pHash (imagehash)
-  - location-based duplicate detection using Haversine formula
-  - image classification using CLIP (if available) with URL fallback
-  - priority detection (keyword-based)
-
-Run:
-
-1. python -m venv .venv
-2. source .venv/bin/activate # Windows: .venv\Scripts\Activate
-3. pip install -r requirements.txt
-4. uvicorn app.main:app --reload --port 8000
-
-Notes:
-
-- The in-memory stores (seen_reports, seen_image_hashes, seen_locations) are ephemeral and reset on server restart.
-- CLIP model download requires internet and may take time; if unavailable, the system uses URL keyword fallback for image labels.
-- data/dataset.jsonl collects all incoming reports and results for later training/audit.